Anthropic, the developer of the Claude AI model, is intensifying its efforts to develop specialized processors, seeking greater autonomy over its technological infrastructure. In this context, the company held talks to acquire the startup MatX for approximately $7 billion, according to Reuters sources.
From Acquisition to Partnership
Despite the initial discussions regarding a full acquisition, more recent reports suggest that the talks have shifted toward a potential strategic partnership. The reasons why the acquisition is no longer in an active stage remain unknown, and neither company has provided an official comment.
MatX, founded by former Google engineers who worked on TPU (Tensor Processing Units) technology, specializes in designing chips for training large language models. The expertise of its team is considered critical for accelerating Anthropic's plans for proprietary hardware.
Strategic Goals and Market Context
Anthropic's move is part of a broader strategy to reduce long-term costs and improve the performance of its models. The company is exploring the development of processors for both training and inference, without abandoning its current suppliers.
- Anthropic continues to utilize hardware from Nvidia, Google, and Amazon.
- MatX is reportedly seeking new funding at a valuation of approximately $4 billion.
- Strengthening its internal chip design team remains a priority for Anthropic.
The shift toward specialized hardware highlights the need for leading AI players to gain full control over the value chain, from code to silicon.
