Software and IT equipment investment added 134 basis points to Q1 2026 US GDP growth. Total GDP growth that quarter was 2.0%. That means roughly two-thirds of all American economic growth last quarter came from a single thing: companies buying AI.
This is not a bubble talking point. It’s a structural shift that’s already showing up in the numbers, and most people are still arguing about whether AI is “useful.” Meanwhile the Bureau of Economic Analysis is quietly logging the largest sector contribution to US GDP growth in recorded history.
If you want to build wealth in this decade, you cannot afford to misread this moment.
The data tells you everything
There are three numbers worth memorizing.
First: AI capex by the hyperscalers is on pace to clear $400B in 2026 — up from roughly $230B last year. NVIDIA’s data center revenue is growing 50%+ year over year. AMD’s data center segment just grew 57% in a single quarter. This is not a few experiments. It’s the largest infrastructure buildout since the railroad era, compressed into 36 months.
Second: Anthropic’s annualized revenue just crossed $30 billion — passing OpenAI’s $24B for the first time. A year ago Anthropic was at $1B in ARR. That is 30x growth in 15 months, and it happened mostly through enterprise API deals, not consumer subscriptions. Over a thousand companies now spend more than $1M a year on Claude.
Third: JPMorgan reclassified its AI spend as core infrastructure with a $19.8B technology budget for 2026 — projecting $2.5B per year in value generation. When the most conservative bank in the country puts AI in the same line item as plumbing and security, the conversation is over.
What “67% of GDP growth” actually means
Most people will read that headline as a stock pick. It’s not. It’s a regime change.
The companies that build the infrastructure (NVIDIA, AMD, Broadcom, the hyperscalers, the data-center REITs, the power and grid companies) are obviously beneficiaries. But the more interesting wealth angle is the second-order effect.
If two-thirds of GDP growth is now AI-related, then for the next several years, owning the productive layer of the economy is going to look very different than it has for the last 30. The companies that use AI to compress costs by 20-40% will out-earn the ones that don’t, regardless of what sector they’re in. JPMorgan’s $2.5B annual AI value generation is the early signal. Walmart, UnitedHealth, the big oils — all of them are next.
This is also why the Fed cannot cut rates the way the equity market wants it to. Oil is at $105 because of the Iran war. Inflation is forecast at 4.2% for 2026. But underneath that, productivity is exploding because of AI. A central bank cutting rates into a productivity boom is exactly how you get an asset bubble. Powell knows it. Warsh knows it. Rates stay where they are.
The contrarian read most people miss
Here is the part of this story that the financial media has not figured out yet.
The biggest wealth concentration of the next ten years will not happen at the model layer. OpenAI and Anthropic will both be huge — likely IPO-bound (Anthropic potentially as early as October). But they’re locked in a brutal capex war that requires literal nation-state amounts of capital. NVIDIA committing $100B to OpenAI is not normal. It’s a sign that compute is the binding constraint, and whoever has it determines who wins.
The wealth, in other words, is migrating down the stack — to the picks and shovels.
The same pattern played out in every prior infrastructure cycle. In the 1990s, the companies that made the most money during the internet build-out were not Pets.com or Webvan. They were Cisco, Sun Microsystems, and the fiber companies that laid the pipes. The application-layer winners (Google, Amazon, Facebook) emerged later, after the infrastructure was already in place.
We’re in the equivalent of 1998 right now. The fiber is being laid. The chips are being shipped. The data centers are being built faster than any other industrial buildout in human history. The wealth that’s getting created at this layer of the stack will dwarf the wealth that gets created at the application layer for at least another 3-5 years.
What this means for your portfolio
This is not investment advice — that’s not what Wealtharian does. But here’s the framework worth thinking about.
Three buckets are worth attention:
1. The picks and shovels. NVIDIA gets all the headlines but the real story is broader: AMD’s data center business is now growing faster than NVIDIA’s. Broadcom’s custom silicon is becoming the alternative most hyperscalers want. Power and grid companies — Vistra, Constellation, the nuclear restart names — are tied to the same demand curve. If you want exposure with less single-name risk, an AI infrastructure ETF gets you all of it.
2. The “AI-pilled” incumbents. The next layer of wealth creation will come from old-economy companies that genuinely deploy AI to expand margins. Watch which companies are quietly hiring AI engineering teams and showing margin expansion in the next four quarterly reports. JPMorgan, Costco, and a handful of insurers are leading. Most of the S&P 500 is still in the experimentation phase. The gap between leaders and laggards will widen fast.
3. Your own income. This is the bucket most people ignore. If 67% of GDP growth is AI-driven, the income you generate from your own work — your career, your side hustles, your business — is also subject to the same dynamic. Learning to use AI tools to compress your own work by 30-50% is, in raw return-on-time terms, the single highest-leverage thing most people can do this year. It is the part of your wealth equation that compounds the fastest.
The bottom line
A single sector contributed two-thirds of last quarter’s GDP growth. That has never happened before. The Fed cannot cut into it. Inflation cannot kill it (it survives inflation, because the productivity gains exceed the input cost). Geopolitics cannot stop it (the war in Iran is fueling oil, but AI capex is being funded out of cash flow, not credit).
This is the wealth-building decade. Most people will spend it arguing about whether AI is real on Reddit. The ones who understand the chart understand they’re early in something that does not have a historical analog.
Position accordingly.
Track your own progress
Want to track your own path to financial independence — and see exactly how every paycheck and investment is moving you toward FU money? The Wealtharian Wealth Tracker lets you monitor your net worth, FU money progress, and investment milestones in one place. Try it free →