Good Companies, Impossible Prices: Palantir, Tesla, and Robinhood

Last week a stock gained more than 12% in a single session without announcing anything. Bitcoin did the talking: it jumped 7.5% in a day on renewed hopes for U.S. cryptocurrency regulation, and Robinhood — the brokerage where retail traders would have bought that move — went along for the ride. A few weeks earlier, Palantir reported a quarter so clean it lifted its stock by nearly 30% in a day. And Tesla, down 22% for the year, still trades at roughly 364 years of its current profit.

That is what an expensive market feels like: not simply high prices, but prices that start selling the story themselves. We are in one. On ten years of inflation-adjusted earnings, the S&P 500 hit its second-most expensive valuation in history this month, with only the dot-com peak richer, and the index's dividend yield fell to its lowest level ever. Even within this year, a speculative pocket has already detonated: a memory-chip mania inflated and popped in about four months, taking a hedge fund down with it, while the broad market barely blinked.

I am not calling a crash, and I do not want to be read as doing so. Aggregate valuation is a poor market-timing tool, and it has been wrong about "too expensive" for years. What a record multiple does is narrower: it tells you where the fragility concentrates. It concentrates in stocks priced off narrative rather than profit, because those carry no cushion — the entire return is the multiple, and the multiple requires the story to keep improving forever.

The test for "inflated" is arithmetic. A stock is inflated when the price stops being a comment on the business and becomes a bet that the story never stops getting better. Three of the most-watched names in tech fall on the wrong side of that test today, each for a different reason. Palantir is a superb company at an absurd price. Tesla is a large story attached to thin, shrinking profits. Robinhood is a real business carrying borrowed excitement. Separate them, because each demands a different kind of discipline.

Palantir: the flawless quarter that still doesn't justify the price

Start with the one that is hardest to dismiss, because everything Palantir said was good. Second-quarter revenue jumped 93% from a year earlier to $1.94 billion, adjusted EPS of $0.41 cleared estimates, U.S. commercial revenue rose 149%, and management raised its full-year revenue guidance by nearly $500 million. The stock surged nearly 30% on the news and has risen more than 40% in the past month.

Here is the price those numbers buy you today: roughly 139 times trailing earnings and 68 times sales. To translate the second one into feeling: the average S&P 500 company sells for about 3.2 times sales, so the market is paying $68 for every dollar of Palantir's revenue — about twenty-one times what it pays for a dollar of the index's revenue. At 139 times earnings, you are handing over 139 years of current profit to own the share.

The usual defense of a high multiple is "the growth justifies it." I would push back harder than usual. That argument is comfortable only when the market is under-appreciating the growth — when skeptics are still fighting the company. Nobody is under-appreciating Palantir anymore. The report was watched by everyone, price targets rose afterward, and the aggregate analyst signal tracked by AInvest still reads Buy. When belief is already unanimous, the multiple has no slack left.

Even on sober long-horizon numbers, the price is ahead of the value. Days before the report, Morningstar put its discounted-cash-flow fair value for Palantir at $153; the stock now trades above that despite the beat and the guidance raise. That is what "priced for perfection" means in practice: the report had to be flawless just to keep the price roughly in place, and even flawless left it above what a long-horizon model says the business is worth.

Can the bull case still work? Yes, and I want to state it in full. If Palantir keeps compounding revenue at 60% to 90% for years, a future revenue base many times today's size makes a current 68-times-sales price look routine in hindsight. That is a specific, testable claim, not a fantasy. But it is now the base case, which is the problem: the market grants no discount for the chance that growth simply slows to merely excellent. The buyer who arrives today, after a 40% snapback in a month, is paying full retail for the best possible outcome. I would wait for the price to come to the business — a real de-rating, a pullback, some actual skepticism — before putting new money here. If you have owned it through the fear months, trimming some into a month like this is discipline, not prophecy.

The number I would watch is the year-over-year revenue line. Hold at 60% plus while the multiple compresses and the bull case survives. Settle toward market speed while the multiple stays rich, and the price becomes the entire risk — there will be no valuation cushion left to break the fall.

Tesla: 364 years of profit, and the bill keeps growing

Tesla is the mirror image: the price fell, and the stock is no cheaper for it. The company delivered a record 480,126 vehicles last quarter, up 25%, and produced revenue of $28.24 billion, up 26%. Then the profit line did the opposite: adjusted EPS of $0.33 came in well below the $0.51 Wall Street expected, GAAP net income fell, and management said to expect further cash burn. Operating margin is about 5%; return on invested capital is barely 3%.

Here is the key insight: a fallen price is not a discount when the earnings fell with it. At roughly $350 — down 22% this year and about 30% off its high — Tesla still carries a $1.39 trillion market cap and 364 times trailing earnings. Because profits are shrinking, the forward multiple is higher, not lower: on estimates, a future dollar of Tesla profit costs more than today's. "It's already come down from $499" is a location, not a valuation. The multiple tells you the market is still pricing the company as a finished autonomous ride-hailing monopoly while the income statement describes a thin-margin car maker funding that bet out of shrinking profit.

I want to be fair about the story, because it is real: unassisted robotaxi service is rolling out and expanding in Austin, and new Full Self-Driving approvals keep landing abroad. Tesla has moved from narrative to operations on autonomy, and I have been on the wrong side of this stock before, so I do not assume the market is wrong. My complaint is structural: a $1.39 trillion price must keep recruiting new believers every day, and every quarter in which the profit line fails to move simply re-prices the same unresolved bet. Caution here is not a forecast of failure; it is symmetry — the same valuation that forces Tesla to become a monopoly later also has to be re-earned daily now.

The condition that would change my mind is the profit line itself: robotaxi and AI spending eventually landing as operating income, with margins and per-share earnings inflecting upward while deliveries recover — not more product demos and delivery records. Until then, the honest posture for a holder is position sizing and price action, and the price action is already sour: the stock sits below both its 50-day and 200-day averages, a broken trend. If you are tempted because the price has come down, read the multiple before the share price. They tell opposite stories.

Robinhood: a good business selling borrowed excitement

Robinhood is the instructive case, because its fundamentals are genuinely strong — and that is precisely why its price is dangerous. The company just reported record second-quarter revenue of $1.31 billion, up 32%, with diluted EPS up 48% to $0.62, record net deposits of $22 billion, and a record 4.8 million paying Gold subscribers. Record trading volume in equities, options, and prediction markets carried the quarter. On those numbers alone, this is a good business.

The inflation is in what makes the numbers what they are. Robinhood's revenue is the crowd's mood — trading volume. When the crowd cooled earlier this year, crypto-related revenue stalled, sales growth slowed to the mid-teens, and the stock fell from a $154 high to a $64 low — losing more than half its value in that swing. The company broke nothing; the volume did. The same surge that tempts people today — up more than 12% in a single session on a bitcoin move, 13% in a week — is the same force that can return it all.

The arithmetic captures the tension. At about $110, Robinhood trades near 48 times trailing earnings and about 20 times sales for a $99 billion market cap. The forward multiple is higher, around 68 times, because the Street itself doubts this quarter's volumes are the new normal. Morningstar's long-horizon fair value for the stock is $57; the market price sits roughly 90% above it. Argue with the model if you like — the direction is the point. The price is a boom priced as a base case.

I would not chase this rally, and I would be explicit about it: paying 48 times boom earnings because the crowd is momentarily excited about the crowd is the surest way to hand back a year of gains. If the business is durable — and I suspect it is — the price will come to you. The same sentiment that lifts the stock 12% in a day on a bitcoin headline can return it within months, and that is the entry you actually want.

The discipline, not the prediction

None of this is a forecast of tops. Palantir can compound at a rate that makes today's price look quaint in hindsight; Tesla can mint the robotaxi outcome the multiple prices in; Robinhood can convert a trading boom into a durable financial franchise. A 139-times or 364-times multiple can persist for years, and fighting it in the short term is usually a losing trade.

The point of the exercise is the test, because it is repeatable. Before buying anything into a market this expensive, ask what the price is telling you: does it reward the business as it is today, or does it require the story to keep improving forever? When the answer is the latter — and for Palantir, Tesla, and Robinhood, the arithmetic says it is — then being right about the company is no longer sufficient. You also have to be right about what the crowd will pay for it later. That is a thin cushion to build a position on, which is why the defensible move is to require the numbers to move first, and to decline to pay full price for a story that is already doing its own marketing. In a market at record valuation, the difference between a good investment and an overpriced one is rarely the company at all. It is the price agreed.