Jakarta, INTI - American artificial intelligence (AI) company Anthropic is exploring the possibility of designing its own chips. Based on Reuters reports on Saturday, April 11, 2026, several sources said the plan was still in its early stages and could be scrapped.
According to one source, the company has not yet fully committed to a specific design or made a specialized team for the project.
Demand for Anthropic's Claude AI model has grown rapidly in 2026, with the startup's annual revenue now surpassing USD 30 billion (approximately Rp 512.9 trillion), up from around USD 9 billion (approximately Rp 153.8 trillion) at the end of 2025.
Anthropic uses several chip types, including tensor processing units (TPU) designed by Alphabet’s Google, to develop and operate its AI and Claude chatbot software.
Earlier this week, Anthropic signed a long-term agreement with Google and Broadcom, collaborating on TPU development. Aside from Anthropic, other technology companies, such as Meta and OpenAI, are also planning to design their own AI chips.
Conclusion
Anthropic is reportedly interested in making its own AI chip design. However, the plan is still in early stages, and there are no specific designs or teams made. Anthropic uses several chip designs, including TPU from Google. Meta and OpenAI also have similar plans with Anthropic.
Read more: South Korea's Government is Planning to Introduce AI Assistance in Administrative Workflows