OpenAI is making another major move to secure its leadership in the generative AI revolution. The company behind ChatGPT has announced a strategic partnership with Broadcom, one of the world’s largest semiconductor companies, to design and produce its own specialized AI processors.
The collaboration, expected to kick off next year, will deliver a staggering 10 gigawatts of computing power—enough energy to power a large city. This step underscores how OpenAI is doubling down on building the massive infrastructure needed to support the next phase of AI growth.
Powering the Next Phase of Generative AI
Since ChatGPT’s launch in November 2022, OpenAI has grown at a record pace, with more than 800 million users engaging with its services weekly. To keep up with that demand, the company is shifting from relying on standard chips to creating its own custom-designed processors.
By working with Broadcom, OpenAI aims to build faster, more efficient, and purpose-built AI hardware that can run advanced models more effectively than off-the-shelf solutions.
A Strategic Partnership With Big Implications
While the financial details of the Broadcom agreement weren’t disclosed, the announcement sent Broadcom’s stock soaring nearly 10%—mirroring similar surges seen with AMD and Oracle after their deals with OpenAI.
Broadcom will manufacture and co-develop these chips, while OpenAI integrates them into its own data centers and those of its partners. According to CEO Sam Altman, this is a “critical step in building the infrastructure needed to unlock AI’s full potential for people and businesses.”
Infrastructure Growth Meets Energy Concerns
The rapid expansion of AI infrastructure brings both opportunities and challenges. Delivering 10 gigawatts of computing power will demand vast amounts of energy, raising concerns about electricity supply and sustainability.
This reflects a broader issue across the AI sector: the race to scale up data centers and chip capacity may strain power grids and resource availability. As the AI arms race intensifies, energy efficiency is becoming a critical factor in long-term strategy.
A Pattern of High-Profile Deals
The Broadcom announcement is part of a larger investment spree led by Sam Altman. In recent weeks, OpenAI has signed major agreements with Nvidia, AMD, Oracle, and South Korean tech giants Samsung and SK Hynix.
Each deal is designed to expand OpenAI’s computing capacity, giving it more control over the technology stack that powers its AI models. This approach also helps reduce dependency on external suppliers—a strategic move as competition in the AI chip market heats up.
A Booming Market, But Questions Remain
Despite the rapid expansion and investor enthusiasm, AI profitability remains uncertain. Many analysts point out that the AI industry, while growing fast, has yet to prove it can achieve sustainable revenue at scale.
Some observers are also drawing comparisons to the dot-com bubble of the late 1990s, warning that the surge in investment could create vulnerabilities if expectations outpace actual returns.
Building the Future of AI Hardware
Still, OpenAI remains focused on innovation. Developing custom chips allows the company to optimize hardware for its specific AI models, potentially making them more powerful and cost-efficient.
Broadcom called the partnership a “pivotal moment” in AI development, signaling a new era where the most powerful AI companies also control their underlying infrastructure.
