Imagine being offered $100 million just to change jobs. OpenAI CEO Sam Altman says that's the signing bonus Meta dangled in front of his staff. That's the AI talent war in a nutshell, with tech giants such as Meta, Google and Microsoft racing to hire a small pool of elite AI researchers and engineers, with pay packages now reaching tens, even hundreds, of millions of dollars.
So why are tech giants paying millions? Building a frontier AI model can cost hundreds of millions, and only a few people can do it well. Demand far outstrips supply, one expert can shape a company's direction, and every hire is one a rival can't make.
If you're still running key processes by hand and unsure where AI fits, salaries like that can feel a world away. They're closer to your business than you think, and you don't need Big Tech's budget to compete.
Here's what's really driving the AI talent skills war, and what it means for your business.
What the AI talent war actually looks like
This isn't one eye-catching offer, it's a pattern that is happening across almost every major AI player, and it plays out less like a hiring market and more like a football transfer window.
- Meta set up its Superintelligence Labs and brought in Scale AI co-founder Alexandr Wang as part of a $14 billion investment in the start-up. It has also reportedly offered up to $250 million to a single 24-year-old researcher.
- Google brought Windsurf co-founder and CEO Varun Mohan into Google DeepMind in a $2.4 billion deal.
- Microsoft quietly hired around two dozen Google DeepMind employees for its own AI division.
Each move is about more than one person. It's about buying a head start.
And it isn't only the elite few who benefit. In the UK, PwC’s analysis found the wage premium for AI skills rose to 34.2% in 2025, up from 11% the year before.
Why tech giants are paying millions for AI talent
Frontier models cost far more than the people who build them:
Training a leading AI model is eye-wateringly expensive. Stanford's 2025 AI Index estimated OpenAI's GPT-4 cost $79 million to build, Google's Gemini 1.0 Ultra $192 million and Meta's Llama 3.1-405B $170 million (Stanford HAI AI Index 2025).
Against that, even an eight-figure salary looks small. As Synthesia's Alexandru Voica told CNBC, when a model costs a billion, $10 million for an engineer is a relatively low investment.
Demand has exploded; supply hasn't:
The number of researchers who can build these systems has stayed fairly flat while demand has rocketed, and that gap is what drives wage inflation. Many of the most sought-after people come from a small group of top universities or have years of hands-on experience with large language models. There simply aren't many of them.
One hire can change a company's direction:
At this level, one person can shape which products get built and which bets pay off. Losing a key researcher can stall a programme for years, and every expert you hire is one your rival can't.
Speed is the real prize:
In AI, months matter. The company that ships first sets the standard, then attracts the users, data and talent that keep it ahead. Hiring people who've already done it cuts the learning curve, which is why the giants poach whole teams rather than training from scratch.
Money alone doesn't win:
Some top researchers have reportedly turned down enormous offers in favour of better computing power, research freedom and a clearer vision.
Pay opens the door. The work, the tools and the mission close the deal. That's encouraging, because it means you can compete on interesting work, not just on pay.
What the AI talent war means for your business
You almost certainly don't need a $100 million researcher. Those salaries sit with the few companies training their own models. Most businesses build on top of existing models, which needs a different and far more affordable skill set.
That changes how you can approach the doubts we hear most often from business leaders.
Unclear return on investment:
McKinsey found 88% of organisations now use AI in at least one function, yet only 39% report any impact on enterprise-level EBIT. The roughly 6% of high performers are nearly three times as likely to have redesigned workflows around AI. So, ROI comes from people who rethink how work gets done, not from tools bolted onto old processes.
High implementation costs:
You're paying for engineers who apply and integrate AI, not researchers inventing it. Starting with a contract specialist or a project-based delivery team lets you prove value before you commit to permanent headcount.
Lack of AI knowledge, hype and confusion:
A good AI engineer cuts through the noise and turns we should do something with AI into two or three realistic, measurable use cases. Not sure which skills you need? Start with what an AI engineer actually does.
Data quality issues:
AI is only as good as the data behind it. Often your first hire should be a data specialist who gets your foundations right before any model goes live.
Security and privacy:
Look for people who've deployed AI under UK GDPR and clear governance rules. Ask candidates how they've handled sensitive data, access controls and model monitoring on past projects. The right hire reduces your risk rather than adding to it.
Fear of operational disruption:
Run pilots alongside existing processes, then scale what works. Experienced engineers plan phased rollouts, not big-bang change, so your teams keep working while the new system proves itself.
Skills gaps:
You don't have to hire everything in. Pair a few experienced AI hires with upskilling your current team or use a recruit, train and deploy model to build capability shaped around your needs.
How to compete for AI talent without a Big Tech budget
You can't outbid Meta. And you don't need to. Here's what works for the businesses we partner with.
- Define the problem before the role. Write down the business problem, the data you have and what success looks like in six months. A sharp brief attracts sharper candidates and stops you hiring a researcher when you really need a builder.
- Sell the impact, not just the salary. In a smaller business, an engineer can own a project end to end and see the results. That matters to good people, and career growth often beats salary.
- Benchmark pay honestly. Know where the market sits before you go out. Our UK Digital and Tech Salary Survey gives you current benchmarks by role.
- Move fast. Strong AI candidates rarely wait around. A slow process is one of the main reasons top tech talent rejects job offers.
- Blend your workforce. Mix permanent, contract and project talent so you can scale up or down as your AI plans evolve.
Also read:How businesses can hire AI talent in the competitive market
Ready to hire AI engineers?
Tech giants pay millions because AI talent now decides who leads and who follows. You don't need their budget to compete, but you do need the right people at the right time.
For more than 45 years, Lorien has connected businesses with the tech talent that moves them forward. As a tech recruitment specialist, Lorien combines specialist AI recruitment with large-scale talent solutions.
Whether you need a single machine learning engineer or a full AI delivery team, the approach is shaped around your goals, not a one-size-fits-all process. Permanent hires, contractors, project delivery and AI skills programmes all flex as your plans evolve.
Talk to our AI recruitment specialists and find the AI engineers who'll move your business forward.
FAQs
What is the AI talent war?
It's the race between tech giants such as Meta, Google and Microsoft to hire a small pool of top AI specialists. Offers now reach tens, even hundreds, of millions of dollars.
How do businesses ensure AI delivers a return?
Start with one measurable use case, redesign the workflow around it and scale only what works.
Why are AI engineers paid so much?
Demand for AI skills is rising far faster than supply, so employers compete hard for the people who can turn AI into real results. PwC found that roles needing AI skills pay an average "62%" wage premium.
Is there an AI talent shortage in the UK?
Yes. The UK Government says closing the AI skills gap could unlock up to "£400 billion" in growth by 2030. Lorien's own salary data shows the AI, machine learning and automation sector has grown "50%" since 2024, so competition for experienced AI talent is intense.
How can a smaller business compete for AI talent?
Write a clear brief, sell the impact of the role, benchmark pay honestly and move quickly. Blending permanent, contract and project talent also helps you scale without overcommitting.
How much does an AI engineer cost in the UK?
It depends on experience, skills and location. According to our AI engineer salary survey, AI engineers at entry level earn “£35,000 to £50,000”, rising to “£60,000 to £90,000” at mid level. Senior specialists earn more, especially when their work drives clear business impact. For a full breakdown by role, and rate, download Lorien’s UK Digital and Tech Salary Survey.
How can Lorien help me hire AI engineers?
Lorien offers specialist AI recruitment across permanent, contract and project-based hiring, plus large-scale talent solutions for growing AI teams.
