Big tech is capable of throwing around some eye-watering amounts of cash. As you may recall, Nvidia announced $46.7 billion total revenue during its Q2 2025 earnings call. That’s not just a lot of moolah, but serious spending power.
As such, this week, Nvidia announced it will be investing $100 billion into OpenAI. Part of this mountain of money will go towards supplying the steward of ChatGPT with data centre chips. Details have yet to be finalised, but a letter of intent signed by the two companies announced plans to deploy 10 gigawatts of Nvidia systems for use in OpenAI’s data centres.
The two companies were hardly strangers to begin with, but this latest deal gives Nvidia a stake in one of its biggest customers. Nvidia’s investment in OpenAI will eventually take the form of non-voting shares in the company. OpenAI will then use the resulting cash flow to buy the aforementioned AI chips. Ultimately, this latest pledge of $100 billion makes last week’s surprising news that Nvidia would be putting $5 billion into Intel look like a drop in the bucket.
Once this most recent deal is finalised, sources close to the company claim the plan is for Nvidia to invest an initial sum of $10 billion, followed by a hardware rollout sometime towards the end of 2026. The first gigawatt of power will likely take to the stage of Nvidia’s upcoming Vera Rubin AI compute platform, which was first revealed back in March.
OpenAI CEO Sam Altman explained in a statement that it was all about maintaining a competitive edge in an increasingly crowded field, saying, “Everything starts with compute. Compute infrastructure will be the basis for the economy of the future, and we will utilize what we’re building with Nvidia to both create new AI breakthroughs and empower people and businesses with them at scale.”
But Nvidia spending money on OpenAI so OpenAI can then buy Nvidia hardware has raised some concerns; if this flow of cash looks a little circular to you, you’re not the only one concerned about the potential shape of things to come.
Speaking to Reuters, Bernstein analyst Stacy Rasgon commented, “On the one hand this [deal] helps OpenAI deliver on what are some very aspirational goals for compute infrastructure, and helps Nvidia ensure that that stuff gets built. On the other hand the ‘circular’ concerns have been raised in the past, and this will fuel them further.”
Though that said, it’s perhaps too early to start throwing around words like ‘antitrust,’ particularly as the US Trump administration is all in on AI. Still though, the proposed 10 gigawatt data centres will demand power equivalent to the needs of 8 million U.S. households; despite Nvidia CEO Jensen Huang’s suggestion that AI customers ‘pace themselves’ and other major big tech players looking to nuclear to meet AI’s power demands, there may come a time when such a power imbalance can no longer be ignored.
Best graphics card 2025
