This week, Cloudflare laid off about 1,100 people â roughly 20% of its workforce. The detail buried in the announcement is the one you should not ignore: internal AI usage at the company jumped more than 600% in the last three months.
That number is the story.
It is not just a tech sector story. It is not even just a labor story. It is the clearest signal yet that the corporate operating model is being rewritten in real time, and the wealth created by that rewrite is concentrating into a very small number of hands. If you understand how it works, you can position to be on the right side of it. If you do not, you will be the side.
The numbers no one is putting together
Look at this week’s headlines as one picture, not separate stories:
- Cloudflare: 1,100 cut (~20%)
- BILL: up to 30% of headcount
- Upwork: ~25% of staff
- Coinbase: ~700 jobs (~14%)
- Ticketmaster: ~350 jobs (8%)
- Meta: 8,000 cuts effective May 20
That is over 11,000 jobs in a single news week, on top of the 128,000+ tech jobs already gone in 2026. Meanwhile Microsoft just raised 2026 capex guidance to $190 billion â well above the $154 billion Wall Street had penciled in. Alphabet, Meta, and Amazon together with Microsoft are on track for roughly $700 billion of AI infrastructure spend this year.
Money is not vanishing. It is moving. Out of payroll lines, into capex budgets â and then into the income statements of a very narrow group of companies on the supply side: NVIDIA, AMD, Broadcom, the hyperscalers themselves, the power producers that feed the data centers, and a small handful of AI-native software companies that benefit when their customers cut humans and buy more software.
This is one of the largest, fastest peacetime reallocations of corporate spending in modern history. It is being framed as a labor story. It is actually a wealth story.
What’s actually happening inside companies
The Cloudflare data point is the tell. A 600% increase in internal AI usage in 90 days does not mean engineers wrote a few more prompts. It means the work itself is being restructured. Support tickets, code reviews, design iterations, security analysis, sales outreach, financial reporting â each of these used to be a queue of humans. Now a meaningful share of each queue runs through AI first, with humans reviewing only the edge cases.
Once you cross that threshold, the math changes permanently. You do not need the same number of people to ship the same output. So you do not hire them back when growth returns. The job did not pause. The job got eaten.
This is exactly what every recent earnings call from Big Tech has been signaling, but companies have been careful not to say the quiet part out loud. Cloudflare’s CEO essentially did. Expect more of it.
Where the wealth is flowing
If you want to track where the money is actually going, follow three flows.
First, the picks-and-shovels. Every dollar a company saves on labor and reinvests in AI capacity ultimately ends up in chips, networking, power, and cloud services. NVIDIA reports earnings May 20, with guidance pointing to $78 billion in a single quarter. Bank of America just lifted its price target to $320. AMD’s last earnings sailed past estimates on data center growth, and Wall Street is no longer treating “AI chips” as a one-name story. The supply chain widens â and so does the universe of stocks that benefit. For the breadth of the buildout, the AI capex story we covered earlier this week goes into the $725B wave in detail.
Second, the productivity beneficiaries. Software companies whose product replaces five seats with one â CRM with embedded agents, customer support automation, finance and accounting automation, security operations â see expansion pricing while their customers shrink. The same SaaS contract, with the same vendor, can grow even as the customer’s headcount shrinks. That is a margin story, not a top-line story, and the market is still learning to price it.
Third, the energy and infrastructure underneath. AI training and inference are absurdly power-hungry. The unsexy compounders here â independent power producers, transmission utilities, data center REITs, cooling and electrical equipment makers â are quietly having one of their best multi-year runs. We covered this angle in our Fed pivot piece earlier this week, because the energy demand story is one of the few things that can keep inflation sticky enough to delay rate cuts.
How to be on the right side of this
The good news: you do not need to be a venture capitalist or sit on a tech board to participate. You need to make three decisions.
1. Your portfolio. If you own a broad index fund, you already have exposure to the winners â but you are also fully exposed to the losers. Consider whether your equity mix actually reflects the world being built. That can mean a small overweight to the AI infrastructure complex (chips + power + data centers), not a YOLO into a single name. Diversification still applies. The point is intention.
2. Your income. The single most under-discussed risk on most people’s balance sheet is concentration in one paycheck doing one job at one employer in an industry being actively automated. If your job is mostly producing things AI can already do at 80% quality â first-draft writing, basic analysis, routine code, scheduling, simple design â your runway is shorter than your performance review suggests. The fix is not panic. The fix is building a second income line that uses AI as a tool, not a competitor. Freelance work, a small product, a paid newsletter, consulting in your domain â the leverage AI gives a motivated individual right now is unprecedented. One person with the right tools can do what five did in 2023.
3. Your skills. The premium is no longer on doing the task â it is on knowing what to ask, validating what comes back, and putting AI output into something a human or business will actually pay for. Judgement, taste, distribution, customer relationships, domain expertise â these are the things that compound. Spend twice as much time on them as you do on technical AI skills and you will be fine.
The honest truth no one wants to say
This transition is going to be uneven and, for a lot of people, deeply unfair. Mid-career knowledge workers are getting hit first because their work is structured, document-heavy, and trainable. The companies doing the cutting are the same ones whose stock prices are setting records. Both things are true at once. Pretending otherwise will not protect anyone.
What this brand believes â and you will find it in everything we write â is that the right response to a wealth reshuffle of this size is not to be angry about it. It is to understand it earlier than most people, position your capital and your time accordingly, and pull the people around you along with you. Wealth created by genuine productivity is good. Wealth created while wages stagnate without anyone noticing the connection is a problem. Both can be true. Wealtharian readers should plan to be on the productive side of that line.
What to watch next week
Three things on the calendar:
- May 20 â NVIDIA Q1’27 earnings. Guidance is the tell. Anything above $80 billion and the AI capex narrative gets another leg. Below $75 billion and the air starts to come out.
- May 20 â Meta layoffs go live. Watch the productivity commentary on the next earnings call. They are explicitly recutting around AI.
- Next CPI print. If inflation stays above 3.5%, the rate cut narrative for 2026 is done â and high-quality dividend payers in energy and utilities become the year’s other story.
The AI labor reshuffle is not coming. It is here. Cloudflare just published the receipts.
Want to track your own path through the reshuffle? The Wealtharian Wealth Tracker lets you monitor your net worth, your FU money number, and the income streams you are building outside your day job. The point of all this work is freedom â start measuring it.