Behind OpenAI’s plan to fabricate their own chips, lies a deeper story rife with ambitions and coups, says Satyen K. Bordoloi
Reuters dropped a bombshell in February: OpenAI is diving into chip development. At first glance, this seems like a no-brainer. NVIDIA, the company that designs the chips powering the AI revolution (not to be confused with TSMC, which actually manufactures them), recently became the world’s most valuable company. So, of course, OpenAI wants a piece of that pie. But while the surface-level reasoning is clear, there’s a lot more going on beneath the hood.
OpenAI’s Chip Ambitions: OpenAI is cooking up its first generation of in-house AI chips. Over the next few months, the ChatGPT maker plans to finalise the design of its debut chip and send it off to Taiwan Semiconductor Manufacturing Co (TSMC) for fabrication. This initial step, called “taping out,” is a big deal in chip development. The goal? Mass production by 2026.
But here’s the catch: taping out is expensive—like tens of millions of dollars. It also takes about six months, and there’s no guarantee the chip will work on the first try. If it doesn’t, OpenAI will have to go back to the drawing board, which means more time and money.
This isn’t OpenAI’s first rodeo in the chip game, though. Back in late 2023, they hired ex-Alphabet veteran Richard Ho and assembled a team of around 40 engineers to design a custom chip optimised for AI. They’ve teamed up with Broadcom to leverage TSMC’s cutting-edge 3-nanometer (N3) manufacturing process. N3 is ahead of everything out there as it offers 15% better performance and 30% energy savings compared to older generations. OpenAI wants that to power more efficient and powerful AI models.
Nibbling at NVIDIA’s Dominance: Let’s face it: NVIDIA owns the AI chip market, holding a whopping 80% share. But their dominance comes with a price—literally. NVIDIA’s chip costs have been rising. In addition, the reliance on a single supplier has led big players like Microsoft, Meta, and now OpenAI to explore alternatives.
OpenAI’s move to develop its own chips is just partly about reducing its dependence on NVIDIA. They plan to iterate and create even more advanced processors if their first chip succeeds. This could give them serious leverage in negotiations with other chip suppliers and, who knows, even help them break free from NVIDIA’s grip.
That said, OpenAI’s initial chip won’t replace NVIDIA’s offerings fully. According to Reuters, the chip will primarily be used for running AI models, not training them, and will have a limited role in OpenAI’s infrastructure. But it’s a start.
One Chip to Rule Them All: Think of AI chips as the One Ring from The Lord of the Rings: whoever controls them controls the world. NVIDIA’s meteoric rise to becoming the richest company in the world is proof of that. Sure, ASML in the Netherlands makes the machines that make the chips, and TSMC in Taiwan actually manufactures them. But it’s NVIDIA that’s reaping the rewards.
OpenAI wants in on that action. By developing its own chips, the company isn’t just looking to cut costs or reduce reliance on NVIDIA—it’s aiming to become a major player in the AI hardware space. If they succeed, they could rise to NVIDIA’s level faster than you can say “Jensen Huang.”
The Most Powerful AI Company in the World: Sam Altman, OpenAI’s CEO, is no stranger to ambition—or drama. He’s been through the wringer with Elon Musk, who co-founded OpenAI but left in 2018 after a power struggle. Since then, Musk has repeatedly tried to undermine Altman, but the OpenAI chief has emerged victorious every time.
Altman’s latest power move? Project Stargate is a $500 billion initiative to build the next generation of AI infrastructure. This project not only pushes OpenAI to the forefront of the US’ AI ambitions but also serves as a checkmate against Musk’s own AI ambitions and political influence.
Altman sees himself as the future czar of AI. He isn’t wrong to think big. With OpenAI’s valuation skyrocketing post-ChatGPT, he’s got the resources and the clout to make it happen. But to truly dominate, he can’t rely on another company for something as critical as chips. By mastering chip manufacturing, Altman could position OpenAI as the most valuable company in the world.
Apple vs. OpenAI: Two Paths, One Goal: It’s interesting to see how OpenAI and Apple are taking opposite routes to the same destination. Apple, traditionally a hardware company, is now diving headfirst into AI, interestingly, with OpenAI’s help. They’re calling this Apple Intelligence. OpenAI, on the other hand, is doing the reverse – branching out from software into hardware.
For Altman, this move is well thought out. OpenAI’s success hinges on having the best AI models, and those models need the best chips. By controlling both the software and the hardware, OpenAI could optimise its systems in ways that no other company can.
Turbulence Ahead for OpenAI’s Flight: Of course, it’s not all smooth sailing. Big tech companies like Microsoft and Meta have been trying to develop their own chips for years, with mixed results. The recent market shakeup caused by Chinese AI startup DeepSeek has also raised questions about whether we’ll even need as many chips in the future.
But OpenAI isn’t one to back down from a challenge. With Altman at the helm and a war chest of funding, they’re betting big on their chip ambitions. Whether they succeed or fail, one thing is for sure: the AI landscape is about to get a lot more interesting.
The Bigger Picture: OpenAI’s foray into chip production isn’t just about saving money or reducing reliance on NVIDIA. It’s about control, power, and the future of AI. By developing its chips, OpenAI is positioning itself as a vertically integrated powerhouse capable of dominating both the software and hardware sides of the AI revolution.
For Sam Altman, this is the ultimate power play. If he can pull it off, OpenAI could become the most valuable company in the world—and Altman could cement his legacy as the undisputed king of AI.
So, buckle up. The AI race just got a whole lot more intense, and OpenAI is gunning for the lead.
In case you missed:
- Unbelievable: How China’s Outsmarting US Chip Ban to Dominate AI
- Apple Intelligence – Steve Jobs’ Company Finally Bites the AI Apple
- You’ll Never Guess What’s Inside NVIDIA’s Latest AI Breakthrough
- DeepSeek not the only Chinese model to upset AI-pple cart; here’s dozen more
- Project Stargate: Dubious Origins in the 1970s to AI Goldrush in 2025
- OpenAI’s Secret Project Strawberry Points to Last AI Hurdle: Reasoning
- Kodak Moment: How Apple, Amazon, Meta, Microsoft Missed the AI Boat, Playing Catch-Up
- AI vs. Metaverse & Crypto: Has AI hype lived up to expectations
- How Lionsgate-Runway Deal Will Transform Both Cinema & AI
- Rufus & Metis Tell Tales of Amazon’s Delayed AI Entry
