Palantir CEO Alex Karp has sharply criticized the developments of OpenAI and Anthropic, calling them "trash" in the context of enterprise tasks. Amidst Palantir's impressive 85% revenue growth, the industry is demonstrating a fundamental shift from simple chatbots to complex autonomous agents capable of executing long-running work cycles.

image
image

What Happened

During an investor call, Palantir CEO Alex Karp stated that OpenAI and Anthropic products lack applied value for serious business due to a lack of deep integration. Simultaneously, Palantir reported an 85% revenue growth in the first quarter of 2026. In the financial sector, JPMorgan Chase has already announced plans to implement AI agents capable of working autonomously for many hours, allocating approximately $20 billion annually to technology development.

Context

The modern AI market is undergoing a transition from the era of "smart conversationalists" to the era of specialized agentic architectures. While research companies focus on the power of base models, the corporate sector demands reliability, controllability, and integration with existing data. Parallel to technological growth, social tension is rising: polls show that 61% of Democrats and 47% of Republicans in the US fear job losses due to automation.

Why It Matters for the Industry

There is active competition between base LLM developers and creators of vertical AI platforms. For the industry, the key technological barrier in 2026 will be the transition from short chat sessions to supporting long-running autonomous cycles. This is stimulating investment growth in agentic infrastructure and agent lifecycle management tools, and is forcing giants like OpenAI to expand their field engineering staff to compete with applied players.

Why It Matters for Users

For users and workers, this means a change in the nature of labor: the focus is shifting from direct task execution to controlling AI agents (the transition from human-in-the-loop to human-on-the-loop). At the same time, automation primarily threatens administrative and operational roles, such as accountants, HR, and data managers, which could lead to structural unemployment in the information processing specialist segment.

What Remains Unknown / Limitations

There are legal risks and questions regarding accountability for the actions of fully autonomous agents, as well as technical difficulties in ensuring observability and evaluation (evals) for long-running AI work sessions.

Sources

Author

Look at AI, Editorial Team