BY COMFORT OGBONNA
Meta Platforms is deepening its push into artificial intelligence with a major new agreement to purchase advanced AI chips from Advanced Micro Devices, in a deal that could ultimately exceed $100 billion in value. As part of the arrangement, Meta will also gain the opportunity to acquire up to a 10% stake in the chipmaker, further aligning the two companies as competition in the AI sector intensifies.
The agreement calls for Meta to purchase AMD’s newest generation AI chips, known as the MI450, which will be used to power the company’s expanding network of AI-focused data centers. The partnership covers a total of 6 gigawatts of computing capacity, with shipments supporting the first 1-gigawatt deployment expected to begin in the second half of this year. Additional shipments will follow as Meta scales up its infrastructure to meet growing AI demands across its platforms.
The announcement comes just days after Meta unveiled a separate long-term partnership with Nvidia, under which it plans to use millions of Nvidia chips and related hardware to support its artificial intelligence operations. By working with both AMD and Nvidia, Meta appears to be diversifying its chip supply while ensuring it has the computing muscle required to train and operate increasingly sophisticated AI systems.
AMD, long a rival to Nvidia in the semiconductor industry, has been working aggressively to close the gap in the AI hardware race. Nvidia established an early advantage by adapting its graphics processing units, or GPUs, originally designed to power video games, into high-performance chips capable of training and running advanced AI models. These GPUs now underpin widely used technologies such as chatbots and image generators, including systems like ChatGPT.
The surge in demand for AI tools over the past two years has fueled an unprecedented appetite for high-powered chips. Technology companies have poured billions of dollars into building data centers and acquiring the hardware needed to train large language models and deploy AI services at scale. However, investors and analysts have also raised questions about whether companies are overspending and whether these massive capital investments will translate into sustainable profits and productivity gains in the long run.
For Meta, the stakes are particularly high. The company, which owns major social media platforms, has been racing to strengthen its AI capabilities amid intense competition from rivals such as Google and OpenAI. In June, Meta invested $14.3 billion in AI data company Scale AI and brought its chief executive, Alexandr Wang, onboard to help lead a team focused on developing what it describes as “superintelligence.” Later in December, the company acquired AI startup Manus as part of a broader strategy to expand AI-driven features across its ecosystem, including Instagram and other services.
As part of the AMD agreement, the chipmaker issued Meta a performance-based warrant covering up to 160 million shares of its common stock at a nominal price of $0.01 per share. The warrant is structured to vest in stages as specific milestones are achieved. The first portion will vest once the initial 1-gigawatt of chip shipments is completed, with additional tranches becoming available as Meta’s total purchases scale toward the full 6-gigawatt target.
Investors responded enthusiastically to the announcement. Shares of AMD surged more than 9% in premarket trading Tuesday, reflecting optimism that the partnership could significantly boost the company’s revenue and strengthen its position in the rapidly evolving AI hardware market.
The deal underscores the extraordinary scale of investment flowing into artificial intelligence infrastructure and highlights how major technology firms are locking in long-term chip supplies as they compete to shape the next era of computing.
Original article


Comments