These 3 AI Boom Stocks Fell as Much as 54%. Two Just Posted Their Biggest Profits Ever
Lam Research, SK hynix and Alphabet all depend on the same budget: money for AI infrastructure. The market is now selling that chain from the end backward, hitting hardest the companies that have just reported record profits. What does each have going for it, and what will determine what happens next?
Key points
Thursday, November 5, was the worst day for AI stocks in years, but the trigger didn’t come from any earnings report.
Lam Research is down 54% from its annual high, despite just reporting the highest profit in its history.
SK hynix trades at a P/E of 6.5 with an operating margin of 80%, and the market clearly doesn’t believe that figure.
Alphabet had its worst day since going public in 2004, losing around half a trillion dollars in value.
After the sell-off, it’s tempting to conclude that these stocks are on sale, but Bulios’s fair value estimates tell a different story for all three.
One budget, three bills and one bad week
Thursday, November 5, was the worst day for AI stocks in years. The S&P 500 lost 4.3%, the Nasdaq fell 5.5% and closed in correction territory, and the semiconductor index SOX plunged 12%. The trigger wasn’t in any earnings report: after the elections, in which Democrats won both chambers of Congress, the market began to fear regulation of data center construction.
Lam Research $LRCX fell more than 13% in a single day, is down -37% over the past month and -54.3% from its annual high. SK hynix $SKHY is 36.3% below its high, and Alphabet $GOOG is 26.3% below.
Company | Price (Nov. 6) | Annual high | Down from high | P/E |
|---|---|---|---|---|
Lam Research | 200.44 USD | 438.50 USD | -54.3 % | 29.7 |
SK hynix | 127.34 USD | 199.87 USD | -36.3 % | 6.5 |
Alphabet | 258.45 EUR | 350.80 EUR | -26.3 % | 15.0 |
One dollar connects these three companies. Alphabet spends it on a data center, SK hynix sells it memory chips, and Lam Research sells SK hynix the machines used to make those chips. The market is now selling that chain from the end backward: the company that is still promising AI profits took the smallest hit, while two companies that have just reported the highest profits in their histories took the biggest.
Lam Research: down 54% despite just posting its biggest profit ever
Lam Research makes etching and deposition machines, without which no advanced chip can be made. When Alphabet or SK hynix puts a factory project on hold, Lam feels it first—and that’s exactly how the market is treating it now.
But the numbers tell a different story. For the quarter ended September 27, the company reported revenue of 8.22 billion dollars, up 22% from the previous quarter and 54% from a year earlier. Its operating margin exceeded 40% for the first time in its history, and earnings per share of 2.20 USD were a record. The company is forecasting revenue of 8.6 billion dollars for the December quarter. Management estimates that the semiconductor equipment market will reach around 150 billion dollars this year and exceed 200 billion next year.
So why down 54%? Because Lam is a leveraged bet on hyperscaler investment: when they increase their budgets by 30%, machine orders can grow by a multiple—and the reverse is also true. On top of that, 35% of revenue comes from China.
What it has going for it: 7.5 billion dollars in cash, record free cash flow of 3.2 billion dollars for the quarter, buybacks of 1 billion dollars per quarter, and a target of returning 85% of free cash flow to shareholders. Even so, at 200.44 USD, the stock remains 35% above Bulios’s fair value of 129.36 USD. A P/E of 29.7 is still a growth-stock valuation, not one for a supplier at the end of a cycle.
We believe the industry is in the early stages of a multiyear investment cycle, with customer visibility extending through 2028.
Tim Archer, President and CEO of Lam Research
SK hynix: P/E of 6.5 and an operating margin of 80%. The market doesn’t believe it
SK hynix is the largest manufacturer of HBM memory, which sits alongside Nvidia’s AI chips. Last year, it had a record year, with operating profit of 47.2 trillion won. This year, it will earn that much in less than two months: in the third quarter, it reported revenue of 104.6 trillion won and operating profit of 83.3 trillion won, an operating margin of 80%. Net cash reached 98.7 trillion won, and the company will return around 100 trillion won to shareholders this year—over 30 trillion in dividends and over 60 trillion in buybacks.
A P/E of 6.5 for a company growing this quickly looks like a data error. But that’s the valuation the market gives cyclical companies at the peak of a cycle. Investors are betting that memory prices are at their peak and that an 80% margin won’t last. It’s like a landlord collecting record rent while everyone around them builds new homes. And SK hynix itself has announced that it will significantly increase investment next year.
The company makes two arguments in response. It says the memory shortage will last until 2030. And long-term contracts with 14 major customers, including cloud giants, cover over 50% of production capacity for 2027 and 2028. Even so, at 127.34 USD, the stock is 35% above Bulios’s fair value of 82.66 USD.
Our principle is to share most of the profits from successful strategic investments with our shareholders. These profits supplement our capacity to return capital to shareholders, which may temporarily decline as we prepare for the future, and allow us to continue investing boldly in sustainable growth while protecting shareholder interests.
Kwak Noh-Jung, CEO of SK hynix
Alphabet: the one paying the bill took the smallest hit
Alphabet is the customer in this chain. In the third quarter, it reported revenue of 127.6 billion dollars (+25%), Google Cloud grew 103% to 30.7 billion dollars, and cloud backlog reached 561 billion dollars. But capital expenditures swallowed 51.4 billion dollars in a single quarter, free cash flow was negative (-9.1 billion dollars), and earnings per share of 2.12 USD fell short of the expected 3.02 USD, partly because of an unrealized loss of 8.2 billion dollars on its stake in SpaceX.
The stock had its worst day since going public in 2004: it lost nearly 13% and around half a trillion dollars in market value. The market also recalled June, when the company announced an 80 billion dollar issuance of shares and convertible securities to finance AI infrastructure, including 10 billion from Berkshire Hathaway. A company with more than 260 billion euros in cash is going to shareholders for money.
Alphabet has the lowest multiple of the three, with a P/E of 15, but its free cash flow yield is only 0.6%. For the first time, the market is seriously questioning whether Alphabet is investing its capex or eating into its profits. At 258.45 EUR, the stock is 15% above Bulios’s fair value of 224.61 EUR—the smallest premium of the three.
Our full-year 2026 capital expenditure outlook range of 195 to 205 billion dollars remains unchanged. We continue to expect our capital expenditures to increase significantly in 2027, and we’ll provide further details on our fourth-quarter earnings call.
Anat Ashkenazi, CFO of Alphabet
What will determine what happens next: three dates and three signals
All three management teams are saying the same thing: the 2027 budget will grow. So the market isn’t betting against 2027, but against 2028 and against policy. We’ll get an answer around the turn of January and February:
Jan. 27, 2027, Lam Research: estimated EPS of 2.34 USD and revenue of 8.6 billion USD. More important will be whether it confirms that the equipment market will exceed 200 billion dollars in 2027.
Jan. 28, 2027, SK hynix: contract prices for memory in the first quarter and the specific amount of investment planned for next year.
Feb. 2, 2027, Alphabet (date unconfirmed): a specific capex figure for 2027 and a return to positive free cash flow.
Three signals to watch for sooner: whether hyperscalers raise capex in January or merely reaffirm it; contract prices for DRAM and HBM in the first quarter of 2027; and the first proposals for data center regulation in the new Congress.
If capex grows in 2027, November’s sell-off was about politics and sentiment, not earnings, and Lam will react the most because it has the greatest leverage. If hyperscalers slow down, SK hynix will lose pricing power and Lam will lose orders, while Alphabet will get some relief on cash flow.
Either way, one thing holds true: even after the sell-off, all three stocks are above Bulios’s fair value. A decline alone doesn’t make a stock a bargain. So far, the price has fallen below what the market expected from these companies, not below what they actually earn.
