November 20, 2025

What Nvidia's $57 Billion Quarter Really Tells Us About the AI Future

Investors exited with $11 Billion before Nvidia announced record earnings — what do they know that others don't?

By Jay Cadmus

What Nvidia's Earnings Actually Proved (And What They Didn't)

Let's start with what we learned yesterday.

What the earnings proved: AI infrastructure demand is real and accelerating. Nvidia's data center segment generated $51.2 billion, up 66% from last year and beating estimates of $49.3 billion. The Blackwell Ultra architecture is driving demand across all customer segments. Q4 guidance of $65 billion signals no slowdown through at least early 2026.

This matters because it directly contradicts the bear case that hyperscalers would cut AI spending due to lack of ROI. Microsoft, Amazon, Google, and Meta are still pouring money into infrastructure at unprecedented rates.

What the earnings didn't prove: That this spending will ever generate profitable returns.

Here's the gap nobody's addressing. Nvidia's success measures input (how much customers are spending on infrastructure). It says nothing about output (whether that infrastructure generates revenue that exceeds costs).

Think about it this way. Nvidia reporting record chip sales is like a construction company reporting record sales of steel and concrete during a building boom. It proves people are building. It doesn't prove the buildings will ever have paying tenants.

And the workforce data released this week suggests the buildings might be emptier than anyone wants to admit.

The Jobs Data Nobody's Connecting to AI Reality

The Bureau of Labor Statistics released September's delayed jobs report on November 19. The tech sector added jobs, but the composition reveals the deployment crisis in stark terms.

Tech job postings grew 1.3% month-over-month in October, a significant deceleration from September's strong performance. Year-over-year, tech job postings declined 13.8% compared to October 2024. But here's the critical detail: 51% of U.S. tech job postings now require AI skills, up from 50% in September and representing a 65% increase from October 2024.

Translation: The tech job market is shrinking overall while simultaneously demanding AI fluency as table stakes. Companies aren't hiring more people. They're hiring different people with different skills.

AI-related job postings surged 71% year-over-year according to tech employment data. But remember: 48,414 workers lost jobs citing AI as the reason through November 2025. The jobs being created and the jobs being destroyed aren't swaps. They're different roles requiring different skills in different locations for different people.

The occupations seeing the strongest growth in October tell the story:

ServiceNow Administrators: +150% month-over-month Lead Solution Architects: +150% month-over-month Marketing Data Scientists: +150% month-over-month iOS Software Engineers: +150% month-over-month Lead Data Scientists: +150% month-over-month Operations Engineers: +140% month-over-month Lead Network Engineers: +140% month-over-month

These aren't entry-level jobs. They're specialized, senior roles managing AI deployments, optimizing enterprise software, and building data-driven capabilities. The Bureau of Labor Statistics projects that over the 2023 to 2033 period, AI will primarily affect occupations whose core tasks can be most easily replicated by generative AI. But software developers are projected to grow 17.9%, database administrators 8.2%, and personal financial advisors 17.1% because these roles use AI rather than get replaced by it.

The gap is clarifying: Roles that deploy, manage, and optimize AI are exploding. Roles that AI can replicate are vanishing.

And this is where Nvidia's blowout earnings become more complicated. Because the $65 billion in Q4 guidance will prove that companies are still buying infrastructure. But the workforce fragmentation proves most companies still can't deploy what they're buying.

Why Smart Investors Sold Before the Rally

Now let's talk about what Peter Thiel, SoftBank, and Michael Burry saw that made them exit billions in Nvidia positions before yesterday's blowout earnings.

They're not betting against AI infrastructure demand. Nvidia's results proved them wrong on that bet if that's what they were making. But I don't think it was.

They're betting against the revenue model that justifies the infrastructure spending.

Here's what they see. Big Tech will spend roughly $300 billion on AI infrastructure in 2025. Generative AI revenue? $69 billion. That's only a 23% return on investment. Not 23% profit margin. 23% revenue capture against total spend.

To justify current and planned AI infrastructure investment, the industry needs to generate approximately $2 trillion in annual revenue by the end of the decade. Not $2 trillion total. $2 trillion per year.

Nvidia's earnings proved the $300 billion spend is happening. Smart money exiting before the rally suggests they don't believe the $2 trillion revenue ever materializes.

SoftBank's $5.8 billion Nvidia sale: They didn't exit AI. They sold off to buy Ampere Computing for $6.5 billion, swapping passive exposure to Nvidia for direct ownership of an AI chip competitor. Masayoshi Son is betting $30.5 billion this quarter, mostly on OpenAI infrastructure and the Stargate project.

Translation: SoftBank isn't bearish on AI. They're bearish on Nvidia's ability to maintain 85% market share and bullish on controlling the infrastructure layer directly. When you see smart money swap $5.8 billion from owning shares in the dominant player to owning a distant competitor, they're positioning for market restructuring. They think Nvidia's monopoly faces real threats within 18 to 24 months.

Peter Thiel's $94 million exit: Thiel Macro is a small fund with $155 million in assets. The Nvidia stake was huge relative to fund size. Selling could mean taking profits after massive gains, rebalancing, or genuine belief that Nvidia has peaked despite yesterday's earnings.

The timing (liquidating right before Nvidia's blowout quarter) suggests Thiel isn't reacting to current fundamentals. He's positioning for what comes after the infrastructure build-out phase. When a billionaire who co-founded PayPal and was Facebook's first outside investor exits a position weeks before the company reports record earnings, he sees past the current quarter to longer-term structural shifts.

Michael Burry's $1.1 billion short positions and cryptic announcement: Burry bought put options giving him the right to sell Nvidia at $110 in December 2027 and Palantir at $50 in January 2027. Nvidia currently trades around $185 post-earnings. Palantir is above $70.

Burry is betting on 40%+ declines in both stocks within 24 months despite yesterday's blowout earnings. This is the investor who called the 2008 housing crash. He's also publicly accused hyperscalers of accounting fraud, claiming they understate depreciation on AI chips by extending useful life assumptions from 2–3 years to 5–7 years, inflating reported earnings by $176 billion that would be between 2026 and 2028.

Then Burry deregistered Scion Asset Management with the SEC and made cryptic social media posts hinting at contrarian research launching November 25th, particularly focused on his bearish stance on the AI boom. He referenced "much better things" and releasing research that challenges consensus views on AI valuations.

When the "Big Short" investor calls out accounting practices, takes massive bearish positions despite record earnings, then teases a major reveal days before Thanksgiving, that's not hedging. That's conviction that the market is fundamentally mispricing something structural.

The Deployment Crisis Nvidia's Earnings Can't Solve

Here's what connects all of this: workforce segmentation, institutional exits, and Nvidia's record quarter.

Only 5% of companies globally are "future-built" for AI, according to Boston Consulting Group's 2025 study of 1,250+ organizations. These future-built companies achieve 1.7x revenue growth, 3.6x total shareholder return, and 1.6x EBIT margin compared to peers.

Another 35% are "scalers" in the process of deploying AI. The remaining 60% are "laggards" reporting minimal revenue and cost gains despite substantial investment.

Let that sink in. 60% of companies investing in AI are getting almost nothing back. They're the ones buying Nvidia chips. They're contributing to that $57 billion quarterly revenue. And they can't figure out how to make the infrastructure generate returns.

This is BCG's "Build for the Future 2025" research published in October 2025, surveying senior executives and AI decision-makers across nine industries globally. The study explicitly defines future-built companies by five characteristics:

Lead from the top with bold, multi-year AI ambition Redesign the business with value-based prioritization and measurable outcomes Adopt an AI-first operating model grounded in human-machine collaboration Secure and develop talent, anticipating new skills and investing in upskilling Build the right tech and data foundations to scale safely

The companies that master these five strategies aren't just incrementally better. They're five times more effective at revenue generation and three times more effective at cost reduction from AI compared to laggards.

The gap is widening, not closing. Future-built firms plan to spend more than twice as much on AI compared to laggards in 2025, and they expect twice the revenue increase and 40% greater cost reductions as a result.

Nvidia's earnings prove companies are buying infrastructure. BCG's research proves most of them have no idea how to use it. And the workforce data proves they're cutting the wrong people while hiring desperately for the roles that bridge the deployment gap.

That's why smart money sold before the rally. Because the continuation of infrastructure demand doesn't solve the enterprise AI deployment crisis. It makes it worse.

The Circular Financing Problem

This is where things get genuinely concerning, and why smart investors are exiting despite record earnings. Pay attention, because this affects the entire economy, not just tech stocks.

Nvidia committed a letter of intent to deploy approximately 10 gigawatts of Nvidia systems to OpenAI, intending to invest up to $100 billion as deployments progress. In plain terms: Nvidia is buying its own chips and giving them to OpenAI on credit to build infrastructure.

Microsoft's OpenAI partnership includes massive Azure commitments, so OpenAI's Azure consumption contributes to Microsoft's reported revenue. Translation: when OpenAI spends money on Microsoft's cloud services, Microsoft books that as revenue even though the money is just moving in a circle between partners.

GPU-rich cloud providers are raising multi-billion dollar debt with GPUs as collateral. They're borrowing money and putting graphics cards up as the asset backing the loan, like using your house as collateral for a mortgage. Except houses don't depreciate 50% every two years. GPUs do.

This creates a self-reinforcing loop: Nvidia invests in OpenAI infrastructure using Nvidia chips. OpenAI's spending on Azure inflates Microsoft's AI revenue. Hyperscalers borrow against GPU assets to fund more infrastructure. Everyone's growth depends on everyone else continuing to spend.

Harvard Business Review flagged this as "emblematic of an accelerating AI arms race" with "speculative capital and circular financing" that "could distort market expectations". In the 2008 financial crisis, we learned what happens when growth is built on circular financing instead of real revenue. Banks were lending to each other, using each other's debt as collateral, and calling it growth. Until it wasn't.

40 to 60% of U.S. real GDP growth in the first half of 2025 is explained by IT and AI investment. If Big Tech pulled spending back to 2022 levels, it would erase approximately 30% of revenue growth currently expected for the S&P 500.

Let that sink in. The American economy has effectively outsourced its growth engine to a handful of corporate boards betting on unproven revenue models. If they're wrong, or if they slow spending, GDP growth collapses.

Oracle's market cap loss tells the story investors don't want to hear. On September 10, 2025, Oracle announced a $300 billion OpenAI infrastructure partnership. Market cap: $615 billion. Then, on November 18, 2025: Market cap, $315 billion.

Oracle lost more market value than the entire deal it signed in the two months after announcing it. The market looked at the numbers, did the math on returns, and concluded the deal destroys value instead of creating it. Nvidia's blowout earnings don't change that calculation. They confirm the spending is real. They don't confirm the returns ever materialize.

Credit-default swaps (insurance against default) on Oracle debt surged following $18 billion in bond sales to fund the partnership. When the market starts hedging against your ability to service debt from an AI deal, that's fear, not optimism.

The Workforce Paradox: Why Companies Are Cutting AND Hiring at Record Rates

This is where the story hits home for most people. Because AI isn't just changing markets. It's fundamentally restructuring who works and who doesn't.

491 people lose their jobs to AI every single day in America. That's 48,414 workers through November 2025. Companies aren't hiding it anymore. They're openly citing AI as the reason for layoffs.

Simultaneously, Forward Deployed Engineer postings grew 800% from January to September 2025. Companies are desperately hiring people who can actually make AI work in production environments.

This isn't contradiction. It's clarification of who wins and who loses in the AI economy.

Entry-level white-collar jobs doing routine tasks are vanishing. Customer service representatives, data entry clerks, junior accountants, paralegal assistants, entry-level programmers doing repetitive coding tasks. Dario Amodei (Anthropic CEO) predicts AI could eliminate half of all entry-level white-collar jobs within five years.

Stanford University researchers found early-career workers ages 22 to 25 in the most AI-exposed occupations experienced a 13% decline in employment relative to less exposed occupations. Entry-level hiring in Big Tech dropped 25% in 2024 versus 2023. January 2025 saw the lowest job openings in professional services since 2013, a 20% year-over-year drop.

AI-native implementation roles are exploding. Forward Deployed Engineers embed with customers, write production code, configure systems, and actually make deployments work. These are the people who take AI from demo to reality. And they're the scarcest resource in the AI economy that just validated itself.

Accenture laid off 11,000 employees that it "can't retrain with AI skills". Salesforce cut 4,000 jobs as it implements AI agents. Meta eliminated 600 AI jobs even while expanding superintelligence labs.

The pattern is brutal and clear. AI isn't replacing workers uniformly. It's segmenting the workforce into those who can deploy AI and everyone else.

Let me give you specific examples of who's affected right now:

Jobs disappearing fast: Customer service reps at insurance companies, junior financial analysts doing routine modeling, legal document reviewers, content moderators, basic software testers, data entry specialists, appointment schedulers, travel agents, simple tax preparers.

Jobs growing fast: AI integration specialists, Forward Deployed Engineers, prompt engineers, AI trainers, human-in-the-loop supervisors, AI ethics and governance specialists, agent orchestration engineers, AI product managers.

Jobs safe (for now): Skilled trades (electricians, plumbers, mechanics), healthcare workers requiring human touch (nurses, physical therapists), creative professionals requiring taste and judgment (designers, writers, strategists), complex problem-solvers, people managers, sales professionals building relationships.

The gap isn't just about skills. It's about who can access retraining fast enough. AI skills command a 23% wage premium, exceeding the value of degrees up until PhD level at 33%. But if you're a 45-year-old customer service rep who just lost your job, how do you become an AI integration specialist in six months?

You probably don't. And that's the workforce crisis that Nvidia's record earnings just intensified. Because more infrastructure spending means more deployment needs. More deployment needs means more specialized roles. More specialized roles means faster displacement of those who can't adapt.

What Happens Next: A Clear Timeline of the AI Future (2026–2030)

Let me walk you through what's coming in specific stages. Not abstract predictions. Concrete scenarios based on Nvidia's validated infrastructure demand, institutional repositioning, and the deployment gap that's widening.

Q1 2026: The ROI Reckoning

Forrester's prediction is already playing out. Enterprises will defer 25% of AI spend into 2027. Here's what that looks like in practice.

CFOs are demanding demonstrable ROI (return on investment) within 6 to 12 months. Projects that can't show returns in 90 days get cut. Only 15% of AI decision-makers reported an EBITDA lift in the past 12 months. Fewer than one-third can tie AI value to P&L changes.

This correction accelerates in Q1 2026. Companies that can't prove business value by March 2026 see budgets slashed despite Nvidia's proof that infrastructure works. Pilot projects vanish. Valuations of startups without implementation credibility reset 30 to 50%.

Specific scenarios you'll see:

A Fortune 500 retailer kills a $50 million AI customer service project after realizing it reduced call volume but increased customer complaints and return rates by 18%. A major bank shuts down 12 of 15 AI pilots after none can prove they generated more revenue than they cost, despite spending $80 million on Nvidia infrastructure. AI vendors without paying customers at scale start cutting staff, pivoting strategies, or seeking acquirers at distressed valuations (60 to 80% below peak). Enterprise software prices jump 40% as vendors embed AI features customers didn't request but now have to pay for, even though most can't deploy them.

Mid-2026: Agent Orchestration or Chaos

By mid-2026, vendor fragmentation forces a majority of enterprises to build "agentlakes" (composable agent architectures that manage and orchestrate fractured AI deployments).

Think data lakes (centralized repositories for all data), but for agents. Multiple vendor agents, custom agents, third-party agents, all needing orchestration, governance, and interoperability.

Here's the problem. Company A deploys a Salesforce agent for customer service, a Microsoft agent for internal IT, a custom-built agent for supply chain, and a third-party agent for HR. These agents start making decisions that conflict with each other. The customer service agent promises a delivery date the supply chain agent knows is impossible. The IT agent locks accounts the HR agent just approved. Chaos.

The companies that solve multi-agent orchestration infrastructure win. Those that don't end up with ungovernable chaos: hundreds of agents with unclear accountability, overlapping responsibilities, and no audit trails.

Gartner predicts 40% of agentic AI projects will be canceled by 2027. Most failures will trace back to orchestration breakdown, not model quality. Nvidia proved the chips work. The orchestration layer is where deployments collapse.

Specific scenarios:

A healthcare system's insurance verification agent contradicts its patient intake agent, delaying critical treatments and triggering $12 million in malpractice lawsuits. A manufacturer's procurement agent buys materials the production agent hasn't been programmed to use, creating $3 million in waste before humans notice. New roles emerge with $180K to $250K salaries: Agent SRE (Site Reliability Engineer), Chief Agent Officer, Agent Governance Analyst.

Q4 2026: The First Major Agent Outage

SAS predicts that by end of 2026, Fortune 500 companies will report agentic systems autonomously resolving over 25% of multi-step customer interactions with measurable revenue impact.

That also means the first major "agent outage" will hit headlines. When autonomous systems drive revenue, downtime has a price tag. And when an agent makes an autonomous decision that costs a company millions, accountability becomes the front-page story.

Imagine this scenario: A major airline's AI agent autonomously reprices 50,000 tickets based on faulty demand predictions, underpricing peak routes by 40% and overpricing off-peak flights by 60%. The company loses $40 million in a single day before humans notice. Customers whose tickets were repriced sue for breach of contract. Regulators investigate. The CEO testifies before Congress about autonomous systems making financial decisions without human oversight. Stock price drops 8% in a week.

Or this: A bank's loan approval agent autonomously approves 2,000 mortgages that violate underwriting standards because a training data error taught it to ignore debt-to-income ratios above 50%. The bank discovers the error six months later when default rates spike to 12% on the affected cohort. Shareholders file lawsuits. Regulators fine the bank $500 million for inadequate AI governance. Three executives resign.

Enterprises discover that agents aren't just tools. They're infrastructure that requires monitoring, governance, and disaster recovery. Companies without governance frameworks built from the start scramble to retrofit controls. Those with audit trails, human-in-the-loop checkpoints, and kill switches adapt. Those without face regulatory walls, lawsuits, and public relations nightmares.

2027: Half of Enterprises Deploy Agents (But Most Fail)

IDC predicts that by 2027, 50% of enterprises will use AI agents to redefine how humans and machines collaborate. Not pilot. Deploy.

Gartner forecasts that by 2028, 33% of all enterprise software will incorporate agentic AI capabilities. Microsoft's AutoGen framework is already adopted by over 40% of Fortune 100 firms.

But here's the gap. Adoption doesn't equal success. The 50% deploying agents splits into two cohorts that look nothing alike:

The 10 to 15% who figured out orchestration, governance, and workforce transformation: These companies see 5x revenue increases and 3x cost reductions. They embedded Forward Deployed Engineers, redesigned workflows around AI capabilities instead of bolting AI onto broken processes, trained workforces on AI-native skills, and built platform infrastructure for agent management.

Example: A global logistics company deploys 47 specialized agents handling route optimization, warehouse management, customs documentation, and customer service. They invested $80 million upfront in orchestration infrastructure, governance frameworks, and workforce training (retraining 2,400 employees and hiring 180 Forward Deployed Engineers). Result: 31% reduction in operating costs, 22% increase in on-time delivery, 40% reduction in customer service headcount, but 15% increase in specialized roles managing agent fleets. Net EBITDA improvement: $220 million annually.

The 35 to 40% who deployed without infrastructure: These companies face agent collisions (agents making contradictory decisions), coordination failures (agents optimizing local objectives that harm global performance), security incidents (agents accessing data they shouldn't), and compliance nightmares (agents making decisions that violate regulations). Their agents work individually but fail systemically.

Example: A retail chain deploys agents across inventory, pricing, marketing, and customer service without central orchestration. The pricing agent drops prices 30% to clear inventory. The marketing agent runs promotions assuming full stock availability. The inventory agent reorders based on historical patterns, not current promotions. The customer service agent promises two-day deliveries the logistics system can't fulfill because inventory is already depleted. Result: $120 million in lost revenue in Q3, customer satisfaction scores drop 18%, and the company spends $40 million unwinding the deployment and rebuilding with proper orchestration.

2027–2028: The Great Consolidation

Gartner already predicts AI market consolidation as vendor supply exceeds demand. With agentic AI vendors proliferating, 2027 to 2028 sees massive M&A activity that reshapes the competitive landscape.

The pattern:

Companies with clear paths to profitability, strong customer retention, and proven implementation capability get acquired at premium multiples by hyperscalers or private equity firms capitalizing on valuations that finally make sense after years of inflation. Saturated categories consolidate rapidly as strategic moats narrow. Ten competitors in AI customer service collapse to three through acquisitions and bankruptcies. Undifferentiated AI vendors without implementation teams or deployment infrastructure vanish or merge at distressed valuations (pennies on the dollar compared to peak).

The gap between early adopters and stragglers widens irreversibly. By 2028, the companies that invested in deployment infrastructure, workforce transformation, and governance in 2025 to 2026 own their markets. Those that piloted endlessly are five years behind technologically and can't catch up organizationally.

Specific scenarios:

Salesforce acquires three specialized agent vendors for $12 billion combined, consolidating CRM agent capabilities and eliminating competition. Microsoft buys a failed AI startup's technology for $50 million (80% below its 2024 peak valuation of $250 million) and integrates it into Azure AI services within six months. 40% of AI startups from the 2023 to 2024 boom either shut down, get acquired for pennies on the dollar, or pivot to entirely different markets. Only 15% survive as independent companies.

2029–2030: Agentic AI Becomes Infrastructure (With Uneven Impact)

By 2030, agentic systems account for nearly half of all AI spending. The World Economic Forum projects 70% of office tasks could be automated by AI with agency.

But here's the nuance. Automation doesn't mean elimination. It means reconfiguration.

New roles emerge faster than old roles vanish, but not for the same people. 12 million jobs in AI integration and organizational transformation emerge by 2030. 8 million jobs in AI training, evaluation, and human-in-the-loop roles.

These aren't one-to-one swaps. Entry-level displaced workers don't automatically become AI integration specialists. Geographic displacement compounds the problem: jobs vanishing in Kansas City, while emerging in San Francisco, Austin, and Seattle.

The result: structural unemployment coexisting with talent shortages. Companies can't find AI-fluent workers willing to relocate or able to command $200K+ salaries. Displaced workers can't access retraining fast enough or can't afford to relocate to where the jobs are. Governments struggle to bridge the gap with programs that launch three years too late and target the wrong skills.

By 2030, the workforce looks fundamentally different:

30% of workers are AI-native: They use AI tools daily, understand limitations instinctively, integrate AI into every workflow, and command 25 to 35% wage premiums. 40% of workers are AI-adjacent: They work in roles AI can't easily replicate (complex judgment, human relationships, creative problem-solving, physical work requiring adaptability). Wages stagnate but employment remains stable. 30% of workers are AI-displaced: They lost jobs to automation and either exited the workforce permanently, accepted lower-paying work outside their original field, or remain stuck in retraining programs that don't match market needs. This cohort faces permanent income loss averaging 40%.

The social and political implications reshape elections, labor movements, and economic policy for a generation. Universal Basic Income pilots expand to 15 states. "Right to human service" laws pass in California, New York, and Massachusetts. Federal AI displacement insurance becomes a 2028 presidential campaign issue.

What This Means For You: No Sugarcoating, No Spin

Let me make this directly relevant to your situation, whether you're running a business, working a job, or just trying to understand how Nvidia's earnings and the deployment crisis affect your life.

If You're Leading an Organization

Stop asking "which model should we use?" You're solving the wrong problem. Your problem is deployment.

The constraint isn't model access. It's deployment capability. Can you move from pilot to production in 90 days? Can you show audited ROI by Q1 2026? Do you have Forward Deployed Engineers or the equivalent? Do you have platform infrastructure for agent orchestration? Have you redesigned workflows around AI capabilities, or are you just adding AI to broken processes?

If you can't answer yes to most of those questions, you're in the 60% of the "laggards" group that will still be figuring this out in 2027 while the 5% of future-built companies pull five years ahead.

Here's what you do now:

Audit every AI project for ROI within 90 days. Kill anything that can't prove business value by March 2026. Nvidia validated that infrastructure spending is real. That makes it even more critical you can justify YOUR spending specifically. Hire or contract Forward Deployed Engineers immediately. They're more valuable than data scientists right now because they bridge the gap between capability and deployment. Budget $200K to $300K per FDE. Build orchestration infrastructure before deploying more agents. If you have more than three agents in production without central governance, you're heading for the chaos scenarios described above. Budget $2M to $10M for enterprise orchestration platforms depending on scale. Invest in workforce transformation, not just technology. The 5% of companies getting 5x returns from AI spent as much on change management, training, and organizational redesign as they did on technology. Budget 1:1 ratios minimum.

If You're in the Workforce

Entry-level jobs doing routine cognitive work are vanishing at 491 per day. Nvidia's earnings just validated the infrastructure that accelerates this. By 2027, half of entry-level white-collar roles could be gone.

Your survival strategy: become AI-native or become AI-adjacent. AI-native means you use AI tools daily, understand their limitations, and integrate them into your workflow instinctively. AI-adjacent means you work in roles AI can't easily replicate: complex judgment, human relationships, creative problem-solving, physical work requiring adaptability.

Skills command 23% wage premiums. Degrees without skills don't. If you're not upskilling in AI fluency right now, you're losing ground every quarter.

Here's what you do now:

Assess your role's AI exposure honestly. If 70% of your job is routine, repeatable tasks, start looking for your next role today. Don't wait for the layoff notice. Your job is on the line. Learn AI tools in your domain immediately. If you're in marketing, master AI content tools (Claude, ChatGPT, Jasper). If you're in finance, learn AI analysis platforms (Tableau AI, Power BI with Copilot). If you're in operations, understand AI automation systems (UiPath, Automation Anywhere). Spend 5 hours per week minimum. Develop uniquely human skills AI can't replicate: Complex negotiation, creative problem-solving requiring taste, building trust-based relationships, strategic thinking across domains, empathy-driven communication, physical work requiring judgment and adaptation. Consider Forward Deployed Engineer paths if you have technical aptitude. These roles are growing 800% and will command $250K+ salaries by 2027. Bootcamps exist. Career switchers are succeeding. This is the gold rush role.

If You're Investing or Watching the Market

Yesterday's Nvidia earnings answered one question: infrastructure demand is real. But smart investor money exiting beforehand suggests they're asking different questions.

Watch for these:

Michael Burry's November 25th reveal: Five days from now, Tuesday before Thanksgiving, Burry teased releasing contrarian research on AI valuations. He's betting $1.1 billion that Nvidia and Palantir drop 40%+ despite yesterday's blowout. What does he know that the others tuned into market doesn't?

Q1 2026 earnings calls: Listen for CFOs demanding AI ROI. Language shifts from "we're investing in AI" to "here's the return we're seeing." Companies that can answer with specifics get more runway. Those that can't see budgets cut despite any proving grounds that infrastructure works.

Forward Deployed Engineer hiring velocity: If OpenAI, Anthropic, Salesforce, and others keep expanding these teams, implementation remains the scarce resource. If hiring slows, they found scalable alternatives (unlikely) or gave up on enterprise (very bad signal).

Vendor consolidation by mid-2026: M&A accelerates. Watch for hyperscalers acquiring specialized vendors, private equity buying implementation capability at distressed valuations, and undifferentiated startups vanishing.

If You're Building with AI

We know that there's high demand on infrastructure buildup. That raises the bar for you. Customers now expect production results, not infrastructure promises.

Implementation-market fit matters as much as product-market fit. Can customers actually deploy your solution? Do you have the human infrastructure (Forward Deployed Engineers or equivalent) to help them? Can you prove ROI in their first quarter, not their first year?

The honeymoon where enterprises paid for pilots that went nowhere is closing out fast. Financial rigor is here. Companies that can ship, measure, and prove value quickly survive. Those stuck in endless pilots don't.

Here's what you do now:

Build Forward Deployed Engineering into your business model from day one. Don't treat implementation as an afterthought or professional services upsell. Make it core product delivery. Budget 1 FDE per 5 to 10 enterprise customers. Price for value, not usage. Customers will pay for measurable business outcomes (cost reduction, revenue increase, time savings). They won't pay for API calls, model access, or infrastructure promises without proven ROI. Shift to value-based pricing. Focus on deployment speed as your competitive advantage. If you can go from contract to production in three weeks while competitors take six months (like Salesforce Agentforce did), you win. Speed to value matters more than feature breadth now. Build governance and orchestration from the start. Enterprises deploying multiple agents need platforms that prevent chaos. If your product requires manual orchestration or creates governance nightmares, you lose to competitors who solved this.

The Uncomfortable Truth About What's Coming

2026 is the year AI loses its hype and gains its infrastructure. Yesterday's Nvidia earnings just validated that transition.

Forrester's prediction is the headline: "AI Moves From Hype To Hard Hat Work". The art of the possible succumbs to the science of the practical. Enterprises stop caring about what AI could theoretically do and start demanding what it actually delivers in production.

For the 5 to 10% of companies with deployment infrastructure, 2026 to 2027 is transformative. They scale agents, show measurable ROI, and pull ahead of competitors stuck piloting.

What's different about this moment: We're not debating whether AI works anymore. We're watching markets, workforces, and organizations re-structure around the gap between what AI infrastructure can deliver and what humans can actually implement at scale.

Everyone else learns the hard way that making AI work in production is harder than making AI work in demos. By the time they figure that out, those entities in the top 5% are untouchable by that point.

It's already happening. And now the timeline has been accelerated.

Which side of the gap will you be on?

References

Yahoo Finance. (2025, November 20). Nvidia stock soars after Q3 earnings, forecasts top estimates with sales for AI chips 'off the charts'. https://finance.yahoo.com

Reuters. (2025, November 19). Nvidia Q3 earnings beat estimates, boost tech stocks. https://www.reuters.com

Advanced Television. (2025, November 19). Nvidia Q3 eases fears of AI bubble bursting. https://www.advanced-television.com

Wall Street Journal. (2025, November 19). Nvidia profits soar, soothing investor jitters over AI boom. https://www.wsj.com

Business Insider. (2025, November 18). Nvidia earnings recap: Stock pops after revenue beat as CEO bats down 'a lot of talk about an AI bubble'. https://www.businessinsider.com

Investopedia. (2025, November 19). Nvidia earnings live: AI chip giant's results blow past Wall Street expectations. https://www.investopedia.com

Fortune. (2025, November 19). Nvidia blows past revenue targets and forecasts trillions in AI infrastructure spending by end of decade. https://fortune.com

NVIDIA Investor Relations. (2025, November 19). NVIDIA announces financial results for third quarter fiscal 2026. https://investor.nvidia.com

CNBC. (2025, November 19). Nvidia shares rise on stronger-than-expected revenue, forecast. https://www.cnbc.com

Al Jazeera. (2025, November 19). Nvidia forecasts Q4 revenue above estimates despite AI bubble concerns. https://www.aljazeera.com

TechCrunch. (2025, November 19). Nvidia's record $57B revenue and upbeat forecast quiets AI bubble talk. https://techcrunch.com

USA Today. (2025, November 19). What AI bubble? Nvidia stock rises as it beats Wall Street forecast. https://www.usatoday.com

SiliconANGLE. (2025, November 18). What AI bubble? Nvidia crushes expectations again, as revenue jumps 62%. https://siliconangle.com

US News. (2025, November 19). Nvidia's strong forecast calms AI bubble jitters, for now. https://money.usnews.com

CNBC. (2025, November 20). Nvidia stock pops 5% in premarket trading after stronger-than-expected results. https://www.cnbc.com

Bureau of Labor Statistics. (2025, November 19). Employment situation news release. https://www.bls.gov

Dice. (2025, August 19). November 2025 jobs report. https://www.dice.com

ETC Journal. (2025, October 24). Nov. 2025: AI developments in the US job market. https://etcjournal.com

Bureau of Labor Statistics. (2025, March 10). AI impacts in BLS employment projections. https://www.bls.gov

Boston Consulting Group. (2025, October 15). Are you generating value from AI? The widening gap. https://www.bcg.com

LinkedIn / Pranay Mehrotra. (2025, October 2). BCG study: 5% of companies are future-built with AI, outperforming peers. https://www.linkedin.com

Yahoo Finance Canada. (2025, October 9). BCG says only 5% of companies are deriving value from AI. https://ca.finance.yahoo.com

Outsourcing Today. (2025, November 17). Only 5% of companies globally are "future-built" for AI. https://outsourcing-today.ro

Supply & Demand Chain Executive. (2025, September 29). 60% of organizations report minimal revenue, cost gains from AI. https://www.sdcexec.com

BCG Press Release. (2025, September 29). AI leaders outpace laggards with double the revenue growth and triple the cost savings. https://www.bcg.com

Yahoo Finance. (2025, November 17). Peter Thiel's fund sold off entire Nvidia stake last quarter. https://finance.yahoo.com

Reuters. (2025, November 17). Peter Thiel's fund offloaded Nvidia stake in third quarter, filing shows. https://www.reuters.com

Investopedia. (2025, November 18). Peter Thiel's hedge fund dumped Nvidia shares just before its big earnings report. https://www.investopedia.com

Barron's. (2025, November 18). Nvidia stock fell after Peter Thiel dumped his stake. Why 'everyone' is selling. https://www.barrons.com

CNBC. (2025, November 17). Peter Thiel's hedge fund dumps Nvidia stake, cuts back Tesla position. https://www.cnbc.com

Barchart. (2025, November 16). Dear Nvidia stock fans, mark your calendars for November 25. https://www.barchart.com

Yahoo Finance. (2025, November 18). Is Michael Burry shutting his fund just before he's about to be proved right? https://finance.yahoo.com

Business Insider. (2025, November 13). 'Big Short' Michael Burry is the only short-seller retail investors like. https://www.businessinsider.com

Trustnet. (2025, November 18). The funds hit hardest in the 'AI bubble fear' sell-off. https://www.trustnet.com

Reuters. (2025, November 13). Michael Burry of 'Big Short' fame is closing his hedge fund. https://www.reuters.com

FX Guys. (2025, November 10). SoftBank divests Nvidia holdings for $5.8B in AI shift. https://news.fxguys.io

ABC News. (2025, November 10). Japan's SoftBank says it has sold its shares in Nvidia for $5.8 billion. https://abcnews.go.com

Tom's Hardware. (2025, November 10). SoftBank offloads entire $5.83 billion stake in Nvidia. https://www.tomshardware.com

S&P Global. (2025, November 16). Institutions shed nearly $43B in US stocks in October as sell-off persists. https://www.spglobal.com

Dataconomy. (2025, November 10). The why behind SoftBank's Nvidia stake sale. https://dataconomy.com

CNN. (2025, November 18). Why some elite investors are turning on the darling of the AI rally. https://www.cnn.com

Hybrid Horizons. (2025, October 3). The AI boom isn't a bubble. https://hybridhorizons.substack.com

The American Prospect. (2025, November 18). The AI bubble is bigger than you think. https://prospect.org

Fortune. (2025, November 11). OpenAI says it plans to report stunning annual losses through 2028. https://fortune.com

Morningstar. (2025, November 17). OpenAI is AI's leading indicator. Does that make it too big to fail? https://www.morningstar.com

Times of India. (2025, November 18). 'Curse of ChatGPT': How Oracle has lost $300 billion in market value. https://timesofindia.indiatimes.com

Bloomberg. (2025, November 19). Companies are warming up to saying AI is the reason for job cuts. https://www.bloomberg.com

CPA Practice Advisor. (2025, November 18). Companies are warming up to saying AI is the reason for job cuts. https://www.cpapracticeadvisor.com

Intellizence. (2025, November 13). Companies that announced major layoffs and hiring freezes. https://intellizence.com

Interview Query. (2025, November 6). Forget data science: This new AI role is exploding with 800% growth. https://www.interviewquery.com

EMCAP. (2025, November 4). AI models are the gold, forward-deployed engineers are the gold miners. https://www.emcap.com

CMS Wire. (2025, September 15). Opticon 2025: Forward deployed engineers step into the spotlight. https://www.cmswire.com

SSO Network. (2025, September 9). Forward deployed engineer: Turning AI promise into progress? https://www.ssonetwork.com

Final Round AI. (2025, May 29). AI job displacement 2025: Which jobs are at risk? https://www.finalroundai.com

EconoFact. (2025, November 14). Fact check: Has AI already caused some job displacement? https://econofact.org

Exploding Topics. (2024, May 25). 60+ stats on AI replacing jobs (2025). https://explodingtopics.com

World Economic Forum. (2025, August 21). Why AI is replacing some jobs faster than others. https://www.weforum.org

World Economic Forum. (2025, October 2). How we can balance AI overcapacity and talent shortages. https://www.weforum.org

Boston Consulting Group. (2025, October 22). AI is moving faster than your workforce strategy. Are you ready? https://www.bcg.com

Kyndryl. (2025, November 12). Q&A: How people readiness will unlock the promise of agentic AI. https://www.kyndryl.com

Salesforce Engineering. (2025, November 13). Accelerating Agentforce deployments: From 6 months to 3 weeks. https://engineering.salesforce.com

Constellation Research. (2025, October 12). Salesforce makes its Agentforce 360 case to be your AI agent platform. https://www.constellationr.com

StartupHub. (2025, November 19). Salesforce agentic AI redefines enterprise capacity. https://www.startuphub.ai

Forrester. (2025, October 27). Predictions 2026: AI moves from hype to hard hat work. https://www.forrester.com

IDC Blog. (2025, October 21). FutureScape 2026: Moving into the agentic future. https://blogs.idc.com

Prism Media Wire. (2025, July 14). Agentic AI: A strategic forecast and market analysis (2025–2030). https://prismmediawire.com

Deloitte. (2025, November 17). TMT predictions 2026: The gap narrows, but persists. https://www.deloitte.com

SAS. (2025, November 6). SAS predictions: The great AI reality check of 2026. https://www.sas.com

Jay Cadmus. (2025, November 12). 65% of enterprise AI deployments are stalling: The real bottleneck. LinkedIn. https://www.linkedin.com

MIT Sloan Management Review & BCG. (2025, November 19). The emerging agentic enterprise: How leaders must navigate a new age of AI. https://sloanreview.mit.edu

McKinsey & Company. (2025, November 4). The state of AI: Global survey 2025. https://www.mckinsey.com

Reuters. (2025, November 19). Wall Street indexes end rocky session higher; Nvidia gains after hours. https://www.reuters.com

Harvard Business Review. (2025, October 24). Designing a successful agentic AI system. https://hbr.org

WebProNews. (2025, November 12). AI's precarious peak: Bubble fears and market tremors in 2025. https://www.webpronews.com

Reuters. (2025, October 31). The great AI buildout shows no sign of slowing. https://www.reuters.com