Anthropic partners with Samsung to develop custom AI chip
Jakarta (ANTARA) - The artificial intelligence (AI) company Anthropic is reportedly exploring a strategic partnership with Samsung Electronics to produce custom AI chips.
This move marks intensifying competition within the global semiconductor industry amidst the surging demand for AI-based computing.
According to a report by technology media GSMArena on Saturday, both companies are currently in the early stages of negotiations. If an agreement is reached, Samsung would be tasked with manufacturing specialised chips designed to support the development of Anthropic’s generative AI models.
This collaboration reflects a significant shift in the semiconductor industry. The continuous rise in demand for AI chips is prompting many technology companies to seek alternative suppliers outside of Taiwan Semiconductor Manufacturing Company (TSMC), which has long dominated the advanced chip manufacturing market.
This situation presents a fresh opportunity for Samsung and Intel Corporation. Furthermore, recent reports suggest that Apple Inc. is reconsidering Intel as one of its chip suppliers.
Samsung is also said to be in talks with Apple regarding potential cooperation within the semiconductor sector.
Despite the plans to develop custom chips, Anthropic emphasised that its data centre operations will continue to rely on several key technology partners in the near future. The company remains a user of Amazon Web Services’ Trainium chips, Google’s Tensor Processing Units (TPU), and graphics processing units (GPU) from NVIDIA Corporation.
However, the increasing complexity of AI models means that Anthropic requires more specific and efficient hardware to enhance its future computing capacity.
The trend of developing custom AI chips is also being pursued by other major players. OpenAI, for instance, is reportedly negotiating with Broadcom Inc. to develop its own AI semiconductors.
The steps taken by Anthropic and Samsung demonstrate that the AI race is no longer confined to the development of language models, but has expanded into the competition for the chip technologies that will serve as the fundamental foundation of the future artificial intelligence industry.