Jensen Just Pulled a Satya

Nvidia reportedly just agreed to buy Hugging Face for $12.9 billion.

And the more I sit with it, the more familiar it feels.

Like GitHub familiar.

I live in Hugging Face now. But I've certainly spent my share of time on GitHub too. Not casually. I use this stuff daily. That's why this deal immediately made me think of Microsoft buying GitHub.

GitHub Was Never Just GitHub

When Microsoft paid $7.5 billion for GitHub in 2018, a lot of people looked at the deal and saw code repositories.

That was never really the asset. GitHub was where developers lived.

It was where projects started. Where collaboration happened.

Where software got forked, broken, fixed, improved, copied, shipped and discovered. And if you were actually using it every day, that distinction was obvious. You didn't "visit GitHub."

You worked there.

That was the genius of Satya Nadella's move.

Microsoft already had Windows, Azure, Visual Studio, .NET and one of the biggest enterprise footprints on Earth.

What it didn't own was the neutral ground where developers were already spending their time.

So instead of trying to drag everyone into Microsoft's ecosystem, Microsoft bought the ecosystem they were already standing in.

That's what this Nvidia move feels like.

Hugging Face Is That Place Now

Hugging Face is often described as the GitHub of AI.

That comparison gets thrown around so much that it almost sounds lazy.

But if you actually use both, it's pretty damn accurate.

Hugging Face is where you go to find out what's happening.

A new model drops.

Someone quantizes it.

Someone makes a smaller version.

Someone fine-tunes it.

Someone benchmarks it.

Someone wraps it in a new inference stack.

Someone figures out how to run it on hardware it was never designed for.

Then somebody else builds on that.

It's messy, fast, global and constantly moving.

That's what makes it valuable.

Models live there.

Datasets live there.

Libraries live there.

Benchmarks live there.

Researchers live there.

Startups live there.

Open-source communities live there.

A huge amount of the open AI ecosystem effectively passes through the place.

And Nvidia apparently wants to own it.

This Isn't a Revenue Deal

Reported annual revenue is around $150 million.

The reported purchase price is $12.9 billion.

That's roughly 86 times revenue.

Nobody at Nvidia accidentally misplaced the decimal point.

They're not buying Hugging Face for what it makes today.

They're buying it for where it sits in the graph.

That phrase keeps coming back to me because it's the cleanest way I know to look at these things.

Where does the asset sit?

What flows through it?

What becomes easier if you control it?

What can you influence without having to explicitly own?

GitHub sat between developers and software.

Hugging Face sits between developers, models, datasets, inference and compute.

And that last word is the whole thing.

Compute.

Everything happening on Hugging Face eventually touches it.

You train a model.

You fine-tune one.

You benchmark one.

You run inference.

You build an agent.

You deploy something.

You scale something.

Eventually the software asks for silicon.

And right now, an enormous amount of that silicon is Nvidia.

This Is Why the Deal Makes Sense

Nvidia doesn't need to own every AI model.

That would actually be a terrible strategy.

Let OpenAI build theirs.

Let Anthropic build theirs.

Let Meta build theirs.

Let China build theirs.

Let a 22-year-old in a dorm room release something terrifyingly good at 3 a.m.

Nvidia benefits from all of it.

More models means more experimentation.

More experimentation means more inference.

More inference means more compute.

More compute means more demand for GPUs, networking, systems and infrastructure.

So what do you buy?

The place where all that activity is already happening.

That's why I keep coming back to GitHub.

Microsoft didn't need to own every piece of software.

It wanted to own the place where software got built.

Nvidia doesn't need to own every AI model.

It wants to own the place where AI gets built, shared, discovered and increasingly deployed.

That's a much bigger idea.

The Smart Move Is to Barely Touch It

There's also a lesson here from Microsoft.

The smartest thing Microsoft did after buying GitHub wasn't wrecking it.

That sounds obvious now.

It didn't feel obvious at the time.

A lot of developers absolutely hated Microsoft back then.

There was real fear that Microsoft would turn GitHub into a funnel for Azure, Windows and its own developer stack.

Instead, it mostly preserved what made GitHub valuable.

The front door stayed familiar.

The platform stayed useful across ecosystems.

Then Microsoft gradually wrapped more value around it.

Actions.

Codespaces.

Copilot.

Security.

Identity.

Enterprise tooling.

Azure integration.

That's exactly what Nvidia should do here.

Don't turn Hugging Face into Nvidia Hugging Face.

Don't make it feel captured.

Don't make AMD developers feel like second-class citizens.

Don't make researchers wonder whether the platform is being tilted under their feet.

Keep the front door neutral.

Then make Nvidia ridiculously easy behind it.

Better CUDA support.

Better optimization.

Better inference.

Better deployment.

Better enterprise tooling.

Better DGX workflows.

Better access to Nvidia infrastructure.

Let everyone keep showing up.

Just make the easiest path increasingly run through Nvidia.

That's the Satya play.

And I Recognize the Move Because I Use the Stuff

This is the part that's hard to see if you only look at the companies from the outside.

You can read financial statements.

You can look at multiples.

You can compare valuations.

You can talk about TAM and margin expansion and strategic synergies.

That's all useful.

But sometimes you understand an acquisition better by watching your own behavior.

Where do I go when I want to find a model?

Hugging Face.

Where do I go when I want to see what people are actually running?

Hugging Face.

Where do I go when somebody tells me some obscure model is suddenly punching above its weight?

Hugging Face.

Where do I end up when I'm figuring out whether I can run something locally on one of my machines?

Usually Hugging Face.

That's not brand awareness.

That's habit.

And habit at developer scale becomes infrastructure.

GitHub figured that out years ago.

Now Hugging Face has it.

Nvidia apparently sees it too.

From Possibility Machines to Possibility Infrastructure

I wrote previously about Nvidia building what I called possibility machines.

That has always been their genius.

Build incredibly powerful programmable hardware and let the world figure out what it can do.

Gaming.

Simulation.

Scientific computing.

Crypto.

Machine learning.

AI.

They didn't need to predict every use case.

They created the conditions for use cases to emerge.

But the company isn't just building possibility machines anymore.

It's building possibility infrastructure.

CUDA.

Networking.

DGX.

AI factories.

Inference.

Robotics.

Simulation.

Software.

And now potentially one of the places where the open AI ecosystem actually organizes itself.

That's a very different kind of company than the graphics-chip business Nvidia once was.

Jensen Just Pulled a Satya

That's still the cleanest way I know to say it.

Satya Nadella realized Microsoft didn't need to own all software.

It needed to become indispensable to the people building software.

GitHub gave Microsoft that position.

Jensen Huang may be making the same move for AI.

Not: we need to own every model.

But: we need to be wherever the builders are.

And if you've spent enough time living in these systems, you know why that matters.

Developers don't move because a corporate strategy deck tells them to.

They move because one place becomes easier.

More useful.

More connected.

More familiar.

More alive.

Eventually, that place stops feeling like a product.

It becomes where the work happens.

Microsoft bought that place for software.

Nvidia may have just bought it for AI.

Jensen just pulled a Satya.

---

Rich Washburn is a technologist, strategist, and Founder & Chief AI Architect of ARIA AI Labs, working at the intersection of AI, infrastructure, communications, and capital. He also serves as Managing Partner and Chief AI Officer at Eliakim Capital.