The damage that spent most of October dismantling semiconductor stocks arrived in enterprise software last Friday, and it arrived fast. AI-focused software shed $260 billion in value as Goldman Sachs raised alarms about debt financing, triggering a sharp momentum reversal on October 16. The Nasdaq Composite fell 2.14% on Friday, bringing its weekly losses to roughly 4.5%, its worst week since June. Palantir, which had set a record close of $209.05 just five trading days earlier, bore some of the sharpest pain.
Palantir closed Friday, October 16, at $186.03, falling 7.22% in a single session and extending its loss since the record close to 11.01%. It took nearly a year to reclaim the old record. The victory lasted five trading sessions. At the center of the selling was a Goldman Sachs prime brokerage note documenting the crowding and the exits: global fundamental long-short funds lost money every single day this week, the Goldman Sachs prime brokerage unit said in a note to clients, with losses for the week to October 15 coming to 2.9%. Goldman added that global AI stocks collectively had been sold by long-short fund managers for six straight sessions.
The Bear Case: Debt Is the Story Now
The bear argument is that Friday was not noise. The selloff began in credit, not equities, and credit markets are harder to dismiss as short-term sentiment. CoreWeave’s shares and bonds slumped as investors questioned the borrowing behind the AI buildout. The company has about $50 billion in debt, and the cost of insuring it against default rose to around 1,000 basis points, implying a nearly 63% chance of default in five years, up from about 30% a year ago. CoreWeave ended the week down 25% after Goldman Sachs released a note saying that hedge funds sold AI stocks at the second-fastest pace in over a year.
The logic that connects CoreWeave’s balance sheet to Palantir’s multiple is uncomfortable but coherent. AI is no longer just a technology story. It is tied into rates, credit markets, and global supply chains, which is why stress can show up simultaneously in U.S. growth stocks, high-yield corporate debt, and semiconductor-heavy markets in South Korea and Taiwan. When the financing layer of the AI buildout looks fragile, the customers funding that buildout look fragile too, and software vendors that depend on enterprise AI budgets inherit that uncertainty. A valuation approaching 100 times forward earnings earns a premium only as long as growth remains unquestioned. A gap that wide also means sector selloffs land hard.
The Bull Case: Earnings Don’t Move With Sentiment
The counterargument is structural, and it is not weak. Palantir does not finance GPU clusters. It does not carry $50 billion in debt. Palantir’s second-quarter revenue grew 93%, highlighting the contrast between its strong business performance and the falling share price. The latest articles frame the move as part of a broader AI software valuation reset, while also pointing to Palantir’s still-strong fundamentals, including record deal activity and continued momentum in sovereign AI initiatives. None of that changed between October 12 and October 16.
Palantir’s spread across U.S. defense, intelligence, and commercial clients gives it a fundamental anchor that pure-sentiment bears currently ignore. The company is scheduled to report third-quarter results on November 2, which means the market will have actual numbers before the debate runs much longer. Despite the correction, the Wall Street consensus remains positive: according to forecasts from 32 analysts compiled by S&P Global, Palantir carries a Buy consensus rating and an average 12-month price target of $206.43, implying roughly 11% upside from the October 16 close. Mizuho raised its target to $245 on the same day the stock fell 7%.
Palantir is not the only defense-AI name whose fundamentals have decoupled from sentiment. Leidos carries a $48 billion backlog in government AI and defense contracts that the market has largely priced as if it does not exist.
Where the Evidence Leads
The bull and bear cases are really arguing about two separate things. Bears are right that the financing architecture of the AI buildout is under genuine stress, and that stress is not yet resolved. CoreWeave has become the bellwether for whether debt-funded AI data-center buildouts can generate enough cash flow to service their borrowing, and its widening credit spreads are now moving the broader AI-infrastructure trade. Sentiment at these multiples is fragile by construction.
But the bear case for Palantir specifically relies on a transmission mechanism that remains unproven: that CoreWeave’s credit problems translate directly into weaker Palantir revenue. The stock’s retreat from its record is not, by itself, evidence of weakening demand for Palantir’s products. The company has not yet reported third-quarter results, and the latest analyst revisions remained largely positive before Friday’s session. A Goldman credit note can move a crowded trade in a single afternoon. It cannot change what Palantir invoices next quarter.
The transmission question cuts across the entire AI software layer, not just Palantir. Datadog’s AI-native customer spending has been accelerating even as broader AI sentiment turned negative, which tests whether enterprise demand is truly insulated from infrastructure credit stress.
The verdict, at current evidence: the multiple was stretched and deserved some compression. A stock that trades at nearly 100 times forward earnings has no margin for macro fear. Friday delivered that compression efficiently. Whether it continues depends less on Goldman’s next note than on what Palantir actually reports on November 2. That number, not the sentiment behind Friday’s selling, is the one worth watching.
