In a recent interview, Jonathan Ross, founder of Groq, echoed Sam Altman’s thoughts by saying, "Compute is the new oil." As the CEO of OpenAI, Altman is determined to solve the challenges surrounding energy and compute, essential components for scaling generative AI services to a broader customer base. Altman believes that compute will become “the currency of the future,” possibly the “most precious commodity in the world.” This is why he advocates for heavy investments to massively increase compute capacity.
However, he also acknowledges that the real challenge lies in addressing the massive infrastructure and energy demands required for this future.
To meet the growing demands of AI, OpenAI and Microsoft have joined forces on a massive $100 billion data center project named ‘Stargate,’ which is expected to launch by 2028. This project is poised to consume vast amounts of energy, potentially reaching 5 gigawatts—equivalent to running several large data centers.
Altman’s approach to solving the energy puzzle goes beyond just infrastructure. He has made significant investments in energy, notably with Helion Energy, a nuclear fusion startup. Helion’s Project Polaris aims to produce electricity from nuclear fusion by 2024. Microsoft has already committed to purchasing energy from Helion's first fusion power plant, with plans to generate 50 megawatts for its own operations. Given the scale of the Stargate data center, it's likely that Helion’s fusion power will play a key role in providing the energy needed.
Altman is optimistic about the potential of nuclear fusion, noting, “Helion’s doing the best work, but I’m happy there’s a race for fusion right now. Nuclear fusion is also quite amazing.” In addition, Altman is backing another nuclear startup, Oklo, which is developing smaller-scale fission reactors.
Altman’s investment interests are not limited to energy but also extend to hardware. His portfolio includes a potential collaboration with Emerson Collective and Thrive Capital on an AI device project led by former Apple designer Jony Ive. Ive is aiming to raise up to $1 billion for the venture, which builds on Altman’s previous investment in Humane, the company behind the AI Pin wearable device released in November 2023.
With OpenAI’s continued development of advanced AI models, such as GPT-5, the need for dedicated hardware and data centers becomes ever more critical. For instance, OpenAI’s latest text-to-video generation model, Sora, requires massive compute resources, which could explain its limited availability due to the high costs and energy demands.
Looking ahead, Altman envisions a world where access to compute is a key factor in determining how AI will shape our future. As he put it, “If it’s really cheap, I’ll have it like reading my email all day, giving me suggestions about what to work on, and trying to cure cancer. If it’s really expensive, maybe I’ll only use it to try to cure cancer.” Altman believes that AI will eventually be more like energy in how it’s utilized, with the value derived from the amount of compute one can access and the price at which it is available.
Altman’s vision aligns with the broader industry trend. Amazon is investing $150 billion over the next 15 years to expand its data center infrastructure to support a growing demand for AI and other digital services. Meanwhile, Elon Musk has pointed out that AI compute is expanding exponentially, increasing by a factor of 10 every six months. Musk also joked that we’ll need “transformers for the transformers” to support the power demands of AI's neural networks. This comment underscores the scale of infrastructure required for AI’s future.
Altman is also eyeing AI hardware as a key piece of OpenAI’s future. Despite relying on NVIDIA GPUs for some operations, OpenAI has made it clear that it plans to explore alternatives, such as AMD and Microsoft’s Maia 100 AI chips, for use in its Stargate project. In a move to further reduce its dependency on NVIDIA, OpenAI has even told Microsoft that it doesn’t want to use NVIDIA’s proprietary InfiniBand cables for Stargate’s supercomputer.
Altman’s interest in launching his own AI chip venture has been widely reported, and he’s currently seeking approval from the U.S. government for this new initiative. Additionally, discussions with investors from the Middle East and chip fabricators like TSMC indicate that OpenAI is gearing up to take on a larger role in developing its own hardware to meet the growing demands of AI.
