STXU

SUNDAR Pichai led with the biggest number he had. Alphabet, he said in the statement released after Wednesday’s (Oct 28) closing bell, was “now on a half-trillion dollar revenue run rate”. Google Cloud revenue had doubled to US$30.7 billion; the Gemini app had passed a billion users.

Within the hour none of it mattered. The shares slid from US$322 to around US$288 in after-hours trading, erasing roughly US$400 billion of market value, as reported earnings fell well short of forecasts.

During the call, Anat Ashkenazi, the finance chief, reported US$51.4 billion of capital expenditure in three months and free cash flow of minus US$9.1 billion, against the minus US$2.7 billion analysts had pencilled in, and said spending would “increase significantly in 2027”.

Thomas Kurian, who runs Google Cloud, conceded his business “could obviously serve a lot more demand with capacity”. Then Philipp Schindler, the chief business officer, explained part of why search grew only 12 per cent, down from 19 per cent in the first quarter of 2026: Queries were moving to AI Mode, where Google earns less per answer.

By the time Asian markets opened, the argument over the sell-off had split four ways.

One camp blames the spending. Another says search is eroding faster than cloud can compensate. The third points out that a US$8.7 billion paper loss on equity stakes made a solid quarter look like a miss. The fourth fears that artificial intelligence agents, buying and booking on users’ behalf, will leave fewer ads to click.

All four are arguing about one side of the invoice.

Follow the money east. Google Cloud’s capacity shortage translates into Samsung’s pricing power. On Thursday, Samsung Electronics confirmed third-quarter operating profit of 107.4 trillion won (US$80.4 billion) – more than double its profit for all of 2025, at a 55 per cent margin.

The same dollar is a cost in California and a profit in Seoul, and a 55 per cent margin is a derivative of someone else’s balance sheet.

Samsung, though, sits on both sides of the invoice. Outside the group, Alphabet’s spending is its chip division’s profit. Inside, the same chip prices are its phone division’s cost.

Thursday’s figures make the point starkly: The chip business earned 106.5 trillion won of the 107.4 trillion won. Phones, televisions, displays and everything else combined made less than one trillion won.

Memory now accounts for more than 40 per cent of a premium smartphone’s component cost, Counterpoint Research estimates, and Samsung’s Galaxy unit has reportedly asked its own chip division for better prices.

The windfall is a transfer, paid by Alphabet’s shareholders, by phone buyers around the world and by colleagues down the corridor. None of those sources is permanent.

Schindler’s admission is what self-cannibalisation sounds like when a company chooses it. Google is paying to replace the business that made it rich, because the threat arrives with every query.

At Alphabet’s Q3 2026 earnings call, its shareholders saw both halves of the price: the capital spending, and the cheaper answers.

Samsung faces a similar question about what its profits should become, but with none of that immediate pressure. AI is enriching its core business, yet the thinking must begin before the transfer stops.

Fujifilm: Reading a windfall for what it was

Asia has one company that read a windfall for what it was, and it did so at the top. In the year 2000, colour film supplied about 60 per cent of Fujifilm’s sales and two-thirds of its profits, by the account of Shigetaka Komori, who became president that June.

That year world demand peaked; film had never looked better. Within a decade demand had fallen to a tenth. Komori cut 5,000 jobs and moved the company’s film chemistry into flat-panel materials and healthcare before the profits ran out. He spent the last good years of one business buying the next.

Samsung gave its own answer alongside its Q3 results: a quarterly dividend of 30 trillion won, or 4,601 won a share. Total shareholder returns between 90 trillion won and 110 trillion won are expected for 2026.

Earlier in May, on the eve of a planned strike, Samsung agreed to set aside 10.5 per cent of its chip division’s operating profit for employee bonuses, paid in mostly shares, over 10 years. In August, its board approved 15 trillion won of share purchases for staff pay.

SK Hynix wrote a similar formula: 10 per cent of operating profit to employee bonuses for a decade, over the objections of managers who wanted the surplus for fabrication plants.

None of this is reckless. Shareholders own the windfall, employees earn it, and the money left for fabs is still enormous.

TSMC, the most disciplined reinvestor of the lot, spent NT$443 billion (US$13.6 billion) on their plants in Q3 2026, about half its operating cash flow.

The choices differ: A dividend sends the gains out of the company, a bonus pool shares them inside it, and a fab bets that they continue. What none of them buys is a business that would outlive the AI build-out – and demand for leading-edge chips lasts only as long as the cannibals keep building.

Asia’s chipmakers are feeding much of that windfall into the boom, with no obvious plan for what follows.

Wall Street spent a night pricing Google’s decision to pay for its own replacement. Samsung spent the next morning announcing how it would share the proceeds.

In 2000, Komori took charge of a company that earned two-thirds of its profits from film, and began preparing for film’s death. Now, 26 years later, 99 of every 100 won of Samsung’s operating profit comes from chips. Alphabet is burning cash to escape the business that made it. Samsung is collecting it.

Who at Samsung is paid to imagine the company without it?

The writer, a former investment banker, is a technology journalist who writes about Asian business, technology, innovation and capital markets. He previously worked at GE Capital and Korea Development Bank group