Hajra Naz
OpenAI just made one of its boldest moves yet. The company announced it will buy 6 gigawatts of computing power from data centers powered entirely by AMD chips — a direct challenge to Nvidia’s dominance in the AI hardware market.
This massive deal could reshape the future of AI computing.
A Big Shift in the AI Chip Race
For years, Nvidia (NVDA) has been the clear leader in AI chips. Its GPUs power everything from ChatGPT to self-driving cars, helping the company hit a $4.6 trillion market cap — making it the world’s most valuable public company.
Now, OpenAI is shaking things up. It’s betting big on AMD (AMD), whose Instinct processors are gaining serious traction in high-performance computing.
The Deal Breakdown
The agreement gives OpenAI access to 1 gigawatt of power by next year and another 5 gigawatts in later phases. These systems will run on AMD’s next-generation Instinct chips, co-developed with OpenAI to boost AI model training speed and energy efficiency.
OpenAI will also invest heavily in AMD, taking a 10% ownership stake worth billions. AMD issued OpenAI a warrant for 160 million shares, valued at over $26.3 billion, which will vest as key goals are hit — including chip performance milestones and market targets.
Market Reaction
The market reacted fast. AMD shares jumped 36% in premarket trading, while Nvidia shares dipped 2%.
Analyst Dan Ives from Wedbush called it a “massive vote of confidence” for AMD and CEO Lisa Su, saying the move places AMD “right at the heart of the AI spending cycle.”
AMD expects the deal to generate tens of billions in revenue, though the exact number remains undisclosed.
Altman’s Vision: Building the World’s Largest AI Infrastructure
OpenAI CEO Sam Altman said this deal is part of building “the world’s most ambitious AI buildout.”
He added:
“AMD’s leadership in high-performance chips will help us accelerate AI progress and bring these benefits to everyone faster.”
Lisa Su called it a “win-win” — strengthening AMD’s position and fueling OpenAI’s rapid growth.
AI Spending is Exploding
OpenAI isn’t slowing down.
Just last month, it signed a $300 billion deal with Oracle for 4.5 gigawatts of data center capacity. It’s also reportedly partnering with Broadcom in a $10 billion chip design project aimed at reducing dependency on Nvidia’s hardware.
Experts say the AMD partnership is part of OpenAI’s strategy to diversify its chip supply, ensuring it doesn’t rely too heavily on a single vendor.
“AMD is another major player in this space, so this move makes perfect sense for OpenAI,” said Ben Barringer, a global tech analyst at Quilter Cheviot.
AI Giants Dominate Wall Street
Today’s top tech companies are all betting on AI.
The eight most valuable companies—Nvidia, Microsoft, Apple, Google, Amazon, Meta, Broadcom, and Tesla—are worth over $1.4 trillion each and are heavily invested in AI innovation.
From AI PCs to video generators like Sora 2, these companies are shaping the next wave of digital transformation.
But not everyone is convinced.
Some Wall Street analysts warn of an AI bubble, arguing that soaring valuations are driven more by FOMO than fundamentals.
Still, with deals like OpenAI’s new AMD partnership, the race for AI dominance is far from slowing down.
Read More: OpenAI Introduces Sora, a TikTok Competitor, with Its Latest Sora 2 Model
FAQs
1. Why did OpenAI partner with AMD?
OpenAI is expanding beyond Nvidia to secure more chip supply and boost computing power for its growing AI models.
2. How big is the deal?
The partnership is worth over $26 billion, including stock warrants and future investments tied to AMD’s chip milestones.
3. Will OpenAI stop working with Nvidia?
No. Sam Altman said the AMD deal is “incremental” to its ongoing partnership with Nvidia, not a replacement.
4. How will this affect the AI industry?
The deal intensifies competition in the AI chip market, pushing AMD into a stronger position against Nvidia while fueling AI infrastructure growth.
5. Is the AI boom sustainable?
Some experts think it could be a bubble driven by hype, but others say the demand for AI hardware and computing power is only getting started.
