Anthropic signs supply deal with Samsung, SK hynix amid custom AI chip race
Samsung and SK hynix are reportedly providing major AI infrastructure to Anthropic, but information surrounding the recent supply deal details has not been revealed.
- Jul 27, 2026,
- Updated Jul 27, 2026 10:39 AM IST
Anthropic, the US-based AI company, has entered into supply agreements with Samsung Electronics and SK hynix. Anthropic CEO Dario Amodei revealed the news at the AI summit in South Korea, although he did not reveal details of the agreement, including what components or services are to be supplied, the deal value, or how long the deal will last.
Samsung and SK hynix are major semiconductor giants, but each company offers different strengths in memory and chip technologies. Previously in May 2026, Anthropic raised funding at a post-money valuation of $965 billion, and publicly revealed its infrastructure partnership with Samsung, SK hynix, and Micron, naming them as strategic technology partners supporting its AI infrastructure.
Must read: US claims China's Moonshot AI used Anthropic's Fable 5 to build Kimi K3 AI model
But there is no information publicly available on whether the deals include HBM (High Bandwidth Memory) chips, memory chips, storage products, logic chips or semiconductor components.
According to a Reuters report, SK Group Chairman Chey Tae-won also revealed that Anthropic was creating its own "computing power", but he did not explain what he meant. Previously, it was highlighted that Anthropic is exploring the possibility of designing its own AI chips instead of relying entirely on processors from companies like Nvidia or AMD.
Anthropic has not confirmed its chip-designing efforts, but it said that it is continually exploring different computing options. However, it currently uses a variety of chips from different suppliers as part of a diversified computing strategy, rather than depending on a single chipmaker.
Must read: Anthropic wins approval for $1.5 billion copyright settlement
In addition to Anthropic, its competitors like OpenAI, Google, Amazon AWS, and others are aggressively developing in-house custom AI chips to reduce their dependency on dominant suppliers like Nvidia.
OpenAI partnered with semiconductor designers like Broadcom to develop custom silicon and signed multi-gigawatt computing infrastructure. Google continues to lead custom silicon efforts with its proprietary Tensor Processing Units (TPUs). Amazon is scaling its custom chip division, which includes Trainium for AI training and Inferentia for inference tasks.
For Unparalleled coverage of India's Businesses and Economy – Subscribe to Business Today Magazine
Anthropic, the US-based AI company, has entered into supply agreements with Samsung Electronics and SK hynix. Anthropic CEO Dario Amodei revealed the news at the AI summit in South Korea, although he did not reveal details of the agreement, including what components or services are to be supplied, the deal value, or how long the deal will last.
Samsung and SK hynix are major semiconductor giants, but each company offers different strengths in memory and chip technologies. Previously in May 2026, Anthropic raised funding at a post-money valuation of $965 billion, and publicly revealed its infrastructure partnership with Samsung, SK hynix, and Micron, naming them as strategic technology partners supporting its AI infrastructure.
Must read: US claims China's Moonshot AI used Anthropic's Fable 5 to build Kimi K3 AI model
But there is no information publicly available on whether the deals include HBM (High Bandwidth Memory) chips, memory chips, storage products, logic chips or semiconductor components.
According to a Reuters report, SK Group Chairman Chey Tae-won also revealed that Anthropic was creating its own "computing power", but he did not explain what he meant. Previously, it was highlighted that Anthropic is exploring the possibility of designing its own AI chips instead of relying entirely on processors from companies like Nvidia or AMD.
Anthropic has not confirmed its chip-designing efforts, but it said that it is continually exploring different computing options. However, it currently uses a variety of chips from different suppliers as part of a diversified computing strategy, rather than depending on a single chipmaker.
Must read: Anthropic wins approval for $1.5 billion copyright settlement
In addition to Anthropic, its competitors like OpenAI, Google, Amazon AWS, and others are aggressively developing in-house custom AI chips to reduce their dependency on dominant suppliers like Nvidia.
OpenAI partnered with semiconductor designers like Broadcom to develop custom silicon and signed multi-gigawatt computing infrastructure. Google continues to lead custom silicon efforts with its proprietary Tensor Processing Units (TPUs). Amazon is scaling its custom chip division, which includes Trainium for AI training and Inferentia for inference tasks.
For Unparalleled coverage of India's Businesses and Economy – Subscribe to Business Today Magazine