Meta tests self-developed AI training chips, attempting to reduce reliance on NVIDIA

Wallstreetcn
2025.03.12 01:36
portai
I'm PortAI, I can summarize articles.

Meta is testing its self-developed AI training chip, aiming to reduce reliance on NVIDIA. The chip is manufactured by TSMC, with plans to scale production after successful preliminary testing. Meta's previous chip projects have repeatedly fallen short of expectations, and the success of this project still faces challenges. If successful, Meta could lower costs, achieve financial victories, and reflect the tech giant's pursuit of self-developed AI hardware

Meta is actively testing its self-developed AI training chips, taking a key step in reducing reliance on hardware suppliers such as NVIDIA.

According to a report by Reuters on the 11th, Meta has begun small-scale deployment to test its self-designed AI training chips, which are specifically designed to handle AI workloads and are manufactured by TSMC, with plans to scale up production after initial testing is successful.

It is worth noting that Meta has previous experience deploying custom AI chips, but those chips were only used for running models, not training them. The report states that several of Meta's past chip design projects were canceled or scaled back due to failing to meet internal expectations, and the success of this new project still faces challenges.

For Meta, self-developed AI training chips have significant strategic and financial implications. It is reported that Meta previously expected its capital expenditures to reach $65 billion this year, most of which would be used to purchase NVIDIA's GPUs.

This means that if Meta can reduce even a small portion of costs by shifting to self-developed chips, it would be a tremendous financial victory for the social media giant.

Meta's move also reflects the general pursuit of self-developed AI hardware capabilities among tech giants. In the current market environment where NVIDIA GPUs are in short supply, having independent research and development capabilities can effectively reduce supply chain risks while providing greater flexibility and cost advantages for the company's AI strategy.

Risk Warning and Disclaimer

The market has risks, and investment should be cautious. This article does not constitute personal investment advice and does not take into account the specific investment goals, financial situation, or needs of individual users. Users should consider whether any opinions, views, or conclusions in this article align with their specific circumstances. Investing based on this is at your own risk