Palantir Technologies (NASDAQ: PLTR) has become the latest casualty of AI investing’s boom-and-bust sentiment cycle, with its stock falling roughly 40% from a November peak near $207 per share.
Shares now trade near $122 to $123, valuing the company at approximately $300 billion, down from a valuation approaching $500 billion at its height.
The decline follows a surge of roughly 135% in 2025, during which investors were willing to pay a price-to-sales ratio near 110 for what many considered the operational backbone of enterprise AI.
That narrative has since flipped, with critics now dismissing Palantir as little more than an LLM wrapper sitting on top of third-party foundation models from OpenAI and Anthropic.
The financial results, however, tell a markedly different story, with revenue reaching $4.48 billion in 2025, representing 56% growth, and climbing to $1.63 billion in Q1 2026 alone, up 85% year over year.
U.S. commercial revenue growth accelerated from 109% in full-year 2025 to 133% in Q1 2026, while net dollar retention rose from 139% to 150%, signaling that existing customers are expanding their spending meaningfully.
Gross margins held steady at approximately 84% across both periods, and the company raised its full-year 2026 revenue guidance midpoint to $7.65 billion, implying 71% growth, a figure that reflects genuine business momentum.
The wrapper critique misses a critical distinction: Palantir’s AI Platform, known as AIP, is model-agnostic by design, but the platform’s real competitive asset is its Ontology, a decision framework that ties enterprise data, business logic, security controls, and operational rules into one governed system.
UBS analysts, after meeting with management and customers at AIPCon, pushed back on the notion that Palantir’s platform is easily replicated, citing its data integration and governance capabilities as advantages that frontier models alone cannot replicate.
Competition remains a genuine risk, with Databricks, Snowflake (NASDAQ: SNOW), hyperscalers, and AI labs all racing to build comparable enterprise tools that could erode Palantir’s head start over time.
Ironically, the commoditization of foundation models may actually strengthen Palantir’s position, as cheaper and interchangeable AI shifts value toward whoever can most securely connect it to complex enterprise workflows, compliance requirements, and real-world systems.
That structural advantage has taken two decades to build, and it is precisely the kind of infrastructure that enterprise customers in government, manufacturing, and healthcare are reluctant to rip out and replace.
Even so, the stock is far from cheap, trading at roughly 130 to 140 times trailing earnings and carrying a price-to-sales ratio of 56, leaving virtually no margin for operational missteps or guidance misses.
The broader lesson here is that last year’s investors justified almost any valuation because the share price kept climbing, while today’s critics dismiss the business simply because it fell sharply, and neither position reflects the underlying numbers.
Earnings growth, customer adoption, and competitive execution tend to determine long-term outcomes, and on those measures, Palantir’s trajectory continues to point firmly upward.
