The earliest stage of any defense production cycle is not procurement, it’s staffing.

Multiple geopolitical tensions are currently forcing governments to reassess weapons stockpiles and industrial capacity. What is less visible is how the defense manufacturing system itself is responding to that pressure.

There is a persistent mistake in the way most observers attempt to understand the defense industry. They watch the wrong signals.

Public discussion tends to revolve around geopolitical headlines, government budgets, and contract announcements as if those events drive the industrial behavior of the sector. In reality they tend to be the final stage of a much longer process. By the time a procurement contract is announced or a defense company references backlog expansion on an earnings call, the underlying industrial adjustments that make that production possible have already been underway for months.

Weapons production isn’t just policy or political… it is industrial at the core.

We are witnessing the repositioning of the defense industrial base.

Factories must be staffed before production lines can scale.

Companies operating inside the manufacturing pipeline were already preparing their workforce for increased throughput well before the political narrative caught up.

Engineering teams must be assembled before complex systems can be integrated and validated. Testing environments must exist before those systems can transition into operational deployment. The defense economy moves in stages, and the earliest of those stages almost always appears in workforce architecture long before it appears in financial disclosures and it is crucial to understand what is coming.

When companies begin adjusting the composition of their engineering workforce, particularly in highly specialized manufacturing roles, they are straight up telling us where capacity will exist in the future rather than describing what exists today.

Over the past several months I have been examining hiring composition across multiple U.S. defense contractors. Not just the number of engineers being hired, which fluctuates constantly in large technical organizations, but the specific categories of roles those companies are choosing to expand relative to their historical baseline. The workforce is a living, breathing entity and can really tell us cyclical patterns and structural changes. Modern conflict is a contest of industrial capacity.

What emerges from that analysis is a pattern that is difficult to dismiss as routine hiring noise. This is typically the noise you hear surrounding new contracts or just any general posts on social media, especially concerning the overall labor market or even the geopolitical scene. Multiple segments of the defense industrial ecosystem appear to be expanding simultaneously, not just inside one contractor, and not within a single weapons program, but across the upstream manufacturing layer, the midstream integration layer, and the downstream battlefield infrastructure layer.

When those three segments begin adjusting at the same time, the industrial system is usually preparing for something larger than a standard procurement cycle.

Read more

Oracle is about to spend tens of billions on AI data centers and GPUs, and when infrastructure becomes the primary asset of the business, labor gets repriced. The connection between labor and repricing and AI is the #1 question hedge funds are asking me right now.

Bloomberg is now reporting thousands of layoffs and hiring freezes across the cloud division. I flagged those signals again on February 19th.

Being early to these shifts matters. Because the layoffs aren’t the story at all… they’re just the mechanism.

Ok let’s walk it back real quick when I published an analysis Jan 2nd Oracle’s 23 Year Loop arguing that something unusual was happening inside Oracle’s workforce, the signal I was watching wasn’t the headline layoffs everyone focuses on when tech companies restructure. It was the composition of the workforce itself… the percentage shifts in who was being exited out of the system and who was being rebuilt to replace them.

So, when Bloomberg reported that Oracle is preparing to cut thousands of jobs across the company while simultaneously accelerating one of the largest AI data-center expansions in its history, most readers will treat those two developments as separate stories.

They aren’t.

Read more

With tensions in the Middle East rising again and energy markets reacting in real time, the instinct in financial markets is to focus on oil… and that instinct is understandable.

If you want the earliest read on whether this remains a “logistics shock” or becomes a structural repricing, stop staring at crude and start looking at the term structure of operational constraint including the labor-market micro-signals that follow those constraints, including transportation/warehousing softness, job-reallocation pressure in energy intensive sectors, and defense production ramps that pull skilled labor into resilience functions. When that moves together… as they did this week… you are watching the global economy rewrite its org chart in real time, which is exactly where capital allocation becomes destiny for the next 6–18 months.

Roughly one-fifth of global oil and petroleum product consumption, and about one-fifth of global LNG trade has been tied to flows through the Strait of Hormuz in recent years, so even the risk of disruption historically pushes prices higher as markets price tighter supply conditions rapidly.

This first week of March has been the “risk becomes constraint” version of that dynamic with tanker traffic through Hormuz slowed in ship-tracking estimates, while freight and insurance repriced in near-real-time.

However, the more important story for investors rarely sits in the oil price itself. The bigger story is really in how companies react operationally to rising energy costs and supply uncertainty. It’s been a busy week for me looking at various sectors and trying to see which ones have already started to price this in and also which ones will be forced to and how soon.

Energy sits at the foundation of the modern production chain. Transportation networks, aviation fuel, petrochemicals, plastics, shipping systems, manufacturing processes, and logistics infrastructure all depend on it. When oil prices rise or even appear likely to rise, the baseline cost of operating across the global economy shifts upward. Markets react to that change quickly because prices can adjust instantly… corporate systems move more slowly, but the first operational levers like routing, coverage, and scheduling are already getting warmed up.

Inside companies, the first response to rising energy volatility is almost never layoffs. The first response is margin protection. Hiring plans begin to slow, expansion initiatives are re-evaluated, and leadership teams start asking how much additional output can be generated from existing teams rather than how quickly headcount should grow.

Historically this pattern appears repeatedly. During the oil shocks of the late 1970s companies became extremely cautious about hiring because input costs were volatile and difficult to forecast. A similar dynamic appeared during the 2007–2008 commodity surge when oil moved above $140 per barrel. Transportation firms, manufacturers, and logistics operators began tightening hiring plans months before the broader labor market weakened. The point is not the headline unemployment rate… it is the internal pivot from growth budgets to resilience budgets. We are just weeks to months away from seeing this shock depending on each sector.

The workforce layer of the economy tends to react with a lag relative to financial markets. But when it begins to move, it honestly reveals far more about corporate expectations than market commentary does. For long-horizon capital, workforce architecture pretty much determines which systems continue functioning when volatility hits and which ones break.

The signals worth watching in the current environment are not unemployment prints or headline labor statistics. The signals appear first in hiring behavior within sectors that are most sensitive to energy costs and transport continuity, and this week provided unusually direct “live” proxies: booking stops, emergency freight increases, route suspensions, and air cargo capacity constraints rather than formal layoff announcements. Step one.

Airlines, freight operators, logistics companies, and heavy manufacturing firms typically begin slowing expansion hiring when energy volatility rises. Roles associated with growth initiatives, geographic expansion, or new product lines tend to be paused first, while core operational positions remain open because companies still need to maintain production and service delivery. This week’s logistics commentary made the “capacity constraint” point explicit…. the freight forwarder Kuehne+Nagel cited material air cargo capacity being offline and warned that if customers shift from sea to air freight, mismatches can intensify quickly. Let’s not forget the major cost-reduction program they announced the first few days in March where they said they expected to eliminate over 2,000 full-time positions by the end of 2026. These layoffs are part of a strategy to save approximately $193 million.

Interesting timing, or just a coincidence?

This transition we are seeing already reflects a broad shift in corporate posture where organizations move from expansion mode to optimization mode. I’ve watched a few make this shift in the last few months even prior to the war movements hit. Waves were already created long before the first missiles struck. The difference between those two states of optimization/expansion is subtle but economically significant. In expansion mode, companies assume stable conditions and allocate capital toward growth initiatives, new teams, and capacity increases. In optimization mode, leadership teams become more focused on efficiency, productivity, and resilience.

Expansion projects move more slowly, hiring becomes more selective, and the emphasis shifts toward extracting greater output from existing structures. We have watched that through the greater part of 2025.

The optimization phase can persist for several quarters even when macroeconomic indicators appear stable on the surface, and it becomes stickier when war-risk insurance, rerouting, and fuel costs remain in the P&L long enough to force contract renegotiations. This is where we are today, even just a few days into this war.

Economic growth may begin to moderate, yet inflation pressures remain elevated. The reason lies in how deeply energy costs are embedded in production systems. Even if demand begins to soften, companies must still operate with higher input costs. For instance, major aviation carriers, American Airlines, Delta and United, all exited out of their fuel hedging programs about 10 years ago.

This in turns makes transportation more expensive, manufacturing processes become more expensive, and shipping networks become more expensive. This week’s realized costs illustrate the mechanism where we now have war-risk premiums for Gulf transits rising sharply and we have already seen them quoted in ranges that imply millions of dollars per voyage for large hull values, while freight benchmarks repriced to record levels. Shock factor, maybe, but it is workforce impact that I am watching here.

Central banks are particularly sensitive to this dynamic. If policymakers cut interest rates too early while energy costs are still feeding through supply chains, inflation can remain persistent even as growth slows. Europe experienced a version of this dynamic in 2022 when energy prices surged and forced central banks to maintain tighter policy for longer than expected despite weakening economic momentum. From the perspective of corporate behavior, the key transmission channel of this environment is capital discipline. This week’s official commentary reflects that the European Central Bank showed reluctance to label an Iran-war energy spike “transitory” given the 2022 lesson-set, while U.S. policymakers said that the oil impulse was potentially a one-off unless it persisted long enough to bleed into broader prices.

This change in capital allocation tends to appear inside organizations well before it becomes visible in macro statistics. Lagging stats are dangerous. It is important to constantly build out future projections, especially with the workforce. Hiring managers face even more scrutiny when requesting new positions, capital expenditure plans are reassessed over and over and boards begin prioritizing investments that increase productivity rather than investments that expand workforce size. In practice this often means greater spending on automation technologies, artificial intelligence tools, supply-chain software, and operational systems designed to allow existing teams to produce more output with fewer additional resources. Sounds very 2026 already. I’m not seeing it across all companies, but their reasons for layoffs or not backfilling roles is not quite AI. Workforce data doesn’t lie. Some automation sure, but it allows they a bit of runway to reposition to save margin.

Another internal dynamic that becomes more common during energy shocks is workforce redeployment. Instead of hiring externally to fill new positions, companies often move employees between teams, delay replacing departing workers, or consolidate functions within existing organizational structures. The labor market does not immediately weaken in such an environment, but the momentum of hiring begins to slow. We have been watching this intently the past 18-24 months already. High-frequency hiring proxies had already pointed to a “low-hire, low-fire” backdrop entering 2026, with the job postings flattening in early January after cooling through 2025 and yes, that matters because a flat baseline makes shocks transmit faster into selectivity and payback in discretionary hiring.

At the same time, geopolitical tensions can generate the opposite workforce dynamic in sectors tied to national security and infrastructure resilience. Defense contractors, energy infrastructure firms, and supply-chain security providers often expand hiring precisely when commercial sectors begin to slow.

This week’s signals are directionally consistent to all of this… the White House planned meetings with major defense contractors to accelerate weapons production and discussed a supplemental budget request tied to operations, implying multi-quarter demand for engineering, manufacturing, and supply-chain labor. “The Companies represented were the CEOs of BAE Systems, Boeing, Honeywell Aerospace, L3Harris Missile Solutions, Lockheed Martin, Northrop Grumman, and Raytheon. The meeting concluded with another meeting scheduled in two months. States all over the Country are bidding for these new Plants.”- per WhiteHouse

Defense industry hiring patterns illustrate this process quite clearly. Workforce expansion in firms tends to follow procurement contracts, production bottleneck remediation, and inventory replacement rather than the initial news cycle surrounding geopolitical tensions. As programs scale, these firms gradually expand engineering teams, advanced manufacturing workforces, cybersecurity specialists, and logistics planning functions. This week added a European illustration as well… missile and rocket-engine capacity expansion plans were publicly reiterated as a response to constrained Western production capacity, which appears to be a workforce constraint as much as a capex one.

The current environment is increasingly shaped by supply-side volatility. Energy security, shipping routes, critical minerals, and the fragmentation of global supply chains have become major sources of uncertainty. When volatility originates in supply systems rather than demand cycles, corporate strategy begins to evolve. Companies become less focused on optimizing purely for cost and scale and more focused on building resilience within their operating models. This week’s LNG shock is maybe the most shocking example too…

I expect to see even more hiring priorities acorss most sectors gradually move away from purely expansion-oriented roles and toward functions that strengthen operational stability. Supply-chain strategy, procurement oversight, infrastructure engineering, logistics planning, and risk management capabilities become more prominent within corporate hiring plans. Over time these adjustments reshape organizational structures because firms begin designing operations that can continue functioning even when inputs become volatile or supply routes become disrupted. Labor response is as much about reallocation as it is about the net level of jobs, particularly in energy-intensive sectors.

For decades global production systems were optimized for efficiency. However, when supply disruptions become more frequent, those lean systems begin to appear fragile. Firms respond by introducing redundancy into their networks and diversifying supplier relationships. The changes of suppliers is already noticeable in the past 6 months. These changes often require new operational capabilities, which in turn influence hiring priorities and workforce architecture. Consulting has picked up heavily in numerous sectors the past 12 months. That is usually the first sign of these supplier changes and upcoming layoffs or hiring freezes.

Financial markets tend to respond to geopolitical shocks immediately because asset prices adjust continuously as new information emerges. Corporate systems, by contrast, operate through slower decision cycles that include supply agreements, infrastructure investments, workforce planning, and capital budgeting processes. As a result, the repricing of economic structures unfolds gradually inside companies. What changed this week is that several “slow-cycle” inputs like shipping lanes, hull war-risk pricing, and LNG restart timelines… and all of that compressed the usual lag and forced corporate decisions into the open.

The upcoming earnings and planning amongst leadership teams will show us adjusting long-term planning assumptions, altering supply chains, reallocating capital, and modifying workforce strategies in response to persistent volatility. Those adjustments can reshape cost structures across entire industries. The following quarter should be a bit volatile and I do expect more disposal activities, exit costs, and restructuring charges in filings.

In this environment the assets most vulnerable to tighter capital conditions are generally those that depend heavily on external funding to sustain growth. Venture-backed technology firms, speculative growth companies, and segments of emerging markets that rely on continuous global capital inflows often find themselves exposed when financing becomes more selective. This dynamic has historically appeared in areas such as late-stage venture software platforms, highly leveraged private-equity portfolio companies, and capital-intensive growth businesses like electric vehicle startups, unprofitable cloud infrastructure providers, or aggressive expansion models similar to those seen at companies like Rivian, Lucid Motors, Snowflake, Databricks, or highly cash-consumptive software firms that scaled rapidly during the era of inexpensive capital.

These ecosystems typically expand rapidly during periods of abundant capital by building teams around product development, growth marketing, and market expansion. When financing conditions tighten, hiring priorities shift toward operational sustainability, financial discipline, and infrastructure functions that extend the runway of existing capital.

Markets eventually begin favoring companies that can fund their own operations through durable cash flows rather than relying on inexpensive financing. This shift rarely appears immediately, but it shows up in corporate hiring patterns and workforce composition 6-12 months ahead of time and if you know what to look for you can time it all correctly.

Share

For investors attempting to interpret the current geopolitical environment, the key question is not whether markets have reacted to the latest headlines. Markets react quickly, they always do… I’m more focused on the question of whether businesses begin making structural adjustments to their operations. Changes in supply-chain strategy, capital investment priorities, and workforce planning typically unfold over months or quarters rather than days. This week’s strongest confirming signal is that firms are already paying the “second-order” price in freight, insurance, rerouting, and emergency surcharges.

Watching hiring patterns across energy-intensive industries, transportation networks, manufacturing supply chains, and defense sectors can therefore provide valuable insight into how companies are interpreting the current environment. Aviation and Defense sectors are also major areas to watch. If hiring slows in energy-sensitive sectors while resilience-oriented capabilities begin expanding, it suggests that firms are adjusting to a more volatile operating landscape rather than assuming conditions will normalize quickly. The concern I have is that some sectors are already operating that way since the start of the year.

The global economy has entered periods like this before, but the current environment appears increasingly defined by supply-driven volatility rather than demand cycles alone. Let’s also not forget the entire AI wars that are happening. When supply constraints become a persistent feature of the system, corporate behavior evolves in response. I wrote recently about US Vs. China in the AI Wars… Capital allocation changes, operating models shift, and workforce architecture adapts accordingly. Some countries are better poised to handle shocks, even during this current geopolitical tension.

Markets reprice geopolitical shocks within hours and the last seven days are proof, but the lasting repricing is the one embedded in routes, contracts, and headcount architecture that persists long after the first volatility spike fades. We are in this for the long term, even if it all halts in the coming week or two.

They are asking which model is better, which chip is faster, which company will beat quarterly expectations.

That is not how enterprise spend is routed.

Enterprise spend is routed through authority.

And if you are allocating capital across long-duration assets… across currencies, across liquidity regimes, across cycles… authority is the only layer that matters.

So I started somewhere different.

I wanted to know whether enterprise AI is being adopted as a growth catalyst inside business units, or whether it is being embedded inside restructuring mandates and operating-model enforcement.

If it is the latter, then the institutions that sit at the enforcement layer of the capital stack should be building capacity, and the layers dependent on labor expansion should not be in synchronized breakout mode.

That is testable.

So I mapped the entire routing system.

Private equity platforms: Blackstone, Apollo, KKR, Carlyle, Bain Capital, TPG, Warburg Pincus, Silver Lake, Vista, Thoma Bravo.

Strategy and transformation firms: McKinsey, Bain, BCG, Accenture, Capgemini.

Big Four advisory: Deloitte, PwC, KPMG, EY-Parthenon.

Restructuring and performance improvement specialists: Alvarez & Marsal, AlixPartners, FTI.

Capital advisory boutiques: Evercore, Lazard, PJT, Moelis, Houlihan Lokey.

Enterprise application vendors across ERP, HCM, CRM, procurement and workflow.

AI-native data platforms and infrastructure.

Global delivery and labor-arbitrage firms.

If AI were primarily a bottom-up innovation cycle, you would expect to see broad hiring acceleration across implementation firms, offshore delivery, and seat-based SaaS vendors. If instead AI is being embedded inside cost programs written by sponsors and restructuring firms, you would expect to see expansion concentrated in those layers.

The staffing data is not ambiguous.

Between February 2025 and February 2026 (YTD):

Blackstone increased internal headcount by +33.81%.

Apollo expanded +15.43%.

KKR +12.53%.

Carlyle +10.28%.

Bain Capital +8.35%.

Warburg Pincus +7.09%.

Thoma Bravo +7.18%.

Silver Lake +5.49%.

TPG +6.65%.

From a workforce composition standpoint, that magnitude of growth inside private equity is not cyclical enthusiasm. It reflects an institutional decision to intensify internal control. These hires are not junior deal teams chasing IPO windows… they are portfolio operations leads, technology integration officers, centralized procurement analysts, margin modeling specialists, data governance heads and cross-portfolio reporting teams.

When sponsors expand internal operating teams double digits during a constrained exit environment, they are not preparing for optionality… they are preparing for enforcement.

And for a multi-billion allocator, that matters.

More sponsor oversight means:

• Vendor stacks are reviewed centrally.

• Tech budgets are scrutinized centrally.

• Automation targets are standardized.

• Labor cost benchmarks are enforced.

That creates a capital-routing bottleneck…

Now move one layer down.

EY-Parthenon expanded +23.14%.

Alvarez & Marsal +21.17%.

Evercore +14.64%.

Moelis +10.83%.

PJT +10.70%.

Lazard +8.70%.

Houlihan Lokey +7.43%.

Deloitte +11.54%.

This is not broad consulting optimism at all… its just good ol’ fashioned throughput. These firms scale when cost programs move from slide decks to execution schedules, and from a workforce perspective, this represents expansion in:

• Operational analytics teams.

• Restructuring associates.

• Vendor rationalization specialists.

• Carve-out implementation leads.

• Process redesign engineers.

When this layer thickens simultaneously with sponsor operating teams, it means decision velocity increases. For allocators thinking in duration terms, that implies AI deployment is not happening through experimentation budgets. It is happening through mandate pipelines.

Now contrast that with McKinsey at −1.95%, Bain flat at −0.14%, and Accenture at −3.12%.

The growth is concentrated in enforcement-adjacent execution.

Now look at enterprise software.

ERP:

SAP +13.74%.

Oracle +1.71%.

NetSuite −5.54%.

SAP reinforcing core control systems makes sense in an environment where financial reporting and cost visibility matter more. Oracle holding steady is not an expansion story. NetSuite contracting tells you mid-market tightening is real.

HCM:

Workday +3.63%.

ADP +2.57%.

UKG +9.85%.

Dayforce +10.88%.

If companies were preparing for a hiring boom, these numbers would look very different. Instead, this looks like instrumentation… labor tracking, scheduling automation, payroll consolidation.

CRM:

Salesforce +7.75%.

Workflow:

ServiceNow +15.22%.

When processes are standardized and enforced, workflow becomes central.

Procurement:

Coupa +8.82%.

Spend management scaling fits cost discipline cycles.

Now underneath that layer:

OpenAI +74.37%.

Databricks +29.59%.

Snowflake +16.99%.

NVIDIA +12.24%.

The substrate is scaling.

Now look at global delivery.

Cognizant −2.26%.

Infosys +14.55%, but not part of a synchronized offshore explosion.

If this were an outsourcing-led transformation cycle like 2014–2019, offshore hiring would be accelerating aggressively. It isn’t.

Revenue streams tied to measurable cost reduction embedded in sponsor mandates behave differently in tightening liquidity environments than revenue streams tied to hiring growth and discretionary expansion.

When AI is embedded in restructuring decks, it becomes:

• Non-discretionary.

• Capital-aligned.

• Portfolio-standardized.

All of those increases durability for substrate layers.

It also increases concentration risk for peripheral SaaS, and compression risk for seat-based vendors dependent on headcount expansion.

Now, for a moment, consider vendor rationalization across sponsor portfolios.

If a sponsor controls 20 portfolio companies and eliminates two overlapping $1M SaaS vendors per company, that is $40M displaced within one platform.

Across ten major sponsors, that is $400M displaced, and that displacement does not hit the ERP control plane first. It hits peripheral tools, which increases dispersion, and from a capital allocation standpoint, dispersion creates asymmetry.

Infrastructure embedded in cost mandates retains support.

Peripheral expansion tools face consolidation pressure.

Now return to my thesis.

Consultants write the restructuring deck.

Sponsors increasingly control the consultants through expanded internal operating teams.

The workforce data shows sponsor capacity thickening materially.

That means consultants are operating within tighter capital frameworks. If OpenAI partnerships are embedded in that consultant layer, AI routing becomes structural. We have arrived to this point now.

This makes everything mandate driven.

Pretty thematic overall, no?

What we are seeing is a regime shift in how enterprise capital is deployed.

Authority has moved upstream, and the enforcement layer has thickened. We are also seeing that the expansion-dependent layers have not accelerated proportionally.

This is a total reconfiguration of the enterprise operating system, and capital always follows the layer that can enforce change.

Now… let’s push the capital implications further from a workforce POV.

If authority is consolidating at the sponsor and restructuring layer, and if AI deployment is being routed through mandate rather than experimentation, then duration risk across enterprise technology bifurcates.

That bifurcation is not visible in quarterly earnings, but it is visible in how revenue is justified internally. Let’s not forget that revenue justified as cost reduction behaves quite differently from revenue justified as growth enablement. In tighter liquidity, like, persistent high rates, currency volatility, boards will accelerate cost programs - favoring embedded infra (lower cut risk) over edge SaaS (higher discretionary review).

When a restructuring team embeds automation inside a portfolio-wide cost program, the internal approval process shifts. The conversation becomes: “What measurable labor savings does this generate?” rather than “Does this accelerate revenue?” That changes budget classification. Cost programs are rarely optional. Growth programs often are.

Over a five-year horizon, even modest labor compression compounds.

Assume 5 million enterprise employees across large sponsor-influenced enterprises. If AI-enabled automation reduces aggregate headcount by 1% per year over five years, that is roughly 50,000 seats per year, compounding. After five years, roughly 250,000 seats have been removed relative to baseline. At $220 per seat annually, that is $55 million in normalized HCM revenue displacement relative to trend. At 2% annual compression, the compounding effect doubles.

The point is not that HCM collapses.. it’s that marginal seat growth embedded in long-duration valuation models becomes structurally lower.

Now layer in vendor rationalization.

If each major sponsor portfolio eliminates two peripheral SaaS tools per asset over a three-year window, and average ACV is $1 million per tool, the math is straightforward. For a 20-company portfolio, that is $40 million displaced. For ten large sponsors executing similar rationalization programs, $400 million in recurring SaaS revenue is redistributed or eliminated. This does not mean all revenue disappears. Some consolidates into control-plane vendors. Some shifts into standardized workflow. But the edge contracts.

Duration-sensitive capital notices edge contraction before it shows up in broad sector multiples.

Now consider margin durability.

Infrastructure providers and AI-native platforms embedded in cost programs are typically justified by ROI thresholds. If a model deployment reduces back-office labor by $10 million annually, the infrastructure supporting that deployment is unlikely to be cut in a downturn. It becomes embedded in operating efficiency.

Peripheral SaaS tied to incremental hiring or discretionary departmental budgets does not have the same internal protection. During liquidity tightening, those budgets are reviewed first.

From a capital allocation standpoint — especially for allocators managing multi-currency reserves or sovereign capital where preservation and duration matter — this difference in internal budget classification becomes critical.

One layer of revenue is classified as operating necessity.

Another is classified as discretionary enablement.

Over long horizons, that difference widens dispersion.

Now consider currency stability and liquidity regimes.

In tighter global liquidity environments, enterprises prioritize free cash flow generation and margin stability. Sponsor-controlled portfolios intensify oversight. Cost programs accelerate. AI embedded in restructuring mandates becomes a tool for protecting margins against currency volatility and financing cost variability.

Infrastructure revenue tied to measurable efficiency thus carries lower sensitivity to discretionary budget contraction. Seat-based SaaS tied to workforce expansion carries higher sensitivity.

For allocators operating across AED, CHF, USD or other reserve exposures, the key question is not “Which AI company grows fastest next quarter?” It is “Which revenue streams remain justified when financing conditions tighten and boards demand cash flow stability?”

The staffing data suggests that enterprise AI is increasingly embedded in the latter category.

After looking at the workforce composition of these companies, we can also widen the lens further.

Private equity operating teams expanded materially during a period of constrained exits. That says to me this is the signal of value creation. Restructuring and capital advisory firms expanded materially, signaling throughput of cost and capital realignment programs. Workflow and procurement vendors expanded in alignment with process standardization and cost visibility. AI-native platforms scaled engineering and integration capacity. Offshore labor did not explode. Broad seat-based SaaS did not enter any kind of true euphoric expansion.

That alignment indicates a structural pivot.

The productivity lever is changing from labor multiplication to labor compression enabled by automation. And no, I am not talking the AI excuses given for layoffs… I’m just calling it for what it is… a capital allocation claim.

When sponsors standardize AI stacks across portfolios, vendor concentration increases, control plane systems gain stickiness, peripheral tools face review and revenue dispersion widens.

Now think about that five-year duration again.

If sponsor enforced automation reduces labor cost by 3–5% across large enterprises, and if vendor stacks consolidate by even 5–10% at the edges, growth normalization for seat-based SaaS becomes structural rather than cyclical. Infrastructure and workflow embedded in enforcement programs retain budget protection. That divergence compounds over time.

Return to my thesis.

Whoever controls the consultant controls enterprise AI spend.

• Consultants write restructuring decks.

• Sponsors increasingly control consultants through expanded internal operating teams.

• The workforce data shows sponsor capacity thickening materially.

And all of that means consultants operate inside tighter capital mandates.

When AI partnerships are embedded in that consultant layer, AI spend becomes embedded in capital structure decisions rather than discretionary experimentation.

The question is no longer “Is AI growing?”

The question becomes:

• Where is AI embedded inside non-discretionary capital enforcement?

• Which revenue streams sit closest to mandated cost reduction?

• Which revenue streams depend on hiring growth and departmental discretion?

• Which layers gain concentration as portfolios standardize?

• Which layers lose share as stacks compress?

When authority consolidates, spend consolidates.

When spend consolidates, dispersion increases.

The staffing footprint I’m seeing across the institutional stack is coherent.

Not headlines. Not WARN notices. Not quarterly beats. Not even product demos.

Authority.

It’s about authority.

And today…. Whoever controls the consultant controls enterprise AI spend.

Share

*Headcount figures derived from internal tracking, LinkedIn aggregates, and public disclosures (e.g., 10-Ks, earnings calls) - directional patterns hold even if exact % vary by division/role.

*Deloitte State of AI in the Enterprise 2026: Productivity/efficiency gains lead (66% of orgs report them), cost reduction at 40%, but revenue growth is mostly aspirational (only 20% achieved vs. 74% hoped for).

*PwC 2026 Global CEO Survey & AI Predictions - 56% of CEOs report no significant financial benefit yet (neither revenue nor cost gains); only 12% see both cost and revenue wins, often in scaled/embedded programs.

*McKinsey Global Private Markets Report 2026 - PE firms emphasize operational value creation as the primary return driver now (AI as force multiplier for underwriting, ops improvements, decision-making); only 6% see high internal impact today, but 70% expect it in 3–5 years.

In November 2025, two consecutive months of hiring data indicated a material shift in Palantir’s incremental workforce layering. Hiring intensity declined into the low-teens percentage of its prior expansion velocity, implying that incremental hiring had fallen approximately 85–90% relative to the earlier build phase. This does not appear to be an isolated monthly fluctuation. Over the broader trailing period, total hiring remained roughly 50% below its prior 12-month baseline, while net additions declined approximately 46% year-over-year. During that same time, the backfill ratio - hires divided by exits - compressed from approximately 1.79x to roughly 1.50x, representing a decline of approximately 16% in replacement intensity.

Engineering, which had previously represented roughly 53% of incremental hiring, moved modestly higher to approximately 56%, but within a significantly compressed hiring envelope. That shift tells us that the focus here was prioritization under constraint, rather than expansionary acceleration.

Operating margins expanded into the low-30% range during this same period, while revenue per employee increased. Historically speaking, these metrics are internally consistent with labor throttling and selective replacement rather than renewed incremental build.

Taken together, these workforce signals indicate reinvestment compression and margin defense layered on top of prior expansion cycles.

-They do not indicate collapse.

-They do not indicate contraction.

-They indicate optimization.

Workforce architecture suggests that, at present, it is not.

Hiring intensity reflects forward demand expectations and management’s willingness to commit incremental labor capital. Companies hire engineers, deployment teams, and product infrastructure ahead of revenue realization- I know most think it is after the snowball gets going… but it is not. When hiring intensity declines into the low-teens percentage of its prior velocity, the implication is a substantial reduction in forward capacity commitments. This is the pause.

An 85–90% decline in incremental hiring relative to prior expansion cannot be interpreted as routine moderation. It represents a deliberate recalibration of expected demand relative to existing capacity. Such recalibrations typically occur when management believes prior build cycles have created sufficient infrastructure to support projected revenue or when incremental return on new hires has moderated.

When I look over the broader period, hiring remained approximately 50% below trailing baselines. This sustained compression being seen in the trendline data confirms that November’s low-teens reading was not anomalous but rather the sharpest data point within an ongoing deceleration.

Hiring is anticipatory.

If incremental layering remains suppressed for multiple quarters, will forward growth acceleration become structurally constrained unless reacceleration occurs? Historical signs usually point to yes.

Palantir continues to add headcount in net terms. However, net additions have declined approximately 46% year-over-year. This indicates that productive capacity continues to expand, but at roughly half the prior pace. A mix of changes plus a potential change in hiring type- more specialists over generalists as well.

When net additions slow materially while revenue continues to grow, operating leverage expands.

Revenue growth exceeds labor growth.

Revenue per employee increases.

Operating margins improve.

This mechanical relationship is visible in Palantir’s financials. Margin expansion into the low-30%ish range coincides with slowed incremental labor layering.

Operating leverage derived from moderated hiring differs from leverage derived from accelerating scale. In the former, margins expand through cost discipline. In the latter, margins expand through capacity expansion outpacing fixed cost growth. The workforce data aligns with the former dynamic.

Key Workforce Data Summary

Hiring Intensity

Incremental hiring fell into the low-teens % of prior expansion velocity in November 2025 (85–90% reduction vs prior build phase).

Over the broader trailing period, total hiring remained approximately 50% below the prior 12-month baseline.

Net Additions

Net headcount growth slowed approximately 46% year-over-year.

Headcount remains positive, but expansion velocity has materially decelerated.

Engineering Mix

Engineering represented 53% of incremental hiring previously.

Engineering represents 56% more recently.

Increase occurred within a materially compressed total hiring base.

Backfill Ratio

Backfill ratio declined from 1.79x to 1.50x.

Approximately 16% compression in replacement intensity.

Revenue Per Employee

Increased during period of hiring compression.

Revenue growth exceeded labor growth.

Operating Margin

Expanded into the low-30% range during the same period.

Engineering’s share of hiring increased modestly from approximately 53% to roughly 56%. However, because total hiring is materially compressed, this shift tells us that we are seeing preservation rather than expansion.

A genuine expansion phase would involve both an increase in total hiring intensity and a disproportionate increase in engineering’s share of that expanding base. Instead, engineering represents a stable majority within a constrained hiring envelope.

This configuration suggests that existing product infrastructure is being maintained rather than aggressively extended. For a company whose valuation partially reflects expectations of ongoing AI-driven platform expansion, incremental engineering layering is a leading indicator of future monetization surface area.

At present, that layering remains constrained.

Backfill compression?

The decline in backfill ratio from approximately 1.79x to roughly 1.50x represents a material reduction in replacement intensity. Attrition is being tolerated at higher rates relative to replacement.

Backfill compression is one of the most direct mechanisms through which operating margins can expand without explicit restructuring. It signals capital discipline and selective replacement rather than aggressive layering.

The alignment between backfill compression and margin expansion confirms that labor throttling is contributing to improved profitability metrics.

Revenue per employee increased during the period of hiring compression as well. This is consistent with revenue growth exceeding labor growth. However, revenue per employee improvements driven by reinvestment moderation differ from improvements driven by structural productivity gains.

If revenue per employee gains reflect structural AI leverage, duration extends. If they reflect throttled labor growth, gains plateau once labor compression stabilizes.

Without renewed incremental layering, productivity improvements alone cannot indefinitely sustain elevated growth duration.

Consider a simplified duration framework:

If a valuation assumes eight years of elevated revenue growth before normalization, and reinvestment compression effectively reduces that window to six years before normalization begins, intrinsic value declines materially even if margins remain elevated.

Duration compression reduces the compounding base upon which terminal value rests. Even modest reductions in high-growth years produce nonlinear valuation effects.

Hiring intensity compressed into low-teens percentages and sustained below-baseline levels suggests reinvestment velocity has paused. If that pause persists, effective growth duration shortens.

What would change my view from a workforce POV?

This interpretation would be invalidated by sustained evidence of reinvestment reacceleration. Specifically, total hiring intensity would need to rebound materially toward prior expansion baselines rather than remain compressed. Engineering would need to expand as a dominant share of a growing hiring base, not merely represent a preserved majority within a constrained envelope. Net additions would need to reaccelerate meaningfully. Backfill intensity would need to rise back toward prior ~1.8x replacement levels, signaling aggressive layering in anticipation of demand expansion.

A true expansion phase would look like….

In a genuine acceleration phase, total hiring intensity would inflect upward materially. Engineering would represent a dominant and expanding share of that reaccelerating base. Net additions would accelerate, not merely remain positive. Backfill intensity would increase rather than compress. Incremental capacity layering would lead revenue growth rather than revenue being harvested from prior layering.

Current workforce data does not reflect that configuration.

Palantir’s workforce architecture reflects reinvestment compression and margin defense rather than aggressive incremental expansion.

This configuration is consistent with optimization layered on top of prior build cycles.

The current valuation framework implies sustained reinvestment velocity. The workforce data currently reflects optimization. Those two states cannot persist indefinitely without reconciliation.

Growth does not need to decline for valuation to adjust. Duration compression alone is sufficient.

Workforce architecture remains the leading indicator.

Amanda Goodall, The Job Chick

amanda@thejobchick

Almost no one analyzes workforce architecture as a capital-cycle signal… but Amanda Goodall does.

If the largest private employer in the country can grow without adding workers, the benchmark for a “healthy” labor market changes.

Walmart told investors something far more consequential in 2025 than probably anything they could say in their filings… it no longer expects labor to scale with growth.

“When we look out two years, three years, five years… we’ll have roughly about the same number of people we have today and we’ll have a larger business,” Walmart U.S. CEO John Furner said.

That statement should have landed harder than it did.

Walmart is the largest private employer in the United States. For decades, its operating model was linear meaning more stores, more volume, more people. That model has now been explicitly retired. Jaw drop… I know….

Revenue is rising, and so to is digital penetration, automation spend… yet headcount is not and it is all about demand.

Over the past two decades, Walmart’s workforce composition has shifted away from human judgment including management, supervision, and project coordination… and has focused heavily on execution increasingly mediated by software.

Walmart’s EVP of AI Transformation explained this in plain language too. I couldn’t take notes fast enough as he spoke.

Read literally…. not optimistically… because this is one of the clearest labor signals a Fortune 1 company has ever given.

I wanted to focus this as a piece about labor relevance.

To understand what Walmart is doing, you have to stop listening for buzzwords and start listening for control logic and their ideas of risk execution here.

Daniel Danker, Walmart’s Executive Vice President of AI Transformation, Product, and Design, was unusually explicit about the future operating model. Asked what a Walmart store looks like ten years from now, he did not talk about fewer people. He talked about where decisions live.

“The associates that are there all have a device that has an AI agent that powers their entire day.”

WOW.

“This agent is doing everything from helping them help customers… to telling them what the best next thing you can do is.”

That sentence is the story. Seriously.

Historically, deciding the best next thing to do was the job. That is why retail required layers of department managers, assistant managers, shift leads, floor supervisors. Retail is volatile. Shelves empty unpredictably…. well they used to anyways… Someone had to continuously reprioritize labor.

Walmart has now embedded that function into software.

Danker gave a concrete example:

“If there’s a spill on Aisle Five… it will literally kick in and say, ‘Actually, I need to reroute you.’”

Talk about command and control logic.

This is how labor disappears without layoffs. Welcome 2026.

Walmart does not need to fire supervisors en masse. It simply stops replacing them. Attrition does the work. Over time, roles collapse into flatter execution bands. Judgment migrates upstream into centralized systems.

The workforce data I’m looking at already confirms this pattern and they haven’t really kicked this into gear yet.

Looking at Walmart job flows since 2003, growth in business management and program / project management roles is negligible relative to total employment. These categories account for well under one percent of net job creation over more than two decades. Operations roles grow, but slowly… and increasingly as a percentage of a largely static base.

This is not an accident.

Most automation narratives focus on frontline displacement - but that is a distraction and probably one they hoped for.

Walmart is not starting by eliminating cashiers or shelf stockers. It is eliminating the reason those workers ever needed managers.

Store level middle layers exist to resolve conflicts between tasks and they allocate labor under uncertainty.

When Danker describes Walmart’s associate agent, he is describing the removal of that function.

The agent does not just answer questions about inventory… it monitors depletion rates, knows inbound truck schedules and it evaluates urgency. It reprioritizes work mid-task.

Those were management decisions but once encoded in software, they do not come back.

This is why Walmart can cap global headcount around 2.1 million and still expect to grow. Middle layers are cost-dense and judgment-heavy. Removing them produces immediate operating leverage without touching frontline execution.

The effect is subtle but permanent.

Execution-only labor does not scale with revenue. It scales with square footage. Once stores are optimized, incremental growth flows through throughput, automation, and software… not people.

Walmart’s leadership is saying this plainly. They are not promising job growth because they no longer need it.

This matters far beyond Walmart.

The result is not mass unemployment overnight. It is chronic under-absorption. Fewer ladders and more people competing for execution roles with less internal mobility.

That is the workforce signal embedded in Walmart’s strategy.

At roughly $130 per share, Walmart trades near its highs with a valuation multiple that increasingly reflects technology-like expectations: sustained margin improvement, operating leverage, and efficiency gains driven by automation rather than labor expansion.

The Q3 2025 bull case was straightforward: store-level automation replaces incremental hires, efficiency metrics improve as the labor mix tightens, and Walmart earns a higher multiple for discipline.

That case largely played out.

The question now is whether the market is pricing only the upside.

From a workforce perspective, Walmart’s strategy supports margins in the medium term. Fewer middle layers reduce fixed labor costs, centralized decision logic reduces variance and execution labor becomes more interchangeable.

But we must remember that thinner labor systems are also more brittle.

When judgment is centralized, local adaptability declines and errors can propagate faster. Training and churn costs rise at the execution layer. Political and regulatory scrutiny intensifies as absorption capacity falls. I’m not seeing this for now… but that is now.

A pullback into the low-100s would not contradict the thesis. It would reflect a bit of digestion…. or indigestion depending on what side you sat on…. with investors reassessing how much efficiency is already priced in versus how much execution risk remains.

Walmart leadership does not say this explicitly. But the language they are all using is pretty darn telling.

Danker described “four super agents” coordinating “too many nano agents to count.” He framed the future in terms of architectures, orchestration, and control layers…. not people.

That is how technology companies talk.

Walmart is not trying to be Silicon Valley of course, but it is doing something more consequential that we have not seen before in such a meaningful way… they are absorbing technology into the core of labor allocation.

Will millions of people still work for Walmart in ten years?

The wrong question is whether jobs exist.

The right question is how many humans are still required once judgment, prioritization, and exception handling are no longer human bottlenecks.

Walmart is answering that question now. Explicitly… and pretty much ahead of almost everyone else.

Is labor still going to be relevant?

Once the largest employer in the United States proves it works, it will not remain unique to Walmart.

Amanda Goodall

amanda@thejobchick.com

Get more workforce intelligence and risk execution reports.

You know my POV on what is really happening in the workforce. If you are new, don’t worry, you will quickly get up to speed. If something is genuinely transformative, you don’t need to shout about it every earnings call. You don’t need to keep telling people it’s inevitable. It shows up. And structurally, it’s how work actually gets done. It shows up in role clarity, in whether teams stabilize, in whether backfills resume, in whether people feel like they’re building something durable or just bracing.

And right now, when I look inside companies… not at stock prices, not at narratives, not at AI demos… AI does not look like clean productivity. It looks like pressure. It looks like organizations trying to run hotter with fewer people and calling that progress because a model says it should work.

That’s the part that gets me.

What I’m seeing is not companies suddenly operating with more slack or ease. I’m seeing companies compressing the people inside them and then laundering that compression through technical language until it sounds like destiny instead of a choice. And when you start there, the workforce outcomes are predictable… even if the market outcomes lag.

Let me be very explicit about what I actually track, because this matters. I don’t look at headline headcount numbers. Those are blunt, lagging, and easily managed. I look at job architecture. Which roles disappear. Which ones get merged. How spans of control widen. Whether attrition gets backfilled or absorbed. Where contractors appear as pressure valves - because DAMN y’all - it’s such a tell about company health. One entire industry is specifically consistent in their patterns right now, and historically we didn’t see this before.

Across tech right now, the pattern is consistent.

Mid-layer execution roles are really starting to thin out across the board. Managers are seeing twice as many direct reports, and the roles that used to be a bit more distinct are fast becoming hybrid positions with broader scope and less support. Don’t get me started on the ChatGPT style vague job descriptions either… and what I am hearing in the rumor mill is that accountability is harder and harder to trace, which ironically doesn’t help any employee.

Seriously… when productivity is real, you don’t need heroics, you don’t need people doing three jobs. So what I am seeing in the workforce data is consolidation justified by tooling rather than proven output. It’s a weird world we live in right now. AI is being layered on top of work that was already close to capacity, and leadership is assuming the technology will close the gap before the strain shows up.

That assumption is doing an enormous amount of work inside these organizations.

This is where narrative laundering really starts to matter. Narrative laundering is when uncertainty gets cleaned up by a model so no one has to own the downside. It’s not about lying. It’s about taking something operationally messy and reframing it as technically inevitable.

“The model says we have to.”

“This is where the industry is going.”

“We don’t have a choice.”

Those phrases aren’t neutral. You’ve heard CEOs say things just like that recently right? It feels like the new 'cool kids phrase.’ Once these are accepted, responsibility is dunz-o.

AI is being sold as an option-expanding technology. But the way it’s being implemented is option-constricting.

There are some CEOs that shall remain nameless that come across not quite as a villain, but as a signal amplifier. When they goes on stage and talk about the future of compute, markets hear destiny… boards hear inevitability… executives hear cover. Their messaging doesn’t cause layoffs. What it does is synchronize assumptions across companies without coordination. It sets the external clock. It gives internal decision-makers permission to align timing. But what I hear- lies, lies, and more damned lies.

Read more

Snowflake is in one of those moments.

This is not about revenue collapse at all.. frankly it is not even about failure either. It is a story about a company that has reached the phase where growth narratives collide with margin reality, and where the word “innovation”… which was once an asset - just becomes a liability when it no longer translates into operating leverage.

What follows is not a prediction based on rumors. It is an interpretation based on structure.

• Workforce behavior.

• Cost signaling.

• Leadership actions.

If you are looking for a simple claim - “Snowflake will lay people off in February” - you are missing the point. The more important story is why February makes sense, who would be affected first, and how this kind of adjustment actually happens in modern companies that care deeply about optics.

Read more

What’s actually reshaping power isn’t trade flows or GDP rankings. It’s the restructuring of labor, those who keep skilled people in place, those who let institutional knowledge decay, and those who still know how to operate complex systems at scale.

Once you look at the world through a workforce lens, a country that almost nobody is modeling correctly comes sharply into focus - South Korea.

I have no intent on creating a nationalist argument, because this isn’t the growth story most think this argument is. It isn’t even a hype trade. It’s just about execution, and in this current cycle we are in, it is everything. Recently the Korean market has also become unusually volatile as global institutions rotate through it looking for momentum trades. That may look like speculation, but it also reveals something a bit deeper overall. Markets tend to rediscover execution economies late. The factories, engineers, and industrial supply chains usually matter long before capital flows recognize them.

The ones that will win the behind the scenes headlines in 2026 are the ones that have positioned themselves so the system cannot function without them. It’s getting mighty interesting and we are still in January.

Decoupling is on the lips of many economists lately…Trade data and tariff headlines totally take the spotlight. I’m going to be brash enough to say… most are missing the fact that this decoupling is happening inside the labor systems.

In a nutshell, the United States has reorganized its workforce around abstraction. Software, platforms, finance, defense integration, legal frameworks, and services dominate labor allocation. Churn has been normalized and institutional knowledge is thin by design. I’m not a fan of that.

China, by contrast, has doubled down on physical execution. Infrastructure, manufacturing, logistics, energy, and automation are hiring and training labor at scale. Overcapacity is a feature, not a bug. Way different than what we see in US companies right now.

Now, of course, once training pipelines, promotion incentives, and career paths diverge this far, reintegration becomes extremely difficult. Workers don’t simply move back across systems because trade rules soften. Skills atrophy is real, and this is why the current state of “selective decoupling” is the structural labor realignment we are starting to see break out.

And this is where South Korea becomes impossible to ignore.

South Korea’s workforce model is not what people think as it is often mischaracterized as a consumer-tech economy or a lifestyle-driven export state. This is totally missing the point.

Korea’s economic model has been built around workforce continuity. Engineers remain close to production, operators are retrained instead of discarded (ring any bells), and technical labor is not treated as a variable cost. Overall what we are seeing is policy-aligned, corporate-enforced, and culturally reinforced.

Unlike the US, Korea did not hollow out its industrial labor base during globalization. Unlike China, it did not absorb scale at the expense of quality control or international trust. Instead, it preserved a dense, long-tenure technical workforce capable of executing complex systems repeatedly, reliably, and on schedule. It’s impressive.

If my thesis was wrong, the labor flows would contradict it. They don’t.

SK Hynix is the clearest example. It is not a foundry and it does not design AI models. It manufactures memory… specifically high-bandwidth memory that has become the binding constraint in advanced AI systems.

What matters is not the product, but the people.

SK Hynix has preserved long-tenure process engineers, yield specialists, packaging experts, and manufacturing operators through multiple down cycles. As US semiconductor firms retrenched, paused fab investments, and shed institutional knowledge, Hynix absorbed displaced senior talent from Intel, Micron, Western Digital, and Solidigm.

Markets have begun to recognize this. The repricing of SK Hynix is not a speculative AI trade, and the market is assigning value to workforce reliability at a bottleneck layer that cannot be substituted.

Samsung Electronics follows the same logic.

• Engineers remain tied to fabs.

• R&D stays integrated with production.

• Workforce planning aligns with decade-long capital cycles, not quarterly margin pressure.

KEPCO and KHNP tell an even more underappreciated story. Nuclear engineers and grid operators are retained despite pricing constraints and political pressure, not to mention, that the apprenticeship pipelines remain intact. Export-capable EPC teams are preserved even when domestic returns are limited. You cannot surge nuclear talent. You either have it or you don’t. This has been very interesting to watch, and quite possibly the only example I have ever seen in the workforce.

Hyundai Heavy Industries and Hanwha extend the pattern into shipbuilding, defense, and energy. Skilled operators and engineers are cross-trained and pooled, and the workforce continuity allows for a quick pivot between commercial and defense production without rebuilding teams from scratch. Nothing flashy, just load bearing, and impressive.

So why is the US model struggling to translate power?

The United States does not lack talent. It lacks labor alignment.

Its workforce is optimized for invention, mobility, and individual upside. That produces extraordinary innovation. It also produces chronic execution gaps. We have totally lost the plot we once had. Once you study the workforce and challenges, and planning that has happened over the past 130+ years, it is evident that we have fallen afoul of what can keep driving us forward. Some say it is greed to hit the highest market cap, as we see so many companies push forward on… but regardless of what it is, it is holding the US back.

Here’s the thing… fabs can be funded, and we know data centers can be financed… but yield intuition, maintenance discipline, and process stability cannot be accelerated with capital alone. The loss of manufacturing muscle memory is now visible across sectors from semiconductors, grid infrastructure, utilities, defense production timelines. Skilled operators are aging out and those hefty apprenticeship pipelines they maintain in South Korea are becoming seriously thin in the US. Engineering is increasingly detached from production.

This is not a three-year policy problem.

It is a generational workforce problem.

On the other hand, China’s workforce remains unmatched in scale and physical deployment. It can build infrastructure, energy systems, and manufacturing capacity at speeds no other country can match. But that same workforce structure creates a different constraint which is value capture.

China’s GDP growth reflects capacity creation. Its equity markets reflect allocation decisions and the labor-generated surplus is often reinvested, redirected, or absorbed by state priorities rather than distributed to shareholders. Overall we are seeing a structural separation between work performed and returns allocated.

Japan, for instance, stagnated because its workforce aged into maintenance. China’s workforce is still building. Investors simply do not sit at the top of the distribution chain.

Then, we have the debate over AI dominance and if it remains fixated on chips and export controls. This misses the operational reality. Compute does not scale because silicon exists. It scales when people can operate the systems around it… fabs, advanced packaging lines, power infrastructure, data centers, and maintenance cycles that tolerate no churn.

The constraint sits with long-tenure operators like the technicians, yield engineers, and systems specialists who keep uptime high and failure rates low.

South Korea has those people. And it protected them when others did not. In fact, South Korea did not try to dominate every layer of the stack. It positioned itself at the layers no one can remove.

Memory.

Energy execution.

Grid reliability.

Advanced manufacturing.

Defense-adjacent production.

This is all about indispensability.

Korea’s workforce model resembles China’s in discipline and execution. It resembles the US in quality control, standards, and trust. That combination is rare… and increasingly valuable as systems strain.

This is why Korea can supply the US without subordinating itself. and also why it can trade with China without collapsing into dependency. It operates inside both systems while being owned by neither.

Much has been made of allies “hedging” between Washington and Beijing. This really is a miss read on labor. Japan, Korea, and the EU hedge because their workforces are embedded in both systems. Security alignment and economic alignment diverge because labor dependencies diverge. The breaking point is actually domestic though… not a diplomatic situation despite how it looks.

When policy forces factories to close, prices to spike, or skilled workers to become idle, hedging ends. Until then, it is straight up rational.

The common thread across sectors is clear. Companies retaining workforce continuity are being quietly repriced. Those that broke their labor pipelines are being discounted regardless of narrative.

Capital will flow toward entities that demonstrate they can still run complex systems without failure while layoffs will continue in abstracted middle layers. Long-tenure operational roles will become scarce and valuable.

Yield engineers, power systems operators, grid technicians, advanced packaging specialists — these are the roles that matter.

There are investor complications in capital, labor, and it extends through the next decade.

This shift has direct, underpriced implications for capital allocation, labor markets, and sector-level outcomes.

First, capital will continue to migrate away from headline innovation and toward execution certainty. Firms that demonstrate workforce continuity… especially in manufacturing, energy, and infrastructure-adjacent sectors… will command valuation premiums regardless of macro noise. This favors companies embedded in bottleneck layers like memory, power systems, grid hardware, advanced packaging, shipbuilding, and defense-adjacent manufacturing.

Second, labor markets will bifurcate. We are already seeing this. Abstracted middle layers that include program management, duplicated strategy roles, non-operator corporate functions will remain vulnerable to layoffs and automation. Again, does this sound familiar on what we are seeing in the US? In contrast, long-tenure operational roles will tighten further (if that is even possible). Yield engineers, power systems operators, grid technicians, nuclear specialists, advanced manufacturing technicians, and maintenance-heavy industrial roles will see sustained demand and rising bargaining power. This is a big part of the 2026 outlook that was sent out to my institutional clients in December 2025.

Third, reindustrialization narratives that ignore workforce continuity will disappoint. Capital subsidies cannot replace lost institutional knowledge, and countries and firms that broke their labor pipelines will struggle to translate spending into output, creating execution risk that markets are only beginning to price.

Finally, geographic exposure matters less than workforce architecture. Investors focusing solely on national champions will miss the point. Weirdly, I see this often. Honestly, the advantage will go to entities that made themselves non-optional by preserving execution talent through cycles.

This is really not a call to abandon innovation or scale. It is a recognition that the next decade will reward those who kept skilled people in place long enough to make complex systems run. There is a lot to be done.

This is not a race between flags, but because of the society we live in today, it is a competition over whether complex systems can still be operated by people who know how not to break them.

South Korea is not winning headlines. It is winning positioning.

But overall, the more important question is who cannot be removed… that is the only advantage that lasts. Most debates are still asking who will win, in fact I was recently on an X space hosted by Gamma Prime that was focused on China vs AI who wins the AI fight… while I have my opinions, I did stay data neutral in this, possibly irritating those that like a good argument.

Fact is, the way it all stands, various countries have various strengths. South Korea’s role exposes the fact that global power no longer comes from dominance, but from preservation. Different countries are winning at different layers… American innovation, Chinese scale, Korean execution … and the system only works if those layers stay intact.

The mistake many corporations are making, especially in the US, is treating labor as a cost to arbitrage rather than a capability to protect. At the current pace, offshoring and outsourcing are not efficiency moves- not in the slightest… but they are definitely degrading the domestic workforce faster than it can be rebuilt. Rising unemployment is taking the US and permanently detaching it from production. Korea isn’t winning by doing more, but it is winning by not breaking what it already had. And in the next decade, that may be the only advantage that actually compounds.

If you want to understand what’s actually happening inside companies before it shows up in headlines, earnings calls, or “unexpected” announcements, I recommend the Insider Edge Report.

This is where I publish my workforce intelligence in real time, analysis built from internal job-flow data, attrition patterns, and operating behavior that quietly reveal structural stress, silent contraction, and strategic shifts well before they’re publicly acknowledged.

I analyze markets the way serious analysts always have: by tracking repeatable patterns across cycles, grounding today’s signals in historical precedent, and discarding narratives that can’t survive contact with data.

If you want a front-row seat to what companies are actually doing… not what they’re saying… Insider Edge Report is where I publish it first.

Recommend Insider Edge Report

Emerging tech policy in the U.S. is not neutral. It is directional. It is written to optimize systems for scale, cost compression, and centralized control, while treating labor as an adaptive variable rather than a design constraint. This is not because policymakers are hostile to workers. Workforce continuity is just not part of the optimization function… and it should be.

Broadband expansion, artificial intelligence deployment, biotech security, and autonomous vehicles were discussed as separate policy domains. In practice, they are components of a single operating model. That model reduces human discretion, standardizes decision-making, and replaces judgment with systems wherever possible. Once that model is adopted, the employment consequences are not accidental. They are foreseeable. I know… It’s what I watch on the daily.

The reason this has not triggered widespread alarm is that the disruption does not look like disruption. There are no dramatic employment shocks attached to Senate votes. Instead, roles just disappear from future planning. Hiring freezes replace terminations. Attrition replaces headlines. And all the while they say growth is happening. Hmmmm…..

If you are a worker - especially a white-collar worker - and you believe emerging tech policy is being written with your job in mind, you are misunderstanding how institutional incentives operate.

Once these systems are deployed, the impact does not land evenly across the workforce. It concentrates in specific sectors, in specific functions, and in specific job families whose purpose is to mediate human judgment inside large organizations.

To understand what emerging tech policy actually does to employment, you have to stop thinking in terms of industries or headlines and start looking at roles.

That is where the pressure shows up first, and that is where the Senate’s design choices translate most directly into job elimination.

The digital divide discussion centered on affordability and fairness.

Millions of Americans cannot afford reliable internet access, and universal broadband is presented as a prerequisite for participation in the modern economy.

That framing is morally compelling, but it doesn’t focus on the labor implications.

Broadband on its own isn’t a threat to workers, and for many people it feels like a clear win.

The problem shows up in who benefits first and most once work and services move online. When employers no longer have to hire people in the same place the work is done, the people most affected are not executives or highly specialized workers. It’s administrative staff, coordinators, reviewers, support roles, and entry-level professionals whose jobs exist because someone needs to be physically present or locally embedded.

Once those jobs can be done from anywhere, employers gain options that workers don’t. They can automate the role, consolidate it into a smaller team, or hire from a cheaper labor market without changing the service itself. For workers in those roles, broadband doesn’t create new opportunity so much as it increases competition and makes replacement easier. That’s why this shift tends to hit middle-income, white-collar jobs first… not because the work disappears overnight, but because fewer people are needed to do it, and those positions stop being refilled when someone leaves.

In that sense, broadband isn’t harmful by itself. What makes it risky for the workforce is that it’s being rolled out alongside automation and cost-cutting incentives, without any policy requirement to protect jobs, wages, or clear transition paths for the people whose roles quietly become unnecessary.

In many cases, lack of broadband forces companies to hire locally.

Once broadband arrives, that constraint disappears, and local hiring is no longer necessary for the same work.

That doesn’t mean all local jobs disappear. It means the reason those jobs had to be local disappears. That’s the shift.

In areas without strong broadband:

• Work has to be done on-site or nearby

• Employers hire locally because they have no alternative

• Administrative, support, coordination, and service roles exist because of location constraints

Once broadband is available:

• The same work can be done remotely

• Employers can centralize the role elsewhere, automate it, or offshore it

• The job no longer needs to exist in that location… or at that headcount

So broadband doesn’t destroy the work… it just removes the requirement that the worker be local.

That’s the key distinction.

I’m not saying we shouldn’t roll out broadband- I am simply saying watch these areas for disruption and also for key leaders in these areas, you need to be putting efforts in place to maintain the workforce in that area.

Moving to the next subject… one of the most consequential moments on the panel were not about bias or ethics. They were about procurement. The federal government is the largest purchaser of technology in the world. When it adopts AI systems for benefits eligibility, fraud detection, compliance review, logistics coordination, or case management, it is not simply modernizing operations. It is actually just redefining which human roles are considered necessary.

Once an AI system performs a function inside a federal agency, human decision layers shrink or disappear. We’ve seen this already. Workforce reduction becomes reputationally safe. And this right here is how AI becomes labor policy without a single vote on employment. Bias mitigation does not alter this trajectory. A fair system still replaces a human. Transparency does not preserve jobs… it just removes objections to eliminating them.

Then we have the Biosecure Act… which was framed as a national security measure designed to prevent foreign misuse of U.S. genomic data. Securing genomic data domestically concentrates ownership and control. When paired with AI-driven analysis, it eliminates entire categories of human interpretation, validation, and administrative oversight.

Research cycles compress.

Review layers thin.

Output per worker increases dramatically.

Autonomous vehicles were discussed primarily through the lens of accident reduction. That framing avoids the real impact. Driving is not a standalone job. It supports dispatch, supervision, compliance, insurance adjudication, and incident review.

Human drivers require human systems.

Autonomous systems do not.

Once vehicles operate without human discretion, the supporting labor stack becomes unnecessary all at once. Regulatory requirements such as steering wheels and brake pedals enforce the presence of a human operator. Removing them removes the justification for every role built around that operator.

This is not speculative.

Everything above describes the architecture. What follows explains how that architecture behaves once deployed, where the pressure lands first inside the workforce, and why the outcome is unavoidable even if no policymaker intends harm.

Workforce disruption does not occur evenly. Systems remove functions, not job titles, and those functions cluster predictably.

In government, the first pressure point is decision mediation. Benefits eligibility reviewers, fraud analysts, compliance officers, caseworkers, and administrative program managers exist to interpret rules, apply judgment, and ensure consistency across cases.

AI systems adopted through procurement are designed to eliminate variability in those decisions. Once implemented, fewer cases reach humans.

Hiring stops.

Backfills are denied.

Oversight becomes exception-based.

Over time, the role itself becomes unnecessary without a single layoff announcement.

Healthcare follows the same structural logic. Intake coordinators, medical billing and coding staff, utilization review nurses, scheduling teams, and clinical documentation specialists exist to translate between patients, providers, and payers. AI now handles intake, prioritization, documentation normalization, and claims processing.

Human oversight remains legally required, but only at the margins. As systems improve, exception volume shrinks and headcount compresses. Employment appears stable while the nature and availability of roles deteriorate.

In logistics and transportation, the transition is more abrupt. Drivers are the visible role, but dispatchers, fleet supervisors, insurance adjudicators, and compliance coordinators all exist because human drivers introduce variability. Autonomous systems remove that variability entirely. Once autonomy scales, these roles lose their function simultaneously rather than gradually.

In technology and corporate operations, the pressure targets coordination itself. QA analysts, product operations roles, customer support tiers, business analysts, and middle-layer managers exist to monitor systems, translate information, and coordinate across functions. AI absorbs monitoring, testing, and reporting. Remaining roles absorb scope until their function is automated or outsourced. Entire layers disappear without any single role being “cut.”

Role-by-Role Exposure Map

Government

Highest exposure:

• Benefits eligibility reviewers

• Fraud analysts

• Compliance officers

• Caseworkers

• Administrative program managers

AI systems adopted through federal procurement replace decision and review layers. Once implemented, these roles are no longer backfilled. Hiring freezes replace layoffs, and human involvement is reduced to exception handling. Over time, the role itself ceases to exist without ever being formally eliminated.

Healthcare

Highest exposure:

• Intake coordinators

• Medical billing and coding staff

• Utilization review nurses

• Scheduling and triage teams

• Clinical documentation specialists

AI handles intake, prioritization, documentation normalization, and claims processing. Human oversight remains legally required, but it shrinks to margin cases. As systems improve, exception volume declines and headcount compresses without triggering visible layoffs.

Logistics & Transportation

Highest exposure:

• Drivers

• Dispatchers

• Fleet supervisors

• Insurance adjudicators

• Compliance coordinators

Autonomous systems remove the need for human operation. Once variability introduced by humans is eliminated, the supporting labor stack collapses. These roles do not disappear gradually…. but rather, they lose their function simultaneously.

Technology & Corporate Operations

Highest exposure:

• QA analysts

• Product operations roles

• Customer support tiers

• Business analysts

• Middle-layer managers

AI absorbs monitoring, testing, reporting, and coordination functions. Remaining roles absorb additional scope until the function they perform is automated or outsourced. Entire layers disappear - again, without a single role being publicly “cut.”

The Actual Employment Outcome

This policy framework does not produce mass layoffs. It produces:

• Permanent hiring freezes

• Disappearing entry-level pathways

• Middle-layer role extinction

• Credential inflation without wage growth

• Contractor substitution for employees

• Geographic wage arbitrage enabled by remote oversight

Because these changes occur through system design rather than announcements, they avoid political backlash. The labor market does not collapse. It thins.

Emerging tech policy is written to optimize:

• Error reduction

• Cost compression

• Speed

• Centralized control

• Standardization

It does not include:

• Employment preservation metrics

• Role transition mandates

• Wage impact analysis

• Skill displacement timelines

• Labor absorption planning

By excluding these elements, policymakers preserve plausible deniability. Disruption becomes market-driven rather than policy-driven. That framing is incorrect. When systems are designed to replace human judgment, labor outcomes are not accidental. They are foreseeable.

Fact is… Even Congress knew of the issues the workforce was facing back in 2003. This is no different.

6.6+ MILLION AMERICAN JOBS are now gone from the US. Just since 2003.
Congress knew.
This is from a 2003 Congressional Hearing.
They were warned millions of jobs would be offshored by 2015.
It’s now 2025. Ten years later.
The math checks out:
6.6 million jobs. $280 BILLION in lost wages.
STOLEN from American workers by corporations chasing cheap labor.
Safe to say those numbers… at least doubled.
Wanna bet they’re even higher?

Emerging tech policy is being written as if labor will adapt on its own. History suggests otherwise.

The Senate is building systems that work exactly as intended. Workers are expected to adjust to consequences that were never debated. That is not innovative policy.

If you want to understand what’s actually happening inside companies before it shows up in headlines, earnings calls, or “unexpected” announcements, I recommend the Insider Edge Report.

This is where I publish my workforce intelligence in real time, analysis built from internal job-flow data, attrition patterns, and operating behavior that quietly reveal structural stress, silent contraction, and strategic shifts well before they’re publicly acknowledged.

I analyze markets the way serious analysts always have: by tracking repeatable patterns across cycles, grounding today’s signals in historical precedent, and discarding narratives that can’t survive contact with data.

If you want a front-row seat to what companies are actually doing… not what they’re saying… Insider Edge Report is where I publish it first.

Recommend Insider Edge Report

Adam Goldstein missed a CES panel he was scheduled to speak on. The interviewer said he dropped out… but I see it as he avoided a room that would have forced Archer to be priced like every other eVTOL story… timelines, autonomy wish-casting, and certification “roadmaps.”

That panel instead was led by the people who actually live in regulated flight reality including Amazon Prime Air, describing live operations across multiple U.S. metros - Wind River walking through the safety certification stack used in aircraft systems - and a technologist describing how regulators treat “AI” when a human life is on the line. Goldstein’s absence changed the panel’s format, opened up live Q&A, and put constraint front-and-center.

That same day, Archer emphasized a different anchor… the NVIDIA partnership announcement was framed as aviation AI, backed by IGX Thor, safety-capable compute, and a clean story for investors who want a single sentence they can repeat.

If you want my contrarian take: the NVIDIA announcement is less about what Archer can do, and more about how Archer intends to be evaluated. The market is pricing advanced air mobility as a technology race. It’s still clever.

Now I’m gonna unpack the part investors keep missing… like the workforce composition, certification constraints that lock labor demand into short bursts, and why a hiring surge can be a late-stage signal that precedes role consolidation.

If you’re buying eVTOL on “autonomy” and “AI,” you’re modeling the wrong variable.

I like this phrase- Both Archer and Joby are hiring aggressively. That fact alone is meaningless.

Read more

What is actually happening inside the automotive and mobility sector is not exactly a consumer upgrade cycle. It is basically reading like a balance sheet- driven reallocation of capital and labor away from physical iteration and toward software systems that compound cost over time. Per the experts- autonomy, SDVs, and AI copilots… these are not endpoints, but rather you need to think of them as mechanisms.

Capital is no longer chasing moonshots for their own sake.

Conversations at CES made this totally clear, even if unintentionally. Capital is flowing toward architectures that shorten development cycles, reduce recall exposure, and convert one-time vehicle sales into long-lived software platforms. This change matters far more for workforce design than it does for marketing decks.

The defining characteristic of the SDV (software-defined vehicle) is not over-the-air updates or flashy interfaces. It is ownership of the stack.

The interesting game play here is that OEMs that control their software architectures can change behavior without touching the factory line, correct defects without mobilizing dealer networks, and monetize incremental improvements years after delivery. Talk about killer margin stability and operational leverage!!!!!

From a labor perspective, this changes where work happens. Historically, vehicle programs absorbed enormous human effort late in the cycle like physical prototyping, tooling changes, validation loops, and post-sale remediation. Also think along the lines of virtual modeling, simulation, and digital twins. Entire layers of downstream engineering effort are being eliminated or pulled forward into abstraction-heavy phases that require fewer people but higher skill density. That is the key here. Remember when I said specialists over generalists… this is what I am talking about… well one part of it anyways.

Read more

People say they miss their childhoods, simpler interfaces, and a time before everything was connected, tracked, and optimized.

It’s a tidy explanation, but it doesn’t really hold up. What people are reacting to isn’t memory really, it is really a case of comparison. They’re noticing that some products and experiences simply do their jobs worse than they used to, even though the technology underneath them is more advanced. It doesn’t happen everywhere, but when it does, the pattern is hard to miss.

Palmer Luckey put it plainly:

“There are a lot of things that were very hard-won lessons that we then forgot because they weren’t profitable to remember.”

That statement captures the core dynamic here. As systems mature and capital becomes abundant, optimization has totally shifted away from performance and towards extraction. Features have accumulated not because they improve outcomes, but because they support monetization models, data capture, or internal justifications for scale.

What is happening now is not a retro trend. It is a reversion… led by builders who understand where incentive structures broke, and who are deliberately reconstructing systems under constraint.

My report below examines that reversion, why it is rational, and why it is appearing simultaneously in consumer hardware, gaming, and even defense technology… long before most institutions are prepared to take it seriously.

Incase you were wondering about the title - “Quake: Arena... it freaking straight slaps” - that was a quote from Reddit Founder, Alexis Ohanian at CES.

Read more

Early in the conversation, he framed the moment with a precision that should not be ignored:

“…it was very clear that we were in a in a moment where the US had to win this race in AI, there was so much we had to do to make sure that we positioned the country in place where we could actually execute. And what it boiled down to is essentially sort of three key themes.”- Kratsios

I’ll touch on those themes in a moment… but let’s discuss the word choice here.

I’ll start by saying he is a very elegant speaker. He did not choose words on the fly. Execute is not an aspirational word. It is logistical. It implies that there is sequencing, bottlenecks, dependencies, and tradeoffs. When that word is used at the level of national policy, labor is never the headline. It is just used as an adjustment variable.

What followed was not a discussion of products or platforms. It was a description of physical infrastructure, energy constraints, regulatory removal, education pipelines, autonomous mobility, healthcare diagnostics, government modernization, and global technology export. Workforce was mentioned and it was always as something to be retrained, or redirected, never as something to be preserved.

That distinction matters.

Because if you track workforce data (like me)… not headlines, not sentiment, not federal data, but real bonafide job architecture, hiring velocity, skill inflation, and role disappearance…. you recognize this phase immediately. It always comes before the layoffs, before the political fight, before the narrative solidifies.

What follows is not criticism. It is straight up my interpretation (along with his direct quotes)… of what was said, what was emphasized, and what the labor data already shows is underway.

Read more

Most consumer disruptions reallocate spending at the margins. GLP-1 reallocates biology, and spending follows. We’re already seeing the early data.

CES is all about tech… and that includes everything health tech, biotech, and digital health. I have been here since Sunday having some amazing conversations with the top leaders, and across the board the advancements we will see will disrupt life as we know it. There is some good, some bad, and some just different.

I attended a fabulous panel with PwC, a research head from UC Davis, and the SVP of Lilly Direct (Eli Lilly) yesterday and they really dug into the realization of what GLP-1 is doing across the globe to society. A major change is coming and we’ve ain’t seen nothing yet. Me… Well I went further into the consequences of GLP-1 on the workforce and the future of work, and I get into that after. But first… some of the initial consumer insights are mind boggling.

20% of households have atleast one person on GLP-1.

Households with one GLP-1 user reduce spend by roughly 4% across apparel, grocery, and travel.

That headline number understates what’s actually happening. This is straight up behavioral rewiring.

When PwC compared consumer interest in GLP-1s to the iPhone at a comparable point in its adoption curve, the difference was not incremental, and honestly it was exponential. Search interest for GLP-1s is materially higher, steeper, and more persistent.

That matters because GLP-1 isn’t changing what people buy first.

It’s changing how people feel, what they want, and how their bodies signal reward.

Read more

The most important thing I heard yesterday at CES Digital Hollywood wasn’t a product launch or a buzzword. It was a statement of posture.

“We’re done with experimentation.”

That line, coming from Adobe, wasn’t about tools. It was about phase change. And my goodness do we feel it on the workforce front too. Digital media, advertising, and content creation are exiting the sandbox era of AI and entering something more operational, more aggressive, and far more monetization-driven.

The conversation across Adobe, AWS, WPP, NBC, and emerging AI platforms made one thing clear:

the next era isn’t about whether AI works: it’s about who controls speed, IP, fandom, and cash flow when creation timelines collapse from weeks to minutes.

YES. Minutes.

Job wise… are we cooked?

The very fact they said… Personalization wasn’t the goal but rather it was the bottleneck… Oh my!

Read more

If anything, the company has the opposite problem. It can imagine too many futures at once, and it has the industrial capacity to gesture convincingly toward all of them. Micro RGB displays that finally make OLED look conservative. AI woven through televisions, appliances, and home systems in a way that feels cohesive instead of bolted on. A vision of domestic technology that is all about “COMFORT, CONFIDENCE and SUPPORT”…. not loud, not gimmicky, not anxious.

From a product standpoint, Samsung looks exceptional.

I want their entire lineup in my home after seeing their lineup unveiled at CES last night.

But desire is not the same thing as readiness, and conviction is not the same thing as momentum. And you know what I look at… When you strip away the CES theater and look at Samsung through the only lens that does not lie… the workforce… a very different picture emerges. Not a bearish one. Not a broken one. But a picture of a company that has deliberately entered a holding pattern, even as the market celebrates its rebound.

That tension is the story. But it is a GOOD one and it has SOOO many layers.

Read more

While everyone else was losing their minds over those Venezuela oil headlines, tankers, sanctions roulette, regime drama… I was looking somewhere else… the part that actually matters for critical mineral chains isn’t the oil noise. This isn’t a Venezeula story though.

For years, Venezuela’s Orinoco Mining Arc has produced illicit coltan, cassiterite, and associated REE-bearing materials. Extraction has run mostly informal… including armed groups, small-scale operators, smuggling networks, with the output feeding primarily into Chinese refining networks that still control the vast majority of global capacity. That setup lived in a tolerated gray area. It didn’t force real change because it stayed off the main grid.

2025 ended that tolerance.

China’s April export controls turned their processing advantage into active leverage. U.S. sanctions pressure ramped up as I’m sure you remember, hitting logistics that support both oil and informal mineral flows. DoW stepped up funding, offtake structures, and partnerships for aligned, traceable sources. The combined effect narrowed viable informal routes. Supply chains that relied on tolerated leakage started facing higher costs around provenance, compliance, and counterparty risk.

This didn’t turn Venezuela into a major formal rare earth supplier overnight.

It did show how quickly tolerance can disappear once multiple pressures align.

As The Job Chick… you know I track workforce because it shows what companies actually expect operationally, well ahead of public statements or market moves. That’s my thing. When constraints tighten, organizations respond differently depending on how they’re structured. Some absorb the change with little disruption. Others have to reorganize… sometimes gradually, sometimes under stress.

The connection here isn’t direct operations in Venezuela. It’s downstream execution really, watching the overall operating when buyers and regulators demand stricter traceability, longer contracting cycles, and heavier governance load.

That’s why MP Materials and Lynas Rare Earths provide the clearest read.

Neither operates in Venezuela.

Read more

That story is technically defensible… and also fundamentally false.

What actually defined 2025 was not mass layoffs in the traditional sense, but a far more sophisticated and less visible phenomenon - labor arbitrage executed at the role, location, and vendor layer rather than the headline headcount level. The effect was the same, lower labor costs, fewer U.S.-based professional roles, diminished internal capability… but the mechanism was guaranteed to be harder to track, and deliberately engineered to evade public scrutiny. In walks me, The Job Chick… damn they should have known.

Here is the true story of the labor arbitrage disaster we saw in 2025 as I have observed it to be: a full operating model.

Post-2022, large companies learned a painful lesson which was traditional layoff announcements carry reputational, political, and regulatory costs. They attract media attention, damage employer brands, and create internal instability that persists far longer than the cost savings justify. That isn’t fun for corporates, right?

So the mechanism changed.

• Instead of cutting people, companies restructured roles.

• Instead of shrinking headcount, they altered job architecture.

• Instead of eliminating work, they relocated it… geographically, contractually, or functionally… until the original workforce hollowed out on its own.

On paper, this looks like restraint. Stocks soared with this news. In reality, it is precision cost extraction.

The most common pattern was not a sudden reduction in total employees, but a gradual erosion of specific job families… particularly mid-level professional roles in engineering, analytics, operations, support, and corporate services. I have overviewed this countless of times over the past year. Job families is something 99.9%+ of people don’t watch, or have access to see what is really happening. I can say with certainty, that these roles didn’t disappear overnight. They were reclassified, absorbed into “centers of excellence,” (remember that one) or simply replaced by offshore equivalents under different titles.

If you only tracked total headcount like so many layoff trackers and even gov data, you missed the story entirely.

Labor arbitrage in 2025 followed a repeatable playbook.

First, companies froze or slowed U.S. hiring in specific functions while maintaining outward-facing job postings. These postings served multiple purposes: signaling growth to investors, sustaining internal morale, and preserving optionality. Many roles were never intended to be filled domestically (not that they said that point blank.)

Second, internal teams were restructured so that responsibilities previously owned by U.S.-based employees were redistributed across global teams, vendors, or automation layers. This redistribution was rarely framed as replacement. It was framed as “operational alignment,” “global delivery optimization,” “platform leverage,” or my very favorite- AI….

Third, attrition was allowed, and in some cases encouraged (think the UPS, Mercedes, and other voluntary buyouts that have happened) —— to do the work that layoffs once did. When senior or experienced employees left, roles were backfilled selectively, often offshore, or not at all. (it was mostly outsourced and offshored.) The org chart remained intact.

Finally, vendor dependency increased. The crazy thing with some of this is that vendors of outsourcing firms have a much higher level of Unbilled Receivables.

I will be doing a full report on this soon. I fully expect that to cause major backlash and maybe even become a catalyst to this bubble popping.

But for now- let me get back to the point- Work that had historically been internal… particularly knowledge-heavy, process-driven work… migrated to consulting firms, offshore service providers, or managed platforms. The risk did not disappear in the slightest. It was simply moved off the balance sheet.

This is why so many companies could honestly say they were not “laying people off” while simultaneously extracting substantial labor cost savings. Don’t believe what you hear on the news or from base level job data. The workforce data per company looks VASTLY different.

The arbitrage was real. Boards stayed fairly silent.

While this pattern appeared across the economy, it was most visible in four sectors.

Technology firms led the way (obv, right?), having already normalized large-scale workforce restructuring in prior years. In 2025, tech companies focused less on headline cuts and more on role dilution… shrinking seniority bands, collapsing job ladders, and shifting execution-heavy work to lower-cost regions while preserving a thin layer of strategic leadership domestically.

Consulting and professional services followed closely. You think Accenture only laid off the few thousand they mentioned? NOPE. I have a report on that live if you are aiming to learn more. Promotions and titles proliferated, but compensation growth lagged and delivery work increasingly moved offshore. The optics were of career progression — - the reality was margin defense… or rather, margin growth. Companies focused on growing margins, not growing teams.

Aviation and transportation companies used a hybrid model: selective layoffs combined with aggressive outsourcing and role consolidation. Maintenance, analytics, scheduling, and support functions were particularly vulnerable, even as overall staffing levels appeared stable. Aviation has someone come through this very well as you have seen in my X and other reports. Transportation- not as much.

Oil, industrials, and energy services applied arbitrage more cyclically. Rather than mass cuts, they relied on contractor churn, delayed rehiring, and regional labor substitution to maintain flexibility ahead of uncertain demand. We still hit my roughly 50K layoffs in 2025. Isn’t that wild. Then you add in all the economic multipliers that increase layoffs… and DAMN.

In every case, the pattern was the same: stability at the top-line metric, erosion underneath.

To understand why labor arbitrage became the dominant strategy in 2025, it helps to look at what companies quietly removed before they stopped announcing layoffs altogether.

Some of the projected and finalized white-collar workforce reductions discussed throughout 2025:

• Microsoft — 16,000+ roles eliminated across multiple waves; engineering, product, and commercial sales most affected, with significant role elimination masked by reorgs and offshore substitution rather than headline cuts.

• Intel — 15,000 roles targeted globally through restructuring and internal job architecture collapse; engineering and manufacturing support functions disproportionately impacted.

• Amazon — 27,000 cumulative layoffs since late 2022, with continued 2025 white-collar erosion through role elimination, vendor displacement, and hiring suppression rather than mass announcements.

• Google (Alphabet) — 12,000 roles cut initially, followed by ongoing 2025 workforce thinning via reclassification, internal mobility constraints, and offshore hiring concentration.

• Meta — 21,000 roles eliminated across efficiency waves; 2025 marked by stabilization in headcount but continued internal role compression and automation-driven displacement.

• Accenture — tens of thousands affected indirectly through slowed hiring, delayed promotions, bonus suppression, and delivery offshoring despite headline promotion optics.

• Halliburton — 3,500–5,000 projected workforce reductions tied to restructuring, margin compression, and global demand softness; later validated by earnings disclosures and profit declines.

• Ford — projected 8,000–13,000 white-collar role eliminations through 2025–2026, driven by software, EV, and corporate restructuring signals rather than single-event layoffs.

• United Airlines — targeted white-collar reductions and leadership thinning paired with insider stock sales and operational consolidation.

• JAMF (Apple enterprise vendor) — dozens of U.S.-based engineering and corporate roles eliminated while work shifted to Europe and India, illustrating vendor-layer arbitrage rather than direct Apple headcount cuts.

Most analysts and commentators focused on how many people companies employed, not what kind of work those people were doing. But a workforce with the same headcount and a radically different skill composition is not the same workforce… operationally, strategically, or even financially.

Another missed signal was leadership shrinkage. In many organizations, management layers really thinned months before any visible cost action occurred. Fewer directors, fewer senior managers, broader spans of control. This was in complete preparation.

Visa pipelines were another really big tell and the American public has become hyper aware of what is going on here. Companies that claimed domestic hiring was “difficult” continued to expand early-career and visa-dependent intake channels, effectively locking in lower-cost labor for future cycles while reducing reliance on experienced domestic talent. AKA the whole push for visa candidates is a total lie.

Don’t forget that vendor concentration increased. When companies rely more heavily on a smaller set of external providers, it often signals internal capability erosion- which is a form of arbitrage that boosts short-term margins while increasing long-term fragility. Recognize that, right?

None of these show up cleanly in earnings headlines. All of them show up in the operating reality.

The most important mistake observers continue to make is treating labor arbitrage as a temporary response to inflation, interest rates, or post-pandemic excess.

It isn’t.

What changed in 2025 was not corporate behavior… it was corporate capability. This is global. This is not just happening in the USA.

Once organizations learned how to extract labor costs without triggering layoffs, they had no incentive to revert. Why would they? The tools now exist to reshape work continuously…. including global delivery platforms, role modularization, vendor substitution, automation layers, and visa-dependent pipelines that lock in long-term cost advantages.

This is not cyclical at all.

The workforce that existed prior to 2020 that included stable career ladders, geographic concentration, predictable internal mobility… is not coming back.

Companies no longer optimize for employee continuity. They optimize for labor optionality.

That means future downturns will not look like past recessions. They will arrive quietly, role by role, region by region, function by function … often without a single press release.

Labor arbitrage is not a phase, but rather the new operating systems and boards are happy. What this does to the workforce looking for jobs is unclear, but we have a big hill to climb, that is for sure.

Once an operating system changes, everything built on top of it behaves differently.

Why this matters going into 2026…..

The labor arbitrage of 2025 did not resolve structural workforce issues. It just deferred them. I feel like kick the can down the road was a BIG phrase of 2025. It will bite us in the butt soon enough. Companies bought time. The cumulative effect is a thinner, brittle internal workforce, increasingly dependent on external execution and vulnerable to disruption when conditions change. None of this is being addressed.

For investors, this means headline efficiency gains may not translate into durable performance. For workers, it explains why job security feels worse even when layoffs slow. For policymakers and analysts, it highlights why traditional labor metrics increasingly fail to capture what is actually happening.

Most importantly, it sets the stage for 2026. My clients have already seen my layoff and strong performer chart for 2026 by sector. I will be releasing this slowly over the next couple weeks. Please utilize it and if you are looking for a job (which you always should be) please be hyper aware of what is happening in all sectors and how it could impact yours.

The same mechanisms that reshaped the workforce in 2025 will determine which organizations can adapt when growth, regulation, or technology forces the next adjustment. Those who understand this dynamic early will not be surprised by what comes next. They will already be watching the right signals.

If you want to understand what’s actually happening inside companies before it shows up in headlines, earnings calls, or “unexpected” announcements, I recommend the Insider Edge Report.

This is where I publish my workforce intelligence in real time, analysis built from internal job-flow data, attrition patterns, and operating behavior that quietly reveal structural stress, silent contraction, and strategic shifts well before they’re publicly acknowledged.

I analyze markets the way serious analysts always have: by tracking repeatable patterns across cycles, grounding today’s signals in historical precedent, and discarding narratives that can’t survive contact with data.

If you want a front-row seat to what companies are actually doing… not what they’re saying… Insider Edge Report is where I publish it first.

Recommend Insider Edge Report

If you’re reading this from an Oracle IP… welcome.

And if you understand how subprime actually worked, the parallels should be uncomfortable.

Oracle isn’t just laying people off. It’s repricing human labor… again.

What’s happening inside Oracle right now isn’t a workforce event or an AI transition story… it is straight up a a risk transfer. The same kind of risk transfer that defined the subprime era: volatility moved off the balance sheet, smoothed by models, and pushed downstream until it no longer appeared where decisions were being made. The layoffs are not the story. They are simply the mechanism by which labor is being abstracted, sliced, and reclassified.

To understand what’s unfolding at Oracle, you have to stop counting exits and start looking at composition. Layoff headlines are noise. I mean, you have probably followed me for years, so you know the drill. See the headline- don’t believe a word they say- we dive into the workforce data instead. This is the same thing with Oracle and why I am going through this now.

The signal I’m looking at is how the workforce is changing as a percentage of the whole, which functions are being removed versus rebuilt, and how responsibility is redistributed once humans are taken out of the loop. That’s where the subprime parallel stops being metaphorical and starts being structural.

By the end of 2025, what Oracle experienced was not a single layoff event but a multi-quarter workforce contraction that cumulatively removed roughly ten to twelve percent of the company’s human labor capacity, depending on how you measure internal versus externalized roles.

Read more