In the rapidly evolving world of artificial intelligence, Anthropic is making significant strides to secure its position as a leader in AI hardware development. The company has been in discussions to acquire MatX an AI chip startup, for approximately $7 billion. This move is part of Anthropic’s broader strategy to accelerate the development of custom hardware tailored to its AI models.
The talks with MatX, founded by former Google tensor processing unit (TPU) engineers, highlight Anthropic’s ambition to build in-house expertise in chip design. This initiative aims to enhance the performance and efficiency of the Claude family of AI models, reducing the company’s dependence on third-party chip providers like Nvidia.
Anthropic’s Strategic Moves in AI Chip Development
Anthropic’s plans to develop custom chips are driven by the need to meet the voracious appetite for data crunching required by its AI models. The company is expected to go public this year, with a valuation target of $2 trillion, hinging on a projected revenue of up to $200 billion by 2028. To achieve this, Anthropic is expanding its in-house silicon team and exploring partnerships with various chip startups.
Designing custom chips is a complex and costly endeavor. It can take a year or more to produce a viable piece of hardware, with design costs running into the hundreds of millions of dollars. However, acquiring a startup like MatX could provide Anthropic with the necessary expertise to lower costs and accelerate the development process. MatX has been working on a chip designed for building large AI models, a critical component in the training process.
Partnerships and Investments in AI Infrastructure
Anthropic’s aggressive scaling efforts include significant investments in computing power. The company has signed a $45 billion deal with Nscale to rent AI compute, utilizing Nvidia’s state-of-the-art Vera Rubin chip system. This deal, spanning six years, will provide compute capacity starting in late 2027. Additionally, Anthropic has secured deals with SpaceXAMDAmazon and Google to expand its computing capacity.
The company’s strategy includes a multi-chip approach, working with providers ranging from Nvidia to Google. This diversification helps mitigate risks associated with the tight supply of Nvidia’s processors, which are expected to remain in short supply through 2027. By developing custom chips, Anthropic aims to create significant performance and economic advantages, similar to OpenAI’s recent success with its custom chip, Jalapeno.
Talent Acquisition and Industry Trends
Anthropic has been actively hiring engineering and executive talent to accelerate its chip design process. Recent hires include Google chip veteran Amir Salek and former OpenAI chip engineer Clive Chan. These acquisitions underscore the company’s commitment to building a robust in-house team capable of designing cutting-edge AI hardware.
The leading AI labs, including Anthropic and OpenAI, are increasingly focused on custom chips tailored to their specific models and workloads. This trend is driven by the need for greater efficiency and performance in AI computations. By developing custom chips, these companies hope to gain a competitive edge in the rapidly growing AI market.
As Anthropic continues to explore partnerships and acquisitions in the AI chip space, its strategic moves are likely to shape the future of AI hardware development. The company’s ambitious plans reflect its commitment to innovation and its determination to lead the AI revolution.
