
Your Phone Charger Pulls Like 20 Watts. This Deal Is Measured in Billions.
Quick gut check: your laptop probably pulls somewhere around 60 watts when you're actually pushing it.
Nvidia just backed a data center that needs 8 gigawatts.
That's enough electricity to run roughly six million American households at the same time. Not phones. Not laptops. Entire houses.
Okay, What Actually Happened
Nvidia just agreed to guarantee up to $105 billion in financing for a new data center in Pike County, Ohio.
The site is being built specifically to run OpenAI's next-generation models, and Nvidia will be the exclusive chip provider for the whole thing.
OpenAI's signed a 20-year lease. The first chunk of power comes online in 2028. The full 8-gigawatt buildout is the long-term plan.
Why This Deal Is Actually a Big Deal
The raw scale is genuinely absurd â this single site will out-power the largest data center currently running in the US, and it's set to surpass the nation's biggest nuclear plant in output.
Nvidia is locking in OpenAI's entire infrastructure stack â this isn't just selling chips anymore, it's Nvidia financing the buildings, the power, and the lease so OpenAI has nowhere else to go for compute.
This is what "next-gen models" actually costs â every leap in AI capability from here on out is going to demand this level of energy, not less.
Is This Visionary or a Problem?
Here's the actual tension: this is either the infrastructure that gets us genuinely sci-fi-level AI, or it's two companies quietly becoming the electricity gatekeepers of the entire internet.
Nvidia's CEO is already pushing back on people calling this "circular financing" â where the chip maker, the customer, and the investor all end up being the same handful of companies.
Worth noting: local communities near these builds are already dealing with strained power grids and rising electricity costs before a single GPU turns on.
So Where Do You Land
This is either the coolest infrastructure story of the decade or a genuinely unsustainable energy grab, depending on who you ask.
Does the scale of this actually worry you, or are you just here for the faster models it's going to produce?
About
Featured Posts






