Hajra Naz
OpenAI is teaming up with Broadcom to design and build 10 gigawatts of custom AI chips and systems. That’s an enormous amount of power, roughly the same as what a large city uses. This move shows just how energy-hungry the global AI boom has become.
A Major Step in AI Infrastructure
This new deal is the latest in a series of high-profile partnerships for OpenAI. The company has already worked with Nvidia and AMD to secure more computing power for its fast-growing user base.
Today, ChatGPT has over 800 million weekly users, and an OpenAI executive recently said that the new Sora video generation app is growing even faster than ChatGPT did.
“Partnering with Broadcom is a critical step in building the infrastructure needed to unlock AI’s potential and deliver real benefits for people and businesses,” said Sam Altman, OpenAI’s co-founder and CEO, in the press release announcing the deal.
Deployment Set for 2026
The rollout of these new AI accelerators and network systems is expected to begin in the second half of 2026.
According to Reuters, the Broadcom deal could use as much power as 8 million U.S. homes. This has raised new concerns about the growing environmental impact of AI.
A 2024 U.S. Department of Energy report found that data centers could use between 6.7% and 12% of America’s total electricity by 2028, compared to just 4.4% in 2023.
AI’s Growing Energy Appetite
Sam Altman has already acknowledged the massive energy needs behind AI. He once said that a single ChatGPT query uses as much power as keeping a lightbulb on for a few minutes.
Creating realistic AI-generated videos with advanced tools like Sora 2 will likely require even more electricity.
Big Boost for Broadcom
The announcement had an instant impact on the stock market. Broadcom’s shares jumped 12% on Monday morning following the news.
The company’s CEO, Hock Tan, had hinted earlier that Broadcom had secured a mysterious $10 billion customer, now confirmed to be OpenAI.
Custom Chips Give OpenAI More Control
OpenAI has long relied on partners like Nvidia to supply the chips that power its AI data centers. But with this new move, OpenAI plans to develop its own custom AI accelerators.
These specialized chips will give the company more control over the hardware running AI tools like ChatGPT and Sora.
In its press release, OpenAI said that by designing its own chips, it can “embed what it’s learned from developing frontier models and products directly into the hardware, unlocking new levels of capability and intelligence.”
This means the hardware will be optimized specifically for OpenAI’s needs, making it faster, smarter, and more efficient over time.
A Look Toward the Future
The Broadcom partnership marks a major milestone for OpenAI. It’s not just about speed or performance; it’s about control, innovation, and sustainability.
As AI continues to expand, managing power use and building smarter infrastructure will be key challenges for every tech company.
OpenAI’s plan to design its own chips could set a new trend in how AI systems are built, from the inside out.
FAQs
1. Why is OpenAI partnering with Broadcom?
OpenAI is working with Broadcom to create custom AI chips that can handle more complex computing tasks and reduce dependence on third-party suppliers like Nvidia.
2. How much power will the Broadcom partnership use?
The project could use about 10 gigawatts of electricity, roughly equal to the energy used by 8 million U.S. homes.
3. When will the new AI chips be ready?
OpenAI plans to start deploying the new AI systems in the second half of 2026.
4. How does this affect Broadcom’s business?
The deal caused Broadcom’s stock to jump 12% after the announcement. The company also confirmed OpenAI as its new $10 billion customer.
5. What are the environmental concerns?
AI’s growing power use is raising alarms. A U.S. government report says data centers could soon consume up to 12% of all electricity in the U.S. by 2028, highlighting the need for cleaner, more efficient technology.
