Anthropic just did something almost no one in software history has ever done. The maker of Claude went from a roughly $375 million revenue run rate at the start of 2025 to a $30 billion run rate this spring. That is 80x growth in a single year. By the time you finish reading this paragraph, Anthropic is collecting more revenue than most public companies make in a quarter.

This is not a fluke. It is the loudest signal yet that the AI boom is past the “interesting demo” stage and well into the “we are now rewiring the entire economy” stage. And whether you have $500 in a brokerage account or $5 million, the way you respond over the next 24 months will decide which side of the wealth divide you end up on.

What actually happened

A quick recap of the last two weeks in AI, because the headlines have been moving faster than most people can read them:

OpenAI shipped GPT-5.5, took its annualised revenue past $25 billion, and is openly preparing for a public listing as early as late 2026.
Anthropic confirmed the 80x growth number and a near-trillion-dollar valuation, with Claude Code alone hitting $1 billion in annualised revenue inside six months of launch.
Elon Musk’s xAI launched Grok Build, a coding agent that runs eight parallel sub-agents at once, throwing itself directly into the same enterprise budget that Anthropic and OpenAI are eating.
NVIDIA reports earnings on May 20. Guidance above $80 billion confirms the AI capex thesis. Below $75 billion and half the market gets repriced overnight.

This is what the early innings of a real industrial buildout look like. The dot-com era took five years to compound to this level of revenue. AI did it in eighteen months.

Why this matters for your money

The instinct most people will have is to read these numbers, feel late, and do nothing. That is the wrong move. Wealth is almost never built by being early to a story. It is built by understanding the second-order effects while everyone is still arguing about the first.

Three of those second-order effects matter right now.

One: AI infrastructure is the picks-and-shovels trade of this decade. Anthropic, OpenAI, and xAI are not running on vibes. They are running on chips, data centers, networking, and electricity. NVIDIA committed $30 billion to OpenAI alone. Hyperscalers are pouring something like $400 billion in combined capex into AI infrastructure this year. If you cannot pick the winning model, you can still own the road every model has to drive on.

Two: software margins are about to get a lot weirder. When Claude Code can write code at a fraction of the cost of a junior developer, the question is not whether engineering teams shrink. It is what happens to the gross margins of any company that successfully internalises this. The best-run software businesses of 2028 will look more profitable than anything we have seen since the early days of Microsoft. That is a setup, not a threat.

Three: a new class of solo operator is forming. Anthropic itself is the proof. They built a $30 billion run rate with a few hundred employees. The same leverage is now available, in miniature, to anyone willing to learn the tools. AI-native one-person businesses with seven-figure revenue are no longer the exception. They are the new normal for the next five years.

The contrarian read

Here is the unpopular take. Everyone is treating this as a story about which AI company wins. That is the wrong question. The more interesting question is: who in the broader economy gets rich because all of them succeed at once?

The answer, repeatedly through history, is the picks-and-shovels providers, the people running adjacent infrastructure (energy, real estate around data centers, cooling, security), and the early adopters who use the technology to compress their own cost base. The companies in the middle, the ones that still need ten employees to do what one Claude-augmented operator can do, get squeezed.

If you are an individual building wealth, three positions worth thinking about:

A core allocation to the obvious infrastructure names (NVIDIA, the hyperscalers, the power and grid plays). This is the boring trade that pays.
A learning budget. Spend an hour a day for the next ninety days getting fluent with the top two AI models. The ROI on this beats any side hustle.
A career hedge. If your job description has not changed in the last twelve months and your employer has not started building AI into the workflow, that is a flashing yellow light. Move toward roles where AI is a force multiplier for you, not a replacement.

What we are watching this week

The single biggest event on the calendar is NVIDIA’s earnings on May 20. The print itself matters less than the guidance. Above $80 billion in forward-quarter revenue and the entire AI infrastructure thesis stays intact. Below $75 billion and the market starts asking very different questions about whether the capex cycle has overshot.

Either way, this is the kind of week where being engaged with the market beats being passive. The next four trading days will produce more clarity about the trajectory of AI wealth creation than the previous six months combined.

Bottom line

The story of the next decade is being written right now, in real time, on quarterly earnings calls and in the API revenue lines of three or four companies. Anthropic’s 80x year is not the end of the AI surprise. It is the warning that the rest of the surprises are coming faster than anyone is positioned for.

Wealth in the next decade will be built by the people who treat AI the way the smartest investors treated the internet in 1998 — not as a question of belief but as a question of allocation.

Related Wealtharian reading:

Want to track your own path to financial independence while the AI economy reshuffles everything? The Wealtharian Wealth Tracker lets you monitor your net worth, FU money progress, and investment milestones in one place. Try it free →