Meta has struck a deal to purchase tens of millions of Amazon's Graviton chips to power its next-generation AI models, a move driven by the ongoing scarcity of Nvidia processors that have become central to large language model development.
The agreement marks a significant shift in Meta's hardware strategy as the company seeks alternative suppliers to reduce its dependence on Nvidia chips, which remain in limited supply despite massive demand across the AI industry.
Graviton chips, Amazon's custom-designed processors, will complement Meta's existing infrastructure as the company scales up AI capabilities. The deal underscores broader industry trends of tech giants developing or securing alternative chip sources to maintain competitive advantage in the AI race.
Meta has invested heavily in AI infrastructure, including its own chip development initiatives. The addition of Graviton processors provides flexibility in sourcing while the company continues expanding computational capacity for training and deploying advanced models.
The shortage of Nvidia chips has forced major AI players to diversify their hardware portfolios. Other companies have similarly pursued custom chips or alternative processors to sustain their AI development timelines.
Amazon has positioned Graviton as a cost-effective and efficient alternative for cloud workloads. The chips have gained traction among AWS customers seeking to reduce computing costs while maintaining performance.
Meta's purchase of tens of millions of units signals confidence in Graviton's capabilities for AI workloads. The scale of the deal reflects the massive computational requirements driving modern AI model development.
The agreement also highlights Amazon's expanding role in the semiconductor market, leveraging custom silicon to strengthen its AWS business and secure major customers.
Neither company disclosed financial terms or specific deployment timelines. The deal is expected to begin delivering chips in coming quarters as Meta ramps up its AI infrastructure.
The European Union is considering sweeping restrictions on young people's social media access, including potential age limits and outright bans. European Commission President Ursula von der Leyen said new legislation could be proposed within months.
Meta is committing an additional $40 billion to its Louisiana data center campus, bringing total expected investment beyond $250 billion as the company expands its AI computing infrastructure.
Washington is considering restrictions to prevent Chinese companies from using American AI models to train their own systems, reviving longstanding debates over tech export controls and intellectual property.
MiniMax Group Inc.'s stock plummeted following JPMorgan Chase's second price target reduction in under a week. The downgrade cites value dilution concerns from the Chinese AI model maker's recent fundraising round.