Internationally recognised art and literary critic, writer and cultural producer. Published in ELLE, Glamour, Esquire and Cosmopolitan; member of the Union of Writers.
I was sitting in a coffee shop in Shoreditch the other day, listening to two screenwriters panic. They weren’t panicking about plot holes or difficult directors. They were panicking about “compute.”
It’s a word that didn’t used to belong in the arts. We worried about funding, or lighting, or whether the audience would actually turn up. But now, we worry about compute. And if you want to know why, you have to look at the man in the leather jacket.
Jensen Huang, the CEO of Nvidia, stood up this week and effectively told the world to stop worrying and love the machine. There’s been a lot of noise recently – nervous, jittery noise from Wall Street – that this whole AI thing is a bubble. That we’re staring at the Dot Com crash part two, just with better chatbots.
But Huang isn’t having it. He came out with earnings that frankly defy gravity. Sales up 60%. A projection of $65 billion in revenue for the next quarter alone. He looked the analysts in the eye and said, “From our vantage point, we see something very different.”
Now, usually, when a tech CEO tells you everything is fine, you should check your wallet. But here’s the thing. I think he might be right. And that terrifies me more than if he were wrong.
Because if this isn’t a bubble – if this is, as the money men are calling it, a “build-out” – then the cultural sector is in for a shock that makes the streaming revolution look like a mild hiccup.
The Concrete of the Cloud
Let’s strip away the jargon. Nvidia makes the chips. The GPUs. These aren’t just computer parts anymore; they are the bricks and mortar of the 21st century.
Colette Kress, Nvidia’s CFO, dropped a number that should make every artist, curator, and theatre director freeze. She anticipates $3 trillion to $4 trillion in AI infrastructure spending by the end of the decade.
Three. Trillion.
That is the GDP of the UK. That is the money being poured into building the physical reality of the next digital age.
Why does this matter to us? To the arts?
Because we live in their house now.
For the last decade, culture has been slowly migrating to the cloud. Our archives are digital. Our distribution is streaming. Our marketing is algorithmic. We thought we were just upgrading our tools. But actually, we were moving into a rented apartment where the landlord is currently tearing down the walls to build a server farm.
If $4 trillion is going into “infrastructure,” that means the pipes, the roads, and the power grid of the internet are being redesigned solely to facilitate AI. It’s not being built for nuance. It’s not being built for the weird, inefficient, human mess that is great art. It’s being built for speed, prediction, and scale.
The Efficiency Trap
There was a line in the earnings call that really stuck with me. Kress was bragging about Salesforce – how their engineering team is now “30% more efficient” because they use AI for coding.
It sounds great, doesn’t it? Efficiency.
But art has never been about efficiency.
If you make a writer 30% more efficient, you don’t get 30% better novels. You get 30% more novels. You get glut. We are already drowning in content – a slurry of mediocre TV shows, playlist-filler pop music, and films that feel like they were written by a committee of robots (because they probably were).
Nvidia is selling the shovel that digs this hole.
The danger here isn’t just that AI will “replace” artists. That’s the sci-fi fear. The real, boring, gritty danger is that the economic model of culture shifts entirely to volume.
If the infrastructure is designed to process billions of tokens per second, the system demands to be fed. It needs “data.” It needs “content.” It doesn’t care if that content is a lovingly crafted play at the Royal Court or a million generated images of cats in spacesuits. It just needs flow.
And when the tech giants – Meta, Google, Microsoft – are spending $400 billion a year just to stay in the race, they are going to need to monetize that flow. Aggressively.
This squeezes the margins for anything slow. Anything difficult. Anything that requires a human being to sit in a room and think for six months without producing a “deliverable.”
The Bubble That Won’t Burst
I’ll be honest, part of me was hoping it was a bubble.
I remember the crypto craze. I remember NFT art. (God, remember that? The bored apes?) It felt feverish and stupid, and deep down, we all knew it would pop. And when it did, there was a grim satisfaction in watching the grifters pack up their tents.
But Nvidia feels different.
Huang made a point that I can’t shake. He said that even if the “Generative AI” stuff – the ChatGPTs and the Midjourneys – doesn’t make as much money as people hope, Nvidia is still safe. Why? Because the entire world’s computing system is upgrading.
“The world has a massive investment in non-AI software,” he said. He’s talking about data processing, science simulations, engineering. The boring stuff that keeps the lights on.
This means the hardware isn’t going away. The data centres aren’t going to be abandoned like WeWork offices. They are the new railways.
So, if the rails are laid, the trains will run.
For the arts, this means we can’t just wait for the hype to die down. We have to accept that the environment has fundamentally changed. The tools we use to create, distribute, and even talk about culture are now running on Nvidia’s engine.
The Circular Economy of Hype
There is, however, a scent of something rotten in the state of Denmark. Or rather, Silicon Valley.
Some of the analysts – the brave ones – are pointing out that this money is going in a circle. Nvidia invests in OpenAI. OpenAI buys chips from Nvidia. It looks great on a balance sheet, but is actual real-world value being created?
Daniel Morgan, a portfolio manager, called it “circular funding deals.” It’s a polite way of saying they’re passing the same ten-pound note back and forth and calling it growth.
If this circle breaks, the fallout for the arts will be brutal.
Think about where cultural funding comes from these days. It’s not the Arts Council (well, not entirely). It’s corporate sponsorship. It’s tech philanthropy. It’s the trickle-down from the booming economy.
If the tech sector sneezes, the arts catch pneumonia. We saw it in 2008. We saw it during the pandemic. If these companies realize they’ve spent $4 trillion on a party that no one attended, the first thing they will cut is the “soft” stuff. The gallery sponsorships. The outreach programs. The daring commissions.
But there’s a darker possibility. What if the circle doesn’t break? What if they force it to work?
To make that $4 trillion investment pay off, they need AI to be in everything. They need it in your word processor, your editing suite, your camera, your streaming service recommendation algorithm.
They need to normalize the synthetic.
The Homogenization of Taste
This is where I get really worried. Not about the money, but about the soul of the thing.
Kress mentioned that at Meta, AI recommendation systems are leading to “more time spent on apps.”
Great for Mark Zuckerberg. Terrible for culture.
“More time spent” usually means “more passive consumption.” It means the algorithm feeding you things it knows you will like. It removes the friction of discovery.
I remember wandering into a record shop in Camden in the 90s and buying an album just because the cover looked weird. It turned out to be terrible, but it was my mistake. It was a human error.
The Nvidia-powered future wants to eliminate error. It wants to predict your taste with such accuracy that you are never challenged, never bored, and never surprised.
If you are an artist trying to make something abrasive, or quiet, or strange, you are fighting against a $4 trillion machine designed to smooth you out.
The Return of the Physical?
Is there a silver lining? Maybe.
I’ve been noticing something interesting in the galleries recently. A return to the visceral. Paint that looks like paint. Performance art that involves sweat and spit. Books that are beautifully, impractically bound.
It makes sense. If the digital world is being colonized by the machines – if the screen becomes a place of slick, AI-generated perfection – then the physical world becomes the premium product.
Nvidia can simulate light bouncing off a surface perfectly. But it can’t simulate the smell of oil paint or the feeling of a bass drum hitting your chest in a basement club.
Perhaps this “build-out” will force a bifurcation in culture.
On one side, you’ll have the “Content Slop” – infinite, personalized, AI-generated entertainment, running on Nvidia’s chips, delivered through your headset. It will be cheap (or free), addictive, and utterly forgettable.
On the other side, you’ll have “Human Art.” It will be expensive. It will be scarce. It will be flawed. And it will be the only thing that feels real.
The “Human Premium”
I was talking to a musician friend about this. He’s terrified of Suno and Udio – those AI music generators that can spit out a radio-ready pop song in thirty seconds.
“How do I compete with that?” he asked.
“You don’t,” I said. “You sell the fact that you have a pulse.”
We might be entering an era where the “Human Premium” becomes the defining feature of the art market. Just as we pay more for organic vegetables or handmade furniture, we will pay more for art that contains the evidence of human struggle.
But this creates a class issue. If “Real Art” becomes a luxury good for the wealthy, while the masses are fed AI slop, we are heading for a cultural dystopia that Orwell didn’t even predict.
The View from the Cheap Seats
So, when Jensen Huang talks about the “next industrial revolution,” he’s not just talking about factories. He’s talking about the imagination.
He’s telling us that the future of creativity is going to be powered by silicon, not caffeine and anxiety.
It’s easy to get swept up in the numbers. $65 billion. 60% growth. It sounds like success. But we have to ask: Success for whom?
For the shareholders? Absolutely. For the tech giants? Probably.
But for the playwright trying to capture the nuance of a broken heart? For the painter trying to mix a colour that doesn’t exist in RGB?
For them, the Nvidia earnings report isn’t a celebration. It’s a noise complaint.
It’s the sound of a massive construction project happening right next door to the library. And the builders don’t care if you can’t concentrate. They have a deadline. They have a budget to burn. And they are pouring the concrete whether we like it or not.
I keep thinking about that phrase: “Not to burst your bubble.”
Maybe we need the bubble to burst. Bubbles are messy when they pop, but at least they clear the air.
The alternative – the “solidification” of this AI infrastructure – means we are locking ourselves into a technological trajectory that treats culture as data.
Nvidia has built the engine. It’s powerful. It’s impressive. It’s undeniably brilliant. But an engine doesn’t know where it’s going. It just knows how to go fast.
It’s up to us – the writers, the makers, the critics, the audience – to grab the steering wheel. Or, at the very least, to refuse to get in the car.
Because if we don’t, we’re just passengers in Jensen Huang’s dream. And frankly, I prefer my own dreams. They’re cheaper, they’re messier, and they don’t require a GPU to run.
