For a week, this market celebrated the winners of the AI wave. Dell jumped a third, Anthropic was valued at $965 billion, Micron cracked the trillion, Nvidia declared war on the PC market. It was a party. And last night, on June 1, the bill arrived at the table — handed over, of all people, by the richest guest in the room.
Alphabet, the parent company of Google, announced it would raise $80 billion through the sale of new shares. It’s the company’s first stock issuance since 2005 — since the IPO over two decades ago. The purpose: artificial intelligence. And that very fact raised a question this morning that the market had until now been suppressing: if even Google needs fresh money to fund AI — then who can actually afford it?
What Alphabet actually announced
The $80 billion consists of three parts, and the breakdown is revealing:
- $30 billion through classic, bank-underwritten share offerings — $15 billion in mandatory convertible preferred stock and $15 billion in regular Class A and Class C shares
- $40 billion through a so-called at-the-market program running from the third quarter of 2026 — shares sold gradually at the prevailing market price. Notably: per the filing, these $40 billion serve primarily to cover tax obligations from employee stock vesting, not directly the data-center build
- $10 billion as a private placement to a single buyer: Berkshire Hathaway
The use of the main tranche is unambiguous: investments in AI infrastructure and global compute. Alphabet states openly in the filing that customer demand for AI compute capacity exceeds current supply — they can’t build fast enough.
The historic dimension: for the first time since 2005
Why is this so significant? Because Alphabet owns perhaps the most profitable business model in the world. Google Search, YouTube, the cloud — these are money-printing machines with enormous margins and free cash flow. A company like this normally doesn’t sell new shares, because it simply generates enough money of its own. The last issuance was in 2005, shortly after the IPO.
That Alphabet now, after 21 years, returns to the stock market anyway says more about the scale of the AI investments than any headline before. The company has announced it will spend between $180 and $190 billion on capital expenditures in 2026 alone — and signaled a significant increase for 2027. For context: in 2024, capex was still around $52 billion, in 2025 around $91 billion. The curve is going vertical.
The Buffett angle: why is caution incarnate buying?
Perhaps the most fascinating part of the announcement is Berkshire Hathaway’s $10 billion stake. Warren Buffett, who has been sitting on a record cash pile of around $380 billion for months and is known for his skepticism toward tech hype, is buying in here — and at a discount: $5 billion in Class A shares at $351.81 and $5 billion in Class C shares at $348.20.
This isn’t an accidental purchase. Berkshire has been building its Alphabet position since the third quarter of 2025. So the man who avoids the most expensive AI bets is betting, of all things, on the company building the infrastructure for it — and at a negotiated preferential price. You can read this two ways: either Buffett believes in Alphabet’s long-term cash-flow strength independent of the AI boom. Or he’s securing a cheap entry into the one AI heavyweight he considers soundly valued. Probably both.
The real story: the contagion
The stock market this morning reacted not only to Alphabet, but to the signal behind it. Alphabet itself fell about 2.6% premarket — dilution from new shares weighs on the price short-term. That’s normal.
More interesting is the contagion. Oracle, which had gained nearly 10% the day before, fell 4.6% premarket — and expressly not because of bad numbers of its own, but out of concern over the rising costs of building AI infrastructure. The Alphabet announcement suddenly showed investors what this build-out really costs, and they’re now asking of every AI heavyweight: is the company’s own cash flow enough, or is a capital raise coming here too?
One market observer captured the new mood: if perhaps the best business model in the history of capitalism — measured by scale, growth, margins, and cash flow — can’t shoulder the AI spending on its own, then who can? That very question has hung over the entire sector since today.
What this means mathematically
Let’s calculate soberly. Alphabet plans $180 to $190 billion in capex for 2026. The company generates enormous free cash flow, but evidently not enough to carry this sum plus dividends, buybacks, and operating costs alone without straining the balance sheet. Hence the issuance.
For the broader market, the decisive number isn’t the $80 billion, but the ratio: capex growing faster than revenue. When the capex curve rises more steeply than the earnings curve, then at some point either AI demand must justify these investments — or returns on invested capital begin to decline. That’s exactly the concern that dragged Oracle down today. The AI wave has a revenue side (which we celebrated last week) and a spending side (which became visible today). So far, the market has looked almost exclusively at the first.
Three scenarios
Scenario 1 — Demand justifies the spending (~45%): Alphabet’s bet pays off, AI cloud demand grows fast enough to turn $180+ billion in capex into future earnings. The issuance was forward-looking, not desperate. Buffett’s entry proves wise.
Scenario 2 — Healthy repricing (~35%): The market begins to value AI stocks more discerningly — winners with real earnings get separated from pure capex burners. A rotation, not a crisis. Those with cash flow win; those who only spend suffer.
Scenario 3 — Capex fear spreads (~20%): Concern over AI costs widens, several heavyweights announce similar capital measures, and the market begins to doubt whether the returns will ever come. Correction across the whole AI chain. Dimon’s warning of an “exuberant” market comes true.
What smart money is doing
Here it gets interesting, because the signals diverge. On one side, Buffett — caution incarnate — is buying into Alphabet. On the other, JPMorgan chief Jamie Dimon warned just on May 29 at the Reagan National Economic Forum expressly of an “exuberant” market and that risks were underpriced. The same Dimon hoards his own money in money-market funds.
This isn’t a contradiction but precision: smart money separates. Buffett selectively buys the one company whose cash flow he considers robust enough to survive the AI investments — at a negotiated discount. Dimon simultaneously warns of the broad market that celebrates every AI story indiscriminately. The lesson: in this phase, it’s no longer “AI yes or no” that decides, but which company can ultimately pay the bill.
What investors should concretely do
- Look at the cash-flow side, not just growth: The decisive question for every AI stock from today on: can the company fund its investments on its own, or does it need fresh capital? The latter dilutes you as a shareholder.
- Understand dilution: When a company issues new shares, your percentage stake in the profit shrinks. With Alphabet, that’s bearable due to sheer size — with smaller AI names it can be painful.
- Heed the Buffett indicator, but don’t blindly copy it: That Berkshire is buying into Alphabet is a quality signal. But Buffett buys with a negotiated discount and a 30-year horizon — terms you don’t have as a retail investor.
- Separate the two sides of the AI wave: Last week was about the revenue (Nvidia, Anthropic, Dell). This week about the spending (Alphabet, Oracle). A healthy portfolio understands both sides, instead of seeing only the good half.
- Plan for taxes: Gains and dividends from U.S. stocks like Alphabet are subject to Austrian 27.5% capital gains tax plus possible U.S. withholding. Calculate net.
The honest bottom line
June 2, 2026, was no crash and no shock — it was a reality check. For a week, the market celebrated the AI revolution as a pure growth story. Today it became visible that every revolution also has a bill, and that this bill is large enough even for the richest corporations in the world to return to the stock market.
That needn’t be a bad thing. Alphabet is investing because demand is real — the company can’t build fast enough, which is a luxury problem. And that Buffett is buying in is anything but an alarm signal. But the tone changed today. The market no longer asks only “how big will AI get?” but also “who can afford the path there — and at what price to shareholders?” That’s a more grown-up question. And more grown-up questions are often the beginning of healthier markets — even if they feel more uncomfortable in the short term than a party.
Try TradingView Free for 30 Days
Plus get a $15 discount on your first subscription through this link.
