AI 对冲基金 · 大师信号(存档)

The second derivative is what caught my eye: revenue growth fell to the mid-30s through most of 2025, then jumped back to 63% and 62% in the last two periods. Operating margin also improved steadily, from -50% to -25%. Debt/equity dropped from 2.10 to 0.20 and the current ratio is 5.48, so this isn't the mix of worsening numbers and leverage that blows up accounts. But the per-share numbers keep this from being a big bet. Quarterly EPS has been stuck between -$0.08 and -$0.11 for two years because the share count kept growing (book value per share rose ninefold). Free cash burn per share is also deeper, -$0.18 now versus -$0.09 a year ago. With a $39.3B market cap and no earnings behind it, the market is already paying for much of this acceleration. I'd take a small long on the inflection, not a full-size position, and add only if the margin gains start showing up in per-share cash flow.

I can't put this in any of my boxes yet, because there are no earnings to grow. EPS is still negative (-$0.08 in each of the last two quarters), so the P/E of -246 is meaningless and the PEG test can't be run. The business is heading the right way: revenue growth sped up to 62%, gross margin rose from 27% to 36%, operating margin went from -50% to -25%, and debt-to-equity fell from 2.10 to 0.20 with a current ratio of 5.48. But book value per share jumped from $0.64 to $5.84, which looks like the balance sheet was fixed with new stock rather than profits, and cash flow is still negative. A $39.3B price for a company that loses money means you're paying for a story, so I'd wait for real earnings before buying, though the improving numbers keep me from calling it bearish.

The trends are going the right way: gross margin rose from 27% to 36%, operating losses fell from -50% to -25% of sales, revenue is growing about 60%, and debt/equity dropped from 2.10 to 0.20. But this business has never shown me a profit or positive free cash flow. Return on equity is negative in every period, and since the company is losing money, the tenfold rise in book value per share must have come from outside capital (new stock or conversions), not from earnings it kept. A moat shows up as steady high returns, and I can't see one here. Paying $39.3 billion for a company with a negative P/E means betting on a future that has to go perfectly, and I don't do that.

Flip it around and ask how you lose money here. You pay $39.3B for a business that has lost money in every period shown: ROE averages -29%, net margin averages -35%, and free cash flow per share is negative every quarter, getting worse from -0.04 to -0.18 over the past two years. The 156% annual growth in book value per share and the drop in debt/equity from 2.10 to 0.20 did not come from earnings, because EPS never once turned positive; shareholders put that capital in, and a company that keeps asking for money is not compounding it. Revenue growth of 62% and gross margin rising from 27% to 36% are real progress, but a P/E of -246 means you are paying up today for profits nobody has seen, and that is not a bet I will make.

There is no earning power to pay for. The company lost money in all ten periods on record, with average net margin of -35%, average ROE of -29%, and negative free cash flow every time. The negative P/E of -246 means a 39.3B market cap rests on hope, not results. The balance sheet is sound: the current ratio is 5.48 and debt-to-equity is 0.20. But book value per share rose from 0.64 to 5.84 while the company kept losing money, which means the gain came from new capital rather than retained earnings. Strong liquidity does not create a margin of safety where no earnings exist, so the defensive investor must stay away.

This is a second leg up, not a peak. Revenue growth dipped to 20% in 2025/Q4, then picked back up to 35% and 36%. Gross margin went 59% → 62% → 66% → 68% and operating margin 51% → 60% in three quarters. Year-over-year EPS growth went 39% → 35% → 59% → 78%, with the latest quarter at 136.27 against 76.64 a year ago. A 32x P/E looks rich next to a record 40% ROE, but if you annualize the latest quarter the real multiple is much lower. The market is pricing a great company, not one whose earnings momentum is still building. The balance sheet can take a hit: debt/equity has fallen from 0.56 to 0.45 and the current ratio is about 2.5. What keeps me from betting the ranch: at a $2.26T market cap and record margins, much of the good news may already be in the price. Free cash flow per share fell to 53 from 67 the quarter before and trails earnings badly. So I'd build the position and watch whether cash flow catches up with reported profits.

This is a fast grower, and it's hard to believe at $2.26 trillion: quarterly EPS went from 76.64 to 136.27 year over year, up 78%, and trailing-twelve-month earnings are up about 54%, on revenue growth of 35-36%. At a P/E of 32 against 45-75% earnings growth, the PEG comes out around 0.5-0.7, so I'm not paying for a story. The story checks out: gross margin went from 58% to 68% and operating margin from 47% to 60%, on a balance sheet with 0.45 debt-to-equity and a 2.46 current ratio. What keeps this out of the classic-setup range is that this is chipmaking, a cyclical business, and margins are at a peak that can't keep expanding. Free cash flow (53 a share last quarter against 136 of EPS) trails earnings badly because of heavy spending on plants, and a company this big has to keep beating giant numbers.

This is a wonderful business on the numbers: return on equity has climbed from 29% to 40%, gross margin from 58% to 68%, and book value per share has compounded about 24% a year. It has done that while cutting debt/equity from 0.56 to 0.45 and keeping a current ratio near 2.5, which looks like real pricing power and disciplined management. What gives me pause is that this is a capital-hungry business. Over the last four quarters only about half of reported earnings turned into free cash flow, so the stock's 32x P/E works out to more than 60x free cash flow. That price assumes 30%-plus growth keeps going in an industry that has always been cyclical. I'd gladly own this business for ten years, but at this price I'd rather wait for a better pitch than swing now.

This is a superb business by every measure shown. ROE rose from 29% to 40%, operating margin from 47% to 60%, book value per share compounded at 24% a year, and debt/equity fell from 0.56 to 0.45, so inverting it turns up no rot in the numbers. The problem is that I only have ten quarters, all from an upcycle, and the business eats capital: trailing free cash flow is about 217 per share against about 427 of EPS, and conversion slipped to 39% in the latest quarter. At 32 times record-margin earnings, the free-cash-flow yield is only about 1.6%, which only works if a cyclical chip foundry holds 60% operating margins while capex keeps swallowing half its profits. That's a great business at a price I won't pay without seeing a downturn in the record, so it stays on the watch list rather than in the portfolio.

The business itself passes my tests: the current ratio is 2.46, debt-to-equity is a modest 0.45, and every period shown is profitable, with margins and book value per share rising steadily. The price is another matter. Trailing earnings (the last four quarters, 136.27 + 110.44 + 93.54 + 87.10 = 427.35) against book value per share of 1,248.33 put the stock at about 32 times earnings and roughly 11 times book. That product is about 350, against my ceiling of 22.5. Free cash flow of about 217 per share over the same four quarters is only half of reported earnings, so the buyer pays some 63 times the cash the business actually produces. The defensive investor gets no margin of safety here, only payment in advance for growth that has not yet been earned, and I call that overvaluation a bearish fact however excellent the enterprise.

The top line has clearly turned: TTM revenue growth went from shrinking 18-52% through 2025 to +199% and +185% in the last two quarters. Gross margin rose to 43% and operating margin improved from -12.1x to -5.5x revenue. But the move is already slowing a little, and it comes off a tiny base. The company still loses about 5.5 times its revenue at the operating level, cash burn per share grew from -0.06 to -0.08, book value per share fell from 1.75 to 1.61, and the current ratio dropped from 37 to under 4 in two quarters. At a 5.3B market cap the market is already paying for the turnaround and much more, so this isn't an asymmetric bet I'd make. Leverage is still modest at 0.21 debt/equity, so it's not a short either; I'll pass and wait for the operating losses to start closing on their own rather than just shrinking against a bigger revenue number.

I can't put this in any of my categories. It's not a fast grower, stalwart or turnaround, because there are no earnings to grow: trailing EPS is -$0.16 and operating losses run about 5.5 times revenue. The two quarters with positive EPS (Q1 2025 and Q1 2026) came with operating margins of -14.7x and -5.9x revenue, so that profit came from outside the business. There's no PEG to figure with a P/E of -19. Sales are up 185% and the operating loss is shrinking against revenue, but that's off a base that shrank in 2024 and 2025. The balance sheet has cushion, yet the current ratio has fallen from about 37 to 3.9 and debt/equity doubled to 0.21 in a quarter. That leaves $5.3 billion paid for a story with no earnings under it. That's the whisper stock I've always told people to pass on, however good the story sounds.

I can't tell you what this business will earn in ten years, and the numbers can't either. Operating losses have run between 5 and 15 times revenue in every period shown, return on equity has been negative throughout, and free cash flow per share has been negative every period. The only positive earnings came from swings below the operating line, not from the business itself. Book value per share did rise from $0.37 to $1.61, and the strong current ratio and low debt look comfortable, but with steady losses and cash burn that growth looks to have come from selling new shares rather than from retained profits, so it isn't compounding. There's no moat, pricing power or earnings power in these figures, and a $5.3 billion market cap for a company that has never made money on its operations is a price that demands perfection, so I'll pass.

Flip it around and failure is already here: operating losses have run at 5 to 15 times revenue in every period shown, ROE has been negative every single period, free cash flow per share has never been positive, and gross margin has slipped from 0.51 to 0.43. Book value per share went from 0.37 to 1.61 while the company lost money the whole time, so that growth didn't come from earnings; it has to have come from selling new shares or some other outside capital. That isn't compounding. The quarters with a positive net margin (Q1 2025 at 28.95, Q1 2026 at 7.52) came while operating margins sat at -14.70 and -5.90, so the profit came from something outside the business, and I don't count it as earnings. Paying a $5.3B market cap for a business that loses several dollars for every dollar of sales and has shown one good revenue-growth figure after years of shrinking means believing something stupid, so I pass.

There are no demonstrated earnings here to value. Operating losses ran between 5 and 15 times revenue in every period shown, and free cash flow was negative throughout. The few quarters with positive net income (net margin +7.52 in 2026/Q1, +28.95 in 2025/Q1) came while operating margins stayed deeply negative, so they reflect non-operating items rather than earning power. The P/E of -19.12 and ROE of -0.44 put the $5.3B market cap at roughly 8.4 times book value, which is a price paid entirely for hope. The balance sheet is adequate, with a current ratio of 3.94 and debt/equity of 0.21. But the current ratio has fallen from about 37-41, and the rise in book value per share from 0.37 to 1.61 did not come from retained earnings, since the company lost money over the record. That leaves no margin of safety for a defensive investor.