Palantir Technologies Inc. reports third-quarter results on Nov. 2. Both of the sell-side notes that landed on Oct. 29 arrived before the print, and both pointed the same way: JPMorgan Chase began coverage with an overweight rating, and Rosenblatt Securities lifted its target price to $260 from $225 while keeping a buy.
The shares fell anyway. Palantir traded about 2% lower before the opening bell on Oct. 30, in a session where technology stocks were under pressure after a run of capital-spending guidance that investors read as expensive.
The moves are worth reading closely because of what they say about the quarter rather than the company. An initiation and a target increase four sessions ahead of a report are forecasts about the report. Rosenblatt’s $35 increase, a little over 15%, implies a set of estimates revised upward, and JPMorgan does not start a name at overweight unless its model already supports the rating.
Palantir has told investors to expect revenue of $2.160 billion to $2.164 billion for the quarter. The range is $4 million wide, which is narrow even by the standards of a company that has made a habit of guiding tightly and beating. On an annualized basis, the low end of that guide describes a business running above $8.6 billion a year.
What those sales consist of has changed over the past three years. The government business, Gotham, began with intelligence and defense customers and remains the base: the U.S. Army, allied militaries, and federal agencies that pay for software which fuses data from sources they will not discuss. The commercial business, Foundry, sells the same approach to manufacturers, insurers and energy companies. Apollo, a third product, delivers both into whatever environment a customer insists on, including ones with no connection to the public internet.
The growth engine for the past two years has been a fourth product, the Artificial Intelligence Platform, and the sales method built around it. Palantir runs short, intensive sessions in which a customer’s own data is wired into a working application within days. The format converts a procurement cycle that used to take a year into a demonstration that takes a week, and it has produced the customer-count growth the company now reports each quarter.
The bull case both analysts are making rests on budgets rather than on products. Defense ministries in the United States and Europe are expanding spending on software and autonomy, and the money is appropriated in multi-year programs that do not disappear when a quarter disappoints. Commercial AI budgets, meanwhile, have kept expanding even as buyers have grown more skeptical of generic tools, on the argument that Palantir sells a finished system rather than a model and a prompt library.
The bear case is arithmetic. Palantir trades at a multiple of forward sales that has no close peer among large software companies, and a multiple of that size is a claim about a decade of growth rather than a year of it. At that price, a quarter that merely meets guidance is a disappointment, a pattern the stock has demonstrated repeatedly since it joined the S&P 500 in September 2024 and the Nasdaq-100 later that year.
There are operational questions underneath the valuation debate. Government revenue is lumpy, recompeted and subject to the timing of awards, which is why the new federal fiscal year that began Oct. 1 matters for the fourth-quarter outlook. The international business has been the weakest part of the franchise for several years, growing slowly in Europe and Japan relative to the American segments. And the commercial customer count, while rising, includes many accounts that begin small; the test is whether they renew at larger sizes.
The AIP funnel has a specific vulnerability that bears watch. Bootcamp-driven sales compress the time between first contact and first contract, which flatters new customer counts, but the revenue that matters arrives in the second and third years. If those expansions slow, the growth rate will follow with a lag that is hard to see in any single quarter.
Coverage initiations cluster for a reason that has little to do with insight. A company of Palantir’s size, index membership and retail following is too large for institutional portfolios to ignore, and a rating from a large bank gives portfolio managers who benchmark against the index a documented reason to own or avoid it. The note is a gate pass as much as an opinion.
Rosenblatt’s raise carries a different signal. Analysts rarely move a target by 15% in the week before a print unless their models have moved first, and models move when channel checks, contract awards or currency assumptions change. Whether those revisions prove right is what Nov. 2 will show.
The premarket decline is the least interesting part of the week and the most revealing. A stock priced for years of acceleration trades with the sector on days when the sector is repricing capital spending, and the previous two sessions had done exactly that to the largest technology companies. Palantir’s own news was positive; its correlation was not.
What to watch on Nov. 2 is narrower than the headline. U.S. commercial revenue growth and the count of customers above a million dollars a year describe whether the funnel is converting. Remaining performance obligations, the contracted work not yet billed, describe whether the government pipeline is as deep as the bulls assume. And the fourth-quarter guide will be read against a federal budget that has just been funded and a commercial market where buyers are consolidating tools rather than adding them.
For now the sell side has placed its bets and the company has not yet answered. Two analysts moved first, and the order in which information arrives is itself information: the people paid to model the quarter expected it to be good enough to say so in public before the quarter was reported.
