Venture capitalist Chamath Palihapitiya has publicly warned that Anthropic risks falling behind in the artificial intelligence race after its Claude chatbot declined to complete a financial analysis task that rival systems handled without issue. The incident highlights a growing tension in the AI industry between safety-focused design choices and practical utility, with major investors questioning whether overly cautious guardrails could cost Anthropic market share.

What Happened in Palihapitiya's AI Model Test?

On Saturday, Palihapitiya posted results from a comparative test of four leading artificial intelligence systems. He ran what he described as a "semi-complicated stock screening prompt" across xAI's Grok, Google's Gemini, OpenAI's ChatGPT, and Anthropic's Claude to see how each would handle the financial analysis task.

On Saturday

The first three systems produced comparable results and completed the work as requested. Claude, however, refused to perform the task. Palihapitiya's public criticism of this outcome sparked a broader conversation about how AI companies balance safety considerations with user expectations for functionality. His comparison to Friendster, the social networking platform that once dominated its market before losing relevance, underscores the stakes he sees in this competitive moment.

"Anthropic needs to solve the computer/power problem or they will be the Friendster of the AI era," warned Palihapitiya.

Chamath Palihapitiya, Venture Capitalist

How Are AI Companies Competing for Computing Resources?

The tension Palihapitiya highlighted reflects a deeper competitive dynamic in the AI industry. Anthropic has been aggressively expanding its infrastructure and partnerships to keep pace with rivals. The company recently signed a $1.8 billion cloud computing deal with Akamai Technologies, a major investment aimed at securing the computational resources needed to train and run increasingly powerful AI models.

In a significant development, Anthropic also gained access to SpaceX and xAI's Colossus 1 supercomputer, marking a notable shift in the competitive landscape. This partnership underscores how critical raw computing power has become to AI development, even as tensions between Anthropic and Elon Musk have existed in the past.

Beyond infrastructure, Anthropic has been expanding Claude's capabilities through new features and integrations. The company rolled out Claude Cowork, a platform that now includes specialized legal tools integrating services such as Thomson Reuters' Westlaw, CourtListener, Definely, Courtroom5, Box, and Harvey. These additions enable law firms to automate research, contracts, and complex workflows more efficiently.

Steps to Understanding the Broader Implications for AI Users

• Compute Access Matters: Companies with access to powerful supercomputers like Colossus 1 can train larger, more capable models faster than competitors with limited resources, directly affecting which AI systems can handle complex tasks.
• Safety vs. Utility Trade-offs: Different AI companies make different choices about what tasks their models will refuse, balancing safety concerns against user expectations for practical functionality in areas like finance and data analysis.
• Enterprise Integration Drives Adoption: Partnerships with established platforms like Microsoft Office, legal research tools, and cloud providers help AI companies embed their systems into workflows where users already work, increasing stickiness and revenue.
• Usage Growth Signals Market Momentum: Anthropic reported "80x growth" in usage and revenue, indicating strong demand for Claude despite competitive pressures, though this growth must be sustained through continued capability improvements.

Palihapitiya's critique touches on a real strategic question facing Anthropic. While the company has made significant progress, expanding partnerships and securing computing resources, the incident with the stock screening prompt suggests that design decisions around what Claude will and won't do may be limiting its appeal in certain professional contexts. This is particularly important as enterprises evaluate which AI systems to integrate into their workflows.

The competitive landscape is intensifying rapidly. Grok, xAI's chatbot, has been gaining attention for its willingness to engage with a broader range of requests compared to some competitors. Google's Gemini and OpenAI's ChatGPT continue to set benchmarks for capability and user experience. For Anthropic, the challenge is clear: maintain the trust and safety standards that differentiate Claude while ensuring the system remains practical enough for professional users who need it to handle real-world tasks.

The broader lesson from this moment is that in the AI industry, competitive advantage increasingly depends on three factors: access to computing power, thoughtful design choices that balance safety with utility, and strategic partnerships that embed AI systems into existing enterprise workflows. Anthropic has made progress on all three fronts, but Palihapitiya's warning suggests the company cannot afford to fall behind on any of them.