Alphabet's $80 Billion Equity Raise: Free Cash Flow Fell 82% in Six Months

Alphabet stopped buying back its own stock. In the six months to 30 June 2026 it repurchased $0 of shares, against $28.3 billion in the same period a year earlier. That is the missing context for the $80 billion Alphabet equity raise announced on 1 June. Free cash flow across those six months came to $4.3 billion, down 82% year on year.

The short answer

Alphabet announced an $80 billion equity raise on 1 June 2026 to fund AI infrastructure. Its own filings show why. In the six months to 30 June, operating cash flow of $84.9 billion was almost entirely consumed by $80.6 billion of capital spending, leaving $4.3 billion.

What Alphabet actually sold, and to whom

On 1 June 2026 Alphabet announced a proposed $80 billion equity capital raise to expand AI infrastructure and compute. The total was later upsized to $84.75 billion.

The structure tells you more than the headline number. Each piece was priced by a different kind of buyer.

The four pieces

The offering documents split the raise into four instruments rather than one large share sale. That spreads supply across markets that do not compete.

Sizes are from the company's free writing prospectus of 1 June 2026. Underwriters were granted a 30 day option over a further $2.25 billion of common stock and a 13 day option over a further $2.25 billion of depositary shares. Goldman Sachs, J.P. Morgan and Morgan Stanley were joint book-running managers.

Berkshire Hathaway's $10 billion

The private placement went to an affiliate of Berkshire Hathaway below the public offering price. Read it as an anchor order, not an endorsement of AI capital spending.

My read is that the Berkshire tranche was a large, cheap position in a cash-generative business, not a bet on data centre returns. Those are different theses that share a ticker.

The filing line that explains the raise

Alphabet did not have to explain itself. The quarterly filing for the period ended 30 June 2026 does it in two rows of the cash flow statement.

Free cash flow is not a line in the filing. It is operating cash flow less purchases of property and equipment, the standard definition used in our earlier analysis of hyperscaler capex. Every other figure comes directly from the consolidated statements of cash flows.

Operating cash flow grew 33%. Capital spending grew 103%.

Alphabet's business got substantially better over the year, and it still ran out of internal funding.

Revenue for the six months rose 23% to $229.7 billion and operating cash flow rose 33%. Neither is a company in trouble.

Capital spending over the same period rose from $39.6 billion to $80.6 billion. When the denominator grows 33% and the numerator grows 103%, the remainder disappears. It took one year.

Dividends alone now exceed free cash flow

Alphabet paid $5.2 billion in dividends in the first half of 2026 and generated $4.3 billion of free cash flow.

The dividend is no longer covered by cash left after the build. That is not a solvency problem for a company holding $242.5 billion in cash and marketable securities. It is a signal about where the marginal dollar comes from.

Buybacks went to zero, and that is the cleanest signal

Every commentator covered the dilution. Almost nobody covered the line above it.

Alphabet's cash flow statement records no repurchases of stock in the first half of 2026, against $28.3 billion in the first half of 2025. A buyback programme of that size does not pause by accident.

The usual defence of the raise is that ongoing repurchases offset the new shares within about a year. That defence requires a repurchase programme currently running. In the first half of 2026, there was none.

I would treat the buyback line as the single most honest number in this story: guidance gets revised, a completed repurchase does not.

Reported profit tells the opposite story, and it is not lying

Anyone reading the earnings headline saw a different company. Net income for the six months was $174.8 billion, against $62.7 billion a year earlier.

That is a 179% increase in reported profit alongside an 82% fall in free cash flow. Both come from the same document. Neither is wrong.

A $98 billion gain that never touched the bank account

Alphabet's Q2 results disclose that other income reflected a net gain of $98.0 billion, primarily net unrealised gains on equity securities. That is a mark to market on holdings, not cash received.

Strip it out and the picture inverts. Reported net income for the half was roughly 41 times free cash flow. A year earlier the multiple was under three.

Earnings can be flattered by an asset revaluation. A data centre cannot be paid for with one. That gap is covered in the comparison of dot-com and AI era metrics, and it is the reason profitability alone settles nothing.

Why equity, when Alphabet could simply borrow

Alphabet did borrow. It raised $56.2 billion of debt in the same six months, more than in the whole of the prior year period. That included a $31.51 billion global bond sale in February 2026 with a rare 100 year tranche.

So the question is not debt against equity. Alphabet took both. The question is why it needed the second.

What a mandatory convertible actually is

A mandatory convertible preferred share pays a fixed dividend, then converts into common stock on a set date whether the holder wants it or not. Alphabet's carries a 6.25% rate and converts around 15 May 2029.

It sits between debt and equity. It does not count fully as borrowing for rating purposes, and delays the share count increase by roughly three years.

The rate is the tell. A company holding $242.5 billion in cash and marketable securities agreed to pay 6.25% on $15 billion of it. That is not the price of cheap money. It is the price of speed and balance sheet treatment.

The instrument choice matters more than the size. Splitting the raise across four instruments, one priced well above prevailing investment grade coupons, reads as a company solving for speed rather than cost.

The funding shift is industry-wide, not an Alphabet problem

If Alphabet were alone, this would be a company story. It is not.

Hyperscaler capital spending in 2026 is on pace to consume close to 100% of operating cash flows, against a 10 year average of 40%, according to UBS estimates reported in June 2026. Alphabet's own 95% for the first half confirms that range, calculated from a filing rather than a forecast.

The bond market absorbed the first wave. The five largest hyperscalers issued $121 billion of US corporate bonds in 2025, against a 2020 to 2024 average near $28 billion a year. Amazon alone raised about $37 billion across 11 tranches on 10 March 2026, followed by a 14.5 billion euro sale the next day.

Morgan Stanley forecasts global AI-related debt issuance near $570 billion in 2026, more than double 2025, with roughly $236 billion already issued by the end of May. BofA Global Research raised its hyperscaler debt forecast to $175 billion from $140 billion in March.

Alphabet's raise is what happens when that first wave is not enough. The debt window is crowded and spreads have widened, a composition risk examined in the piece on AI debt and private credit exposure. Equity was the next window.

Where this argument is weakest

The case above rests on one company's filing and a computed ratio. Here is what it cannot carry.

The case that this is strength, not strain

A company with $242.5 billion in cash was never forced to do anything. Alphabet could have funded a year of this build from the balance sheet and kept the buyback running.

Choosing not to is defensible. Issuing equity while your share price is high is the cheapest capital any firm can get, and doing it early is textbook treasury management. The company told investors demand for AI capacity exceeds available supply, and guided to $180 billion to $190 billion of 2026 capital spending with further increases signalled for 2027. If that demand is real, raising early is correct rather than desperate.

A firm that raises $105.8 billion in six months without a covenant problem, at prices institutional buyers accepted, has demonstrated access to capital, not a shortage of it.

What the 82% figure does not prove

Free cash flow falling 82% proves that capital spending outgrew operations. It does not prove the spending was unwise, and it says nothing about return.

Half-year figures are also noisy. Capital spending is lumpy, and one quarter of accelerated equipment delivery distorts a ratio a full year would smooth. The honest form of this claim needs the 10-K, not the 10-Q.

A $0 repurchase line read from a quarterly filing should be checked against the annual report before anyone builds a thesis on it. I am confident in the reading and would still re-verify it in February.

There is also a figure I cannot get, and it is the one that would settle the argument. No hyperscaler discloses revenue attributable to AI capital spending as a clean line. The return side of this ledger is unavailable to everyone outside these companies. Anyone quoting a payback period is estimating, and the record on measurable AI return suggests those estimates deserve scrutiny.

Four signals to read in the next two earnings cycles

The $80 billion headline is already priced in. These four are not, and each is checkable in a filing.

The fourth row is the one I would set an alert for. Equity issuance by a company with a quarter of a trillion dollars in cash is a different category of event. It becomes a trend the moment it happens twice, and where that capital comes from is traced in the analysis of AI's share of venture funding.

Capital raised at a 6.25% coupon has to be earned back, which puts upward pressure on the price of compute inside every AI product you subscribe to. Headline token prices keep falling while the cost structure underneath them does not, a divergence covered in the breakdown of inference costs and AI margins. Budget for the second number, not the first.

Frequently asked questions

Where to start this week

Three things, in ascending order of effort.

Open the next quarterly filing from any hyperscaler you depend on and read four lines: operating cash flow, purchases of property and equipment, repurchases of stock, and proceeds from issuance of equity. That takes five minutes and tells you more than the earnings call.

Then look at your own contracts. Find every AI or cloud line item renewing in the next two quarters and note which vendor is funding its own infrastructure from raised capital. Those are the contracts most likely to see a price increase.

Finally, write down the number that would change your mind. Mine is capital spending exceeding 100% of operating cash flow at two hyperscalers in the same quarter. If you cannot name yours, you are reading headlines, not signals.

Related reading