Palantir CEO says Israel is the only minor exception to U.S.-China AI dominance

The two-player global race

Palantir CEO Alex Karp recently described the global competition for artificial intelligence as a contest dominated almost entirely by the United States and China. In a conversation with Citigroup’s global chairman of banking, Karp stated that these are the only relevant countries in the current landscape. He characterized every other nation as a "weird derivative" that lacks the power to compete on a global scale.

While Karp did single out Israel as a "minor exception" to this rule, he provided no specific details on why the country qualifies. He did not clarify if this status stems from Israel's military capabilities, its cybersecurity industry, or its startup ecosystem. The remark suggests that while Israel holds a unique position, it does not fundamentally shift the primary power dynamic between Washington and Beijing.

Karp noted that the competition is no longer just about individual models but about controlling the entire technology stack. This includes the physical chips, the software that runs on them, and the systems that connect corporate data to AI models. He argued that most of this infrastructure currently exists only in the United States or China, making it difficult for other regions to build independent capabilities.

U.S. and Chinese advantages

The CEO provided a brief assessment of the current standings between the two superpowers. He estimated that the United States currently holds a lead of about six months over its primary rival. According to Karp, the American advantage is rooted in creativity, while China's primary strength lies in its massive scale.

China has developed specific techniques to remain competitive despite different constraints. Karp highlighted China's ability to run models with less computing power and offer lower prices. He also pointed to a technique called distillation, which involves training smaller models on the outputs of larger ones to achieve efficiency.

Despite these differences, Karp believes the core of the race remains a struggle for dominance. He emphasized that the goal for both nations is to deploy infrastructure without forcing users to surrender their intellectual property. This focus on data sovereignty is a key theme in how both powers are approaching the commercial and governmental deployment of AI.

The predicament of Europe

Karp offered a blunt assessment of Europe's role in this technological shift. He suggested that the continent is likely going to have to purchase AI products from either the United States or China. He described this prospect as regrettable, noting that Europe currently lacks the independent infrastructure to offer a third alternative.

The CEO expressed a personal desire to help Europe, citing his own background and time spent on the continent. However, he stated that "you can’t help someone who doesn’t want to be helped," implying a lack of initiative from European leaders. This suggests a looming era of technological dependency for European businesses and governments.

This dynamic creates a significant hurdle for European companies trying to maintain autonomy. If the continent cannot develop its own stack, it may find itself choosing between two different sets of standards and security protocols. Karp's comments underscore the pressure on European policymakers to decide which side of the AI divide they will align with.

Wealth, taxes, and security

Beyond the technology itself, Karp discussed the political risks associated with AI-driven wealth. He argued that AI companies are essentially wealth creation mechanisms, which could lead to significant domestic political friction. He warned that mounting inequality might fuel a populist push for a wealth tax, which he called "crazy" because it could shut off innovation and kill startups.

Karp used his own potential financial gains as an example of this societal problem. He noted that enforcing a wealth tax would require valuing assets in real time and could expose private details of individuals' lives. He mentioned that Palantir's technology could potentially perform such tracking, though he stated he does not want the company involved in such an effort.

In the corporate world, Karp criticized the current way companies pay for AI tokens without seeing meaningful value. He noted that some executives are frustrated by paying for services that do not produce results or that inadvertently migrate their intellectual property to competitors. To solve this, he promoted Palantir's "sovereign" AI, which allows companies to keep their proprietary knowledge under their own control.

The evolving cyber landscape

The CEO also warned about the increasing sophistication of AI in the realm of cybersecurity. He noted that modern AI models are "excruciatingly adept" at identifying security vulnerabilities in software. This capability is leading to a state of "more guerrilla war" in the cyber space, where the winner is simply the side that can patch holes the fastest.

Karp emphasized that no system can offer 100% protection against these automated threats. As AI makes it easier for attackers to find weaknesses, the speed of defense becomes the primary metric of success. This creates a constant arms race where security is a moving target rather than a static goal.

Palantir's work in this area is often conducted under serious security constraints. Karp mentioned the company's role in Ukraine and in certain parts of Israel, where it helps drive the transformation of institutions. While he did not specify the exact institutions, his comments highlight the high-stakes nature of Palantir's involvement in global security and defense.