The Buyback Is Dead. Long Live the Dilution.
The Buyback Is Dead. Long Live the Dilution.
A stream of consciousness from somewhere on the buy side, Tuesday afternoon, June 2.
Okay. So Alphabet just announced it's raising $80 billion in equity.
Sit with that for a second.
Eighty. Billion. Dollars. In stock. From a company that has spent — spent, burned, vaporized — $346 billion buying back its own shares since 2016. A decade of telling shareholders that the best use of capital is shrinking the float. A decade of the implicit promise that Google's cash machine was so stupidly productive, so grotesquely self-sustaining, that the correct move was always to return it rather than deploy it. That was the deal. That was the entire GOOG thesis for an entire generation of long-only holders.
The deal is off.
GOOGL opened down 3.5% Tuesday morning and spent the session bleeding. Intraday low of $366. The stock hasn't seen $366 since March. By close it had clawed back to $367.94, but that's not a recovery — that's exhaustion. The sellers got tired, not converted.
Here's the structure of the thing, because it matters: $15 billion in a straight public stock offering, $15 billion in mandatory convertible preferred, $10 billion in a private placement to Berkshire Hathaway, and then a further $40 billion ATM program launching in Q3. The ATM alone is a slow-drip overhang that will sit above the stock for months, every trading session a potential supply event, every rally a potential flush point for the company's own distribution desk. That's not capital raising. That's capital rationing, served cold, over time.
And Berkshire showed up for it. Greg Abel's first real move as the man running Buffett's money is to triple down on Alphabet. He's not wrong — Google Cloud is doing $20 billion per quarter in revenue and growing at 63%. The $460 billion contract backlog is a real number. But Berkshire committing $10 billion anchors the deal at optics, not valuation. You don't anchor an $80 billion offering with $10 billion and call it an endorsement. You call it a floor.
The number that keeps pulling my attention back is $185 billion.
That's the midpoint of Alphabet's 2026 capex guidance. Compare it to $91 billion last year, which was already a number that caused a selloff in February when they first disclosed it. They've essentially doubled their capital spending in twelve months, and they're financing the incremental difference by diluting shareholders. The AI infrastructure arms race has now consumed the buyback program at the world's most cash-generative advertising business. Let that land.
Microsoft is at $190 billion capex for the year. Amazon at $200 billion. Meta at $135 billion. You could add those numbers up and lose track of the scale of what is being poured into data centers, custom silicon, fiber, cooling infrastructure, and power contracts right now. Combined, the four hyperscalers are spending somewhere north of $600 billion on compute this year. Every dollar of that is a bet that AI demand will not only continue but accelerate. Every dollar spent is a future depreciation charge. Every future depreciation charge is pressure on margins. Every pressure on margins is a future earnings conversation that nobody on the buy side has fully modeled yet.
The AI infrastructure buildout has now crossed from "investment cycle" into "structural obligation." Nobody can stop. Stopping is conceding the race. So everyone keeps spending, and the capital has to come from somewhere.
For Alphabet, it comes from the equity market. For Microsoft, it comes from the bond market and retained earnings. For Amazon, it comes from AWS cash flow and debt. Each of these is a different bet, with different risk profiles, and the market has been treating them as interchangeable. They aren't.
Meanwhile, Kevin Warsh got sworn in as Fed Chair eleven days ago — White House ceremony, first since Greenspan in '87, very much the whole production — and his first FOMC meeting is June 17. Two weeks. The Fed funds rate is sitting at 3.50–3.75%, and the market has priced a 97% probability of no move in June. That's fine. It's probably right.
But the dot plot lands alongside the decision. And Warsh's first press conference. And the FOMC that, as recently as April, produced four dissents in a single meeting — the most fractious policy committee in thirty years. Trump put Warsh in that chair expecting cuts. Warsh told the Senate, publicly and on the record, that Trump never asked him to predetermine any rate decision and that he wouldn't have agreed to it if he had. That answer had the texture of a man protecting himself in advance.
Core PCE is at 3.3%. CPI at 3.8%. Brent crude is back above $97 after Iran suspended talks again Monday. The Atlanta Fed's model has Q2 GDP tracking above 4%. There is no macro argument for a rate cut right now. There is barely an argument for easing bias language. What Warsh inherits is a central bank that needs to signal tightening optionality without spooking equity markets that are up 11% year-to-date on the back of AI euphoria that will crack the moment the cost of capital becomes a genuine conversation again.
That is a very narrow needle to thread at your first press conference. With cameras. With Trump watching Truth Social.
The market is pricing in one or two hikes by March 2027. Not imminent. Not this meeting. But the direction of travel has quietly, unmistakably, shifted. And nobody on the equity side has fully repriced for a world where the risk-free rate moves up rather than down.
Back to Alphabet for a moment, because there's a detail in this deal that deserves more attention than it's getting.
Approximately $30 billion of the ATM program is earmarked to cover 2026 tax obligations on employee equity awards. Three hundred billion dollars in buybacks over a decade, and now you're issuing stock partly to pay the IRS bill on your own compensation structure. There's a sentence in there about late-cycle excess that someone smarter than me should write.
The AI race is real. The demand for compute is real. Google Cloud's 63% growth rate is real. None of that is in dispute. But $80 billion in equity — the largest fundraise in Alphabet's history, the largest strategic reversal of its capital allocation philosophy since the company went public — landed on a Tuesday, was absorbed by a market that barely flinched at the headline, and closed down 2.3%.
In a normal market, a dilutive equity raise of this size sends a stock down 8%. In a normal market, the reversal of a decade-long buyback commitment is a significant event. In this market, GOOGL's intraday chart looked like mild weather. The S&P 500 closed at a new record high. XLK finished green.
When gravity stops applying to individual names in a sector, it's usually because the whole sector has floated free of it. That's fine until something brings it back. It always comes back. The question is only what pulls it there — the Fed, the oil price, the margin data, or something nobody has named yet.
Right now I don't know the answer. What I know is that the single largest piece of financial news on Tuesday was a $4.5 trillion company abandoning its capital return program to finance a spending race with three other companies of comparable scale, and the market traded it like a footnote.
That's either a sign of deep fundamental conviction or a sign that nobody is paying attention to the right things.
I have my view on which one it is. The chart has its own.
Nonfarm payrolls Friday. Warsh's first dot plot June 17. Somebody is going to blink.