Google asked the market for $80 billion. The market said yes but the stock fell anyway

Google asked for $80 billion. The market offered more, and the stock fell anyway. That is the part nobody can explain, and it is the part that matters.

Start with the question everyone skipped: why does the most profitable company on earth need to raise equity at all? Alphabet earns about $174 billion in cash every year, more than the entire economy of most countries on the planet, and keeps $120 billion in the bank. A business like that funds itself. It does not pass the hat.

Except the math finally broke.

• Alphabet's AI build-out will cost between $180 billion and $190 billion in 2026 alone, and the CFO has already said 2027 will be significantly higher.
• Bloomberg's Mandeep Singh estimates that 2027 figure could climb all the way to $300 billion.

Once your spending crosses your entire annual cash flow, you have left the world where a company pays for its own growth.

The raise is not the news. Crossing that line is.

Why it could not simply borrow

There are two ways to fund a big bet, and the one you choose is a confession about how confident you are.

• Debt is what you pick when you know the bet will pay off. You borrow, you build, the profits clear the loan, and your existing shareholders keep all the upside for themselves.
• Equity is the opposite. You hand outsiders a permanent slice of your future. A confident company avoids that, unless it cannot borrow any more, or it quietly wants company on the downside.

Ben Thompson of Stratechery made this point sharply on the morning of the raise, and the facts back him. Alphabet had already taken on roughly **$85 billion of debt **across six currencies in twelve months, pushing total borrowings past $100 billion. The debt window was not shut, but the terms were turning against it.

So equity was not the clever move. It was the next available one. Hold that thought, because it becomes a test you can watch.

What Berkshire's $10 billion actually says

Now the detail that should make you lean in. Berkshire Hathaway, the firm that built its reputation on refusing capital-hungry technology bets, anchored this raise with $10 billion.

This is not a dabble. Look at how fast the conviction built:

• In Q3 2025, Berkshire made its first-ever Alphabet purchase, worth about $4.3 billion.
• By early 2026, under Greg Abel, it had more than tripled that stake to around $15.6 billion.
• After this placement, Berkshire's total Alphabet holding now tops $26 billion, making it one of its largest positions anywhere.

For a house that spent decades saying no to businesses that burn cash to grow, writing that cheque into a dilutive raise is a loud vote of belief.

But hold two ideas at once. Berkshire's conviction is real, and the market sold off regardless.

When the most disciplined investor in the world says yes and the price still falls, the crowd is worried about something the headline is not saying.

The thing nobody wants to say plainly

A big part of the AI boom may just be the same money going in circles.

Here is how the loop works. A tech giant puts billions into an AI startup. The startup spends that money buying chips and cloud, often from the very giant that funded it. The cash comes back home, but along the way it gets counted as revenue.

A Seaport Global analyst called it a parent co-signing their own child's mortgage, pointing to Nvidia's $100 billion bet on OpenAI: the chipmaker funding the customer who then buys its chips.

There is a calmer reading too. The Swiss Transparent Portfolio notes Alphabet makes about $130 billion in normalised profit a year, second only to Nvidia. By that view this is not weakness, it is ambition that simply outgrew even a giant's cash flow. And the demand is real: Gemini has 900 million monthly users, and Google Cloud's $460 billion order backlog is money customers have already committed.

Both can be true at once. The question that matters is simple:

How much of this AI revenue comes from real, outside customers paying for real productivity, and how much is just the same few companies passing money to each other?

If the outside customers do not show up at scale, the growth will shrink as fast as it appeared.

What to actually watch over the next eighteen months

1. Does Alphabet raise serious new debt after this? If it does, management is signalling that the returns are real and close, and the equity was just a bridge to get there. If debt stays quiet while capex climbs toward $300 billion, then the reading that Google wanted to share the risk because it is unsure of the payoff becomes far more convincing.

2. Does that $460 billion cloud backlog turn into paid invoices from genuinely independent companies, rather than from AI firms funded by the same circle of investors? Sundar Pichai keeps saying his constraint is compute, not customers, and this raise is his answer to that queue.

So don't watch how much Google raised. Watch who actually pays the bills.