Palantir's 85% Quarter and the Quiet End of Per-Seat Software

Palantir reported $1.63 billion in Q1 2026 revenue, up 85% year over year, with US commercial revenue growing 133%, according to the company's SEC earnings press release. Manufacturing and supply chain work drove the commercial line, per CNBC's earnings coverage and Futurum's analysis. The print was the cleanest evidence yet that an AI-native vendor can scale on a pricing model that doesn't charge per user. It landed the same week a Palantir market strategist went on record calling the old model dead.

Danny Lutkus, a market strategist at Palantir, told Forbes that "SaaS is dead" in the context of supply chain software, per o.talking.com's coverage of the interview. The Forbes original is not in the public source bundle, and o.talking.com is a Korean newsletter writing in English, so the quote should be read as Palantir's framing of its own thesis, not an industry-wide consensus. The framing is worth taking seriously anyway, because it names a structural change the print makes legible.

SaaS, or software as a service, is the per-seat subscription model that has dominated business software for two decades. Buyers pay a recurring fee per user, vendors grow by adding seats, and product teams optimize for engagement and time-on-tool. The logic only holds as long as humans are the ones logging in. When an AI agent does the work, the seat disappears, and the vendor's revenue model disappears with it.

The model replacing it is outcome-based pricing: the buyer pays for a finished result, not for access. The agent becomes the user. The dashboard becomes optional. In supply chain, that means paying for a replenished warehouse or a rerouted shipment, not for licenses to a planning tool. The unit of sale is no longer access; it is the work itself. Futurum's breakdown of the quarter attributes the commercial acceleration to US demand for the company's AI platform, sold increasingly on results rather than seat counts.

Two things made the shift possible in 2026. Agent reliability has crossed a threshold where vendors are willing to put outcome clauses in contracts, and the unit economics of compute have fallen far enough that vendors can absorb delivery risk on a per-result basis. Both move in the same direction, and both reward vendors willing to reprice first. The risk for incumbents is that the longer they wait, the harder the migration becomes.

The structural change has three sets of consequences, and they don't all land on software vendors. For software buyers, contracts shift from seat-count forecasts to deliverable definitions, and the legal work moves to defining what "done" means. For builders and product teams, the product stops being the interface the user touches and becomes the system that produces the result. Pricing pages, sales motions, and hiring profiles get rewritten around the agent, not the seat. For adjacent service workers, including law firms, consultancies, and agencies, the same logic applies one layer up. If time-on-tool stops being the billable unit for software, the hourly billable hour for human services is the next target. The broader point is that outcome pricing is a category change, not a vendor pitch.

Outcome pricing has documented failure modes: vague deliverables, disputes over what counts as finished work, and buyers absorbing open-ended scope risk. A vendor paid per outcome has an incentive to define outcomes narrowly; a buyer paying per outcome has an incentive to define them broadly. Legacy SaaS vendors aren't going quietly, and procurement teams still need predictable run-rate costs, not open-ended outcome agreements. The 85% quarter proves the model can scale at one vendor under one set of contracts. It does not prove every buyer or every use case will adopt it.

The next signal to watch is contract language. When RFPs start defining deliverables instead of seats, the transition has moved from marketing into procurement. The 85% quarter made the thesis legible; the next earnings call, and the customer disclosure that comes with it, will show whether the model is being bought at scale or only piloted.