US Anthropic Builds In-House AI Chip Design Team

en.Wedoany.com Reported - Anthropic has confirmed it will develop its own AI chips, planning to co-design hardware and models to support its Claude technology running faster and more efficiently. The news was first reported by Business Insider, and Anthropic subsequently confirmed the plans to tech media outlet TechCrunch.

Signs of collaboration in the chip manufacturing stage had already emerged earlier. The Information reported last month that Anthropic was in talks with Samsung to explore the company as a potential contract manufacturer for such chips.

Anthropic made this decision against the backdrop of rising demand for Claude. As AI companies are currently snapping up as many AI infrastructure deals as possible, Anthropic has reached agreements with Amazon Web Services (AWS), Google, Nvidia, and AMD to secure AI computing hardware. However, to truly scale up and meet current demand levels, relying solely on external suppliers is clearly insufficient.

Anthropic is not the first AI company to pursue an in-house chip development path. In June this year, OpenAI released the Jalapeño chip, built by Broadcom, which is specifically designed for inference workloads. Google DeepMind has long relied on Alphabet's TPU chips to power its AI models, while Meta has been developing its own MTIA accelerators for AI workloads.

To advance its in-house chip initiative, Anthropic has posted job listings, seeking engineers with chip design experience for its "custom chip team."