
Why startups might be the only winners in the AI talent war
With top AI graduates commanding £220k salaries at companies like Anthropic and traditional industries unable to compete, Mark Miller – serial entrepreneur and CEO of insurtech InsureVision – talks about the growing talent crisis. His view? The talent war isn't just reshaping big tech – it's creating the biggest opportunity startups have seen in decades.
Anthropic is paying more than £200,000 for new grads. How did we get here?
It’s actually rational if you look at the numbers. Meta committed £40 billion annually to AI infrastructure. When you’re burning that much compute, paying £200,000 to get the best minds is a rounding error. However, this does then create a massive opportunity gap. Entire industries like insurance, healthcare, and logistics can’t compete on salary. They need innovation but can’t access the talent. That’s where startups come in.
We’re looking at a 10x salary differential in some cases. Can traditional industries survive this?
Look at the compute distribution data – Europe went from 10% of global compute in 2010 to just 2% today. The US big tech companies probably control two-thirds. It’s not just salaries; it’s infrastructure. Traditional industries are being stripped of their ability to innovate internally. They’ll have to rely entirely on partnerships and acquisitions
Why would top AI talent ever join a startup over big tech?
Two words: ownership and impact. Yes, Anthropic pays a lot of money, but only the early Anthropic employees made generational wealth – we're talking about 100x returns for those who joined in 2021. More importantly, there’s a certain type of brilliant mind that’s bored by optimising ad clicks. They want to solve real-world problems. In our case, we’re literally trying to save lives with AI-powered risk assessment. You can’t put a price on that mission.
But don’t startups lack the compute resources to compete?
This is the counterintuitive part – you don’t need Meta-scale infrastructure to solve specific problems. We process millions of hours of video with a fraction of their compute because we’re laser-focused on one challenge. Big tech trains massive general models; startups can build specialised ones that outperform in narrow domains. It’s David vs Goliath, but David knows exactly where to aim.
What about the “AI bust” we see mentioned? Should founders be worried?
There will absolutely be a correction. We’re seeing echoes of the dot-com era. But remember – Amazon and Google emerged from that bust stronger than ever. The key is solving real problems for real customers, not just riding the hype. When the music stops, companies with actual revenue and defensible technology will thrive.
The UK government offers compute grants but with massive bureaucracy. The US just hands out thousands with minimal red tape. Can UK/EU startups compete?
The application process for UK compute resources reads like a Kafka novel. Meanwhile, Google Cloud is happy to grant you a few hundred thousands with just a few email exchanges. But, constraints breed creativity. UK and EU startups have to be more capital efficient, more focused. Sometimes that’s exactly what you need to build something transformative rather than just throwing compute at problems.
If you’re a founder in a traditional industry, how do you attract AI talent?
First, stop competing on salary – you’ll lose. Compete on mission, and the promise of solving unsolved problems. Second, be specific about the challenge. “We’re using AI to transform insurance” is boring. “We’re building computer vision that can predict accidents before they happen” is exciting. Third, give them the tools. You might not have Meta’s infrastructure, but you can give them freedom to publish, attend conferences, and work on genuinely hard problems.
Final question – is this current lack of talent distribution sustainable? What happens next?
The current situation is absolutely unsustainable. You can’t have one industry hoarding all the talent while others wither. We’ll see one of two things: either a massive correction that redistributes talent, or a new model emerges where specialised AI companies serve traditional industries. My bet is on the latter. The future isn’t every company becoming an AI company – its AI-native startups solving domain-specific problems for industries that desperately need innovation.
For more startup news, check out the other articles on the website, and subscribe to the magazine for free. Listen to The Cereal Entrepreneur podcast for more interviews with entrepreneurs and big-hitters in the startup ecosystem.