Skip to Content
Opening Bell

Close-UpMonday, August 1011 Min Read

Alphabet's $99 Billion Profit Came From Shares It Never Sold

$99 billion of Alphabet's second-quarter net income, and $53.4 billion of Amazon's, came from remarking stakes neither company has sold. Strip the two out and S&P 500 earnings growth falls from 50.4% to 32.0% — and JPMorgan raised its target on Monday looking at the higher figure.

Last quarter Amazon ran the largest retail network and the largest cloud business on earth and produced $27.5 billion in operating income. In the same quarter it booked $53.4 billion below the operating line — roughly twice as much from remarking its stake in Anthropic as from running the entire company.

At Alphabet the ratio is stranger still. Second-quarter revenue was $119.8 billion. Net income was $112.1 billion. Profit was very nearly equal to sales, for a simple reason: $99.0 billion of it came not from anything Alphabet sold, but from shares it has not sold — and largely cannot sell.

On Monday, JPMorgan raised its year-end target for the S&P 500 to 8,000. The stated reason was profit growth that is "strong and broad-based."

By the Numbers

$99.0B

Alphabet's quarterly gain on equity investments

$53.4B

Amazon's non-operating income tied to Anthropic

50.4%

S&P 500 second-quarter earnings growth

32.0%

What's left with those two excluded

A Rule Written in 2018, Billed in 2026

None of these numbers is a trick. All of them are audited and all of them are GAAP. There is no accounting scandal here. There is an accounting rule, and it is working exactly as drafted.

Before 2018, an increase in the value of shares a company held did not touch the profit line until those shares were sold. Such gains accumulated in a separate compartment called other comprehensive income and never reached EPS. The standard the FASB issued in 2016 and put into effect in 2018 changed that: changes in the value of equity investments now flow straight through net income.

For holdings in companies that are not publicly traded, the rule carves out an exception. Those stakes are carried at cost until an observable price appears. The moment a funding round or an IPO establishes that price, the stake is remeasured — and the entire accumulated difference, however many years it took to build, lands in that single quarter's income statement.

The stake the drafters had in mind was a minority interest in an industrial supplier. In 2026 the same rule ran across a $94 billion line on Alphabet's balance sheet.

The Line That Sat Still for Eleven Years

Google put $1 billion into SpaceX in January 2015 alongside Fidelity; together the two bought roughly a tenth of the company. For eleven years that investment sat on Alphabet's books at cost, because SpaceX had no price.

On June 12, 2026, SpaceX went public and the line acquired one. As of June 30 Alphabet's stake was worth $94 billion, an effective ownership of about 4.9%. Eleven years of accumulated appreciation landed in one quarter's earnings.

Alphabet cannot sell any of it. Roughly $80 billion of the stake sits under the standard post-IPO lock-up; the remaining $14.1 billion is restricted through the third quarter of 2027. The restrictions only began to ease after SpaceX reported its first set of results on August 4.

Both get added on the same line and divided by the same share count. What comes out the other end is a single number: earnings per share.

The Mechanism: One Dollar of Share Price, $550 Million of Profit

One division problem explains why this structure is so volatile.

Here is the uncomfortable half of the arithmetic: the rule is symmetric. What it writes up, it writes down.

A share that was $171 on June 30 was $107.01 on July 28 — a decline of more than 37%. Apply that to a $94 billion stake and you get an unrealized loss of roughly $35 billion, more than a third of the $99 billion booked in the second quarter. An analysis published by the Motley Fool argues that if SpaceX does not recover in the coming weeks, Alphabet could report an outright loss for the third quarter.

That is not a forecast. It is the definition of the rule.

The chart below is live; it shows where the stock stands today, not the event itself.

GOOGLAlphabet Inc
Alphabet — three months back from today

How Much Did Index Earnings Actually Grow?

Volatility inside one company's income statement spills into the whole index if the company is large enough. In the second quarter, it spilled.

According to FactSet's August 7 update, blended second-quarter earnings growth for the S&P 500 was 50.4%. Strip Alphabet and Amazon out of the same calculation and it falls to 32.0%. At the start of the quarter, analysts had expected 23.1%.

S&P 500 Second-Quarter Earnings Growth

Including Alphabet and Amazon50.4%
Excluding the two32.0%
Expected at the start of the quarter23.1%

The gap between those bars says that more than a third of the index's headline earnings growth came from two companies' non-cash markups. The same distortion is sharper in the beat statistics: companies reported earnings 29.2% above estimates in aggregate, the widest surprise FactSet has recorded since it began tracking in 2008. Excluding those two names, the surprise falls to 10.9%.

Amazon's individual share is large on its own. FactSet calculated that the company alone accounted for 76% of the dollar-level increase in S&P 500 earnings. Amazon reported EPS of $5.75 against an analyst estimate of $1.82.

Same Day, Same Numbers, Higher Target

On Monday morning JPMorgan strategist Dubravko Lakos-Bujas lifted his year-end S&P 500 target from 7,800 to 8,000. He raised his 2026 EPS estimate from $350 to $365 and his 2027 estimate from $390 to $420, holding the forward multiple at roughly 20 times. The index last closed at 7,757.64; the Wall Street consensus target sits near 7,845.

Honesty is required here, because there is an easy and wrong connection to draw. What JPMorgan raised is a forward earnings estimate, and forward estimates as a rule exclude non-operating investment gains. The target increase is not built directly on that $152 billion of paper profit.

The problem sits somewhere subtler. The numbers that generate the headlines — "record profits," "50% growth," "the biggest beat on record" — do include those gains. The market's story is fed by the headline, not by the estimate table.

ReadingThe claimWhere it's weak
The number is inflatedMore than a third of headline growth comes from two companies' non-cash markupsEven stripped out, growth is 32% — far above the historical average
Nothing is wrongForward estimates never counted these gains; the target raise doesn't rest on themThe investor reading the headline doesn't make that distinction; the narrative feeds on the same figures
The rule is flawedRunning the daily price of an unsellable asset through EPS drains the measure of meaningThe alternative is never showing a real gain in value at all

Timeline

Timeline

  1. June 12SpaceX goes public at $135. Alphabet's eleven-year-old investment gets a market price for the first time.
  2. June 16The stock peaks at $225.64.
  3. June 30The quarter closes. Alphabet's stake is booked at $94 billion.
  4. July 22Alphabet reports net income of $112.1 billion, of which $99.0 billion is appreciation on equity investments.
  5. July 28SpaceX hits $107.01, its lowest level since the IPO.
  6. July 30Amazon reports net income of $62.6 billion, including $53.4 billion of non-operating income tied to Anthropic.
  7. August 4SpaceX reports its first quarterly results. Alphabet's lock-up begins to ease.
  8. August 7FactSet publishes that growth falls from 50.4% to 32.0% with the two excluded.
  9. August 10JPMorgan raises its year-end target to 8,000.

Two Paper Gains That Will Not Behave Alike

Alphabet's and Amazon's gains appear on the same line but will not move the same way. The difference lies in how the underlying price is formed.

Alphabet now holds shares in a listed company. There is a price every day, so the stake will be remeasured at every quarter end. From here on this line will write up or down every three months; Alphabet's income statement is tied, whether it likes it or not, to where SpaceX closes on the last day of a quarter.

Amazon's Anthropic stake has no market price at all. It gets remeasured only when a new funding round creates an observable one. So Amazon's number will not oscillate quarterly like Alphabet's; it will sit motionless for long stretches and then jump — or drop — like a step.

For an investor the upshot is this: one is a number in permanent motion, the other a one-off jolt after long silences. Neither is cash. Both are inside earnings per share.

It helps to look at the broader market for comparison.

QQQInvesco QQQ Trust
QQQ, which tracks the Nasdaq 100 — three months back from today

What's Left

For Alphabet there is a real gain here: $1 billion committed in 2015 is worth $94 billion eleven years later. That is not an accounting game, it is the result of a good investment. What is contestable is when, and on which line, that result gets shown.

As the lock-up eases, Alphabet may sell part of the stake. The moment it does, the paper gain becomes cash and the argument ends. But trying to sell a 550-million-share position into a stock that has swung violently since its debut is a problem in its own right — the size of the seller is precisely what pushes the sale price down.

Meanwhile the index's own visible multiple is shaped by the same figures. FactSet puts the S&P 500's forward 12-month price-to-earnings ratio at 20.0, just above the five-year average. The trailing multiple, whose denominator is swollen by $152 billion of paper gains, looks cheaper than it is.

The index did have a genuinely good quarter. Even with the two companies removed, 32% earnings growth is not weak by any standard. But the distance between 50.4% and 32.0% is the distance between two different stories: the first describes how much a group of companies sold, the second how much the shares they happened to hold were worth that day.

This quarter, Wall Street used both in the same sentence.

This piece draws on FactSet's earnings season updates of August 3 and August 7, 2026, Amazon's second-quarter results filed with the SEC, Fortune and Motley Fool coverage of Alphabet's disclosures, CNBC and Benzinga reporting on JPMorgan's August 10 strategy note, and Axios's assessment of earnings quality. Alphabet's investment gain is reported variously between $98 billion and $99 billion: FactSet counts $98 billion under "other income," while the company's gain-on-equity-securities line is $99.0 billion. The third-quarter markdown figure is not a disclosed number but an estimate calculated from public prices.