In a major move to fuel its artificial intelligence expansion, a leading social media company has entered into a multiyear agreement with a prominent chipmaker to purchase AI graphics processing units (GPUs) equivalent to a massive 6 gigawatts of power. The deal, announced on a Tuesday, sent the chipmaker’s stock surging as much as 10% in premarket trading.
The agreement includes a performance-based structure where the chipmaker will issue 160 million shares of common stock to the tech firm. These shares will vest in stages as the chipmaker hits specific milestones, the first of which is triggered upon shipping the initial gigawatt of chips.
“We expect this partnership to drive substantial multi-year revenue growth,” a representative from the chipmaker stated. “The performance-based structure also tightly aligns our companies around execution and long-term value creation.”
The first GPUs to be deployed, from the chipmaker’s MI450 line, will be installed in large-scale data center systems in the second half of the year. The tech company also confirmed it will purchase a significant number of additional central processing units (CPUs), highlighting the growing importance of these components for running advanced AI services.
This agreement comes just a week after the same tech company announced a separate, multiyear deal with the chipmaker’s primary rival, which will supply millions of its latest GPUs and host the first large-scale deployment of its own CPU servers.
These enormous purchases are part of a broader spending spree. The social media giant alone is projected to spend over $135 billion in 2026 on its AI build-out, covering everything from data center construction to chip acquisition and model training. It is not alone; several major technology firms plan to collectively invest an estimated $650 billion in AI this year.
These staggering figures have caused some investor concern about whether the massive investments will pay off. Since announcing their spending plans, stock performance has varied significantly among the top companies, with one software giant’s shares declining more than 15%. The chip sector has also cooled as questions about a potential AI bubble and competition from custom, in-house chips developed by the tech giants themselves create uncertainty.
Despite reports of the social media company exploring alternative chip options, analysts suggest it will be difficult for such custom solutions to fully replace the general-purpose AI processors supplied by established chipmakers.
