Palantir (PLTR) CEOAlex Karp just pitched the company’s next big AI moment as a trust crisis.
In a fresh television appearance on CNBC’s “Squawk Box,” Karp said the AI boom has created a problem most businesses aren’t too eager to say out loud.
Investors are chasing bigger models and faster chips. Still, Karp cut through the noise, pointing to an uncomfortable reality about what businesses give up when they hand over their most valuable data and decisions to outside AI providers.
Interestingly, this comes as the stock has gained 11% over the past week, potentially signaling a break from the broader downtrend.
For perspective, shares are still down 30% year to date and 22% over the past month, according to Seeking Alpha.
In a sharp pivot from the typical AI narrative investors have been fed, Karp is reinforcing Palantir’s core pitch that control could become the scarce asset in enterprise AI.
That raises the bigger question for investors: Is Palantir riding the AI hype or exposing the weakness beneath it?
What Alex Karp said about AI’s trust problem
Palantir CEO Karp believes the way AI is being sold may no longer align with what serious customers actually need.
In his CNBC appearance, he identified the issue as a trust gap between companies using AI and the frontier labs supplying the models.
He argued clients have a “level of discomfort and loss of trust,” especially in sensitive sectors where data, intellectual property, and mission-critical decisions cannot be treated like ordinary software inputs.
More AI:
- Goldman Sachs has blunt message for AI stock investors
- Microsoft CEO sends a blunt warning on AI and the tech ecosystem
- The next AI infrastructure race has nothing to do with chips
Karp argues that large language models alone aren’t enough.
In battlefield, manufacturing, clinical, and regulated settings, he said companies need an application layer that makes AI “safe and useful and precise.”
That is exactly where Palantir has the edge, especially in Ontology, the layer that keeps models useful without letting sensitive data, prompts, or business logic leak outside the enterprise.
The bigger concern, in Karp’s telling, is ownership.
“Who owns the data?” he asked. “Where is it cached? Are the prompts secure?” Those questions turn the AI debate from a performance race into a control issue.
In essence, Karp challenged the entire token-based AI model.
If businesses feel they are paying for usage while risking their proprietary edge, Palantir can argue that its value lies not just in access to AI but in control of it.
Those concerns aren’t theoretical, either.
Cisco’s 2025 Data Privacy Benchmark found that 60% of respondents worry GenAI inputs could be shared with the public or competitors, while 58% worry the tools could harm a company’s legal rights or IP.
IBM separately found 97% of organizations with an AI-related security incident lacked proper AI access controls.
Kevin Dietsch/Getty Images
Palantir stock price-target split widens
- Wedbush: $230: Dan Ives kept an outperform rating, arguing Palantir remains a premium AI software asset, despite the pullback.
- Rosenblatt: $225: Rosenblatt reiterated buy, backing Palantir’s Ontology platform as a durable moat in enterprise AI.
- Loop Capital: $220: Loop stayed bullish after Q1, citing AI-driven revenue growth and U.S. revenue up 104% year over year.
- Morgan Stanley: $205: Morgan Stanley pointed to Palantir’s 10th straight quarter of accelerating revenue and raised forecasts.
- Consensus check: MarketWatch shows a $189.87 average, $200 median, $255 high, and $70 low target.
Sources: Wedbush/MarketBeat, Rosenblatt, Capital.com, Business Insider, MarketWatch
Why the Nvidia deal gives Palantir’s warning more weight
Karp went on to discuss the new Nvidia deal, saying it is exactly what Palantir’s technical customers are most looking for: “control over their compute, their models, their data stack, and their alpha.”
For some context, Palantir and Nvidia recently announced a deal that centers on an “intelligent engine” that lets U.S. government agencies and critical-infrastructure operators run Nvidia AI and Nemotron open models in sovereign, classified, air-gapped, or sensitive environments.
Nvidia described it as using Nemotron open models to deliver mission-specific, sovereign AI for government and critical infrastructure customers.
Nvidia brings the AI platform, compute, and open models, while Palantir brings AIP, Ontology, Foundry, and Apollo, the software layer designed to enforce authorization, auditability, and operational control.
In many ways, the partnership is a direct answer to Karp’s warning.
If companies are worried about where prompts are cached, who controls model weights, and whether proprietary insights migrate into closed models, Palantir is pitching itself as the control layer that keeps AI useful without giving away the crown jewels.
What control of the AI stack means for investors
For investors, Karp’s argument effectively reframes Palantir’s AI story from software demand to pricing power.
Interestingly, I covered UBS’s Karl Keirstead’s comments, who also pointed to Palantir’s pricing power, driven by its profitability and moat. He noted Palantir stock trading at 46x 2027 estimated FCF, but it looks undervalued compared to its 55% three-year CAGR.
That strength will only grow if enterprise AI becomes a question of control, and the competition goes beyond model access.
Karp argues that Palantir’s demand will likely sit in the layer between raw models and real business operations.
It also gives Palantir a way to defend premium valuation multiples if its Ontology becomes a required control system for AI deployment.
For perspective, Palantir stock is trading at over 85 times forward non-GAAP earnings, which is 240% higher than the sector median, according to Seeking Alpha.
Though it’s changing hands at a sizeable discount to its five-year average, it still carries a lofty valuation.
However, the risk is all about execution.
Investors need proof that this trust argument translates into larger contracts, faster commercial adoption, and sustained margin strength.
Related: Microsoft reportedly makes another brutal workforce move