The artificial intelligence race is often described as a competition between models, chips, data centres and regulation. But one of the most important stories in AI right now is more human: the fight for talent.

The recent move of Noam Shazeer, Google Gemini co-lead and one of the most influential figures in modern AI, to OpenAI is not just another executive transition. It is a signal. In the frontier AI economy, a small number of exceptional researchers, engineers and technical leaders can shift the strategic balance between companies. Talent is no longer a support function in the AI race. It is the infrastructure behind the infrastructure.

For Europe, this matters deeply. The European Union has positioned itself as the global leader in AI regulation through the EU AI Act. It is investing in AI factories, gigafactories and industrial adoption. It is trying to build a credible AI sovereignty agenda. But none of that will be enough if Europe cannot attract, retain and scale the people capable of building and deploying frontier AI systems.

The next phase of AI competition will not be won by the jurisdictions that only regulate. It will be won by those that combine regulation, infrastructure, capital and talent into one coherent strategy.

Noam Shazeer is not a typical executive hire. He has been part of some of the most important technical shifts in AI, including work connected to the transformer architecture that helped define the current generation of large language models. His career spans Google, Character.AI, Google again, and now OpenAI.

That trajectory tells us something about the structure of the AI market. The most valuable AI talent is highly mobile, globally visible and strategically courted by the world’s most powerful technology companies. The competition is not only for engineers who can implement models. It is for people who understand how to define research direction, build elite teams, scale systems and turn scientific progress into product advantage.

In most industries, losing one senior executive matters. In frontier AI, losing one technical leader can mean losing years of accumulated research judgement, team credibility and strategic momentum.

That is why the talent war should not be treated as HR gossip. It is a market signal.

The AI talent market is no longer functioning like a normal labour market. Compensation packages for top AI researchers have reached levels that would have sounded absurd only a few years ago. Major technology companies are competing not just with salaries, but with equity, research freedom, access to compute, publication prestige, infrastructure, mission and long-term upside.

This is rational from the perspective of Big Tech. If a small team can improve model performance, reduce inference costs, design safer systems, build agentic workflows or create better enterprise products, the commercial upside can be enormous.

But this creates a structural problem for everyone else.

Startups, universities, public institutions and European scaleups are competing against companies with almost unlimited resources. The gap is not only about money. It is also about compute access, brand power, internal tooling, data scale and the ability to work on the most ambitious AI problems.

In practice, the AI talent war creates a concentration effect. The best-funded companies attract the best people. The best people improve the best models. The best models attract more customers, capital and talent. The flywheel becomes self-reinforcing.

Europe cannot ignore this dynamic.

European AI sovereignty is often discussed in terms of regulation, cloud, chips and compute infrastructure. These are all essential. But sovereignty also requires people.

A sovereign AI strategy without a talent strategy is incomplete. You can finance supercomputers, create regulatory frameworks and launch public-private initiatives, but if the most capable AI researchers, MLOps engineers, data infrastructure specialists, safety experts and product leaders leave for U.S. labs, Europe will remain dependent.

The EU AI Act creates a regulatory baseline for trustworthy AI. It provides a structure for risk management, transparency, high-risk systems and general-purpose AI obligations. This is valuable. But compliance frameworks do not build models, deploy agents, optimise inference pipelines or transform industrial processes.

That work is done by people.

Europe’s challenge is therefore not only to regulate AI properly, but to become a place where ambitious AI builders want to stay, build and scale.

The EU AI Act is a milestone. It gives Europe a clear legal framework at a time when many companies are asking for certainty around AI risk, transparency, accountability and compliance. For regulated sectors such as healthcare, finance, energy, public administration and industrial systems, this can become a competitive advantage.

Trust matters. Legal certainty matters. Responsible deployment matters.

But Europe must be careful not to confuse regulatory leadership with market leadership.

If the AI Act becomes the only distinctive pillar of Europe’s AI strategy, the EU risks becoming the world’s compliance layer while other regions build the core platforms. The danger is not regulation itself. The danger is regulation without execution capacity.

The key question is not whether Europe can create good AI rules. It already has. The question is whether Europe can build globally competitive AI companies, retain technical talent and deploy AI at scale across its industrial base.

That is where the talent war becomes central.

The AI talent problem is not limited to frontier labs. It also affects enterprise adoption.

Many companies across Europe are not struggling because they lack curiosity about AI. They are struggling because they lack internal capability. They do not have enough people who understand AI strategy, data readiness, workflow redesign, model evaluation, governance, security and change management.

This is especially important for SMEs. Large corporations can hire consultants, build AI centres of excellence and pay for premium tools. Smaller companies often lack both the budget and the internal skills to move from experimentation to deployment.

That creates a productivity gap. AI adoption becomes concentrated in companies that already have digital maturity, technical teams and access to capital. The rest of the market risks being left behind.

For Europe, this is not only a technology issue. It is an economic competitiveness issue.

If AI adoption remains uneven, Europe’s productivity challenge will deepen. If only a small group of large companies can deploy AI effectively, the broader economy will not capture the full value of the technology.

Europe does not need to copy Silicon Valley. But it does need a more ambitious AI talent strategy.

First, Europe needs to develop frontier talent. That means supporting world-class research environments, competitive compensation, access to compute and stronger links between universities, startups and industry.

Second, Europe needs to retain technical leaders. Many European researchers are trained in excellent institutions but leave when they want to build at scale. The issue is not only salaries. It is the availability of ambitious companies, fast capital, technical infrastructure and cultural permission to pursue large-scale bets.

Third, Europe needs to scale applied AI talent. The biggest opportunity for Europe may not be in building the most famous consumer chatbot. It may be in deploying AI across manufacturing, logistics, mobility, health, energy, finance, public services and cybersecurity. That requires AI product managers, data engineers, solution architects, compliance-aware technologists and business leaders who understand AI well enough to make strategic decisions.

Fourth, Europe needs to upskill non-technical professionals. AI transformation will not be delivered only by machine learning engineers. Legal teams, marketing teams, HR teams, operations teams, public-sector leaders and executives all need AI literacy. The companies that win will be those that distribute AI capability across the organisation, not those that isolate it in a technical department.

The European Commission’s AI Continent strategy, AI factories and AI gigafactories are steps in the right direction. Infrastructure is essential. Europe needs compute capacity, data access, secure cloud environments and industrial deployment channels.

But infrastructure without builders is underutilised capacity.

The strategic question is not only whether Europe can finance AI infrastructure. It is whether Europe can build the talent ecosystem around that infrastructure. Who will train the models? Who will fine-tune them for European industry? Who will evaluate risks? Who will build products on top of them? Who will help SMEs adopt them? Who will connect AI capability with real business value?

This is where public policy, education and the private sector need to operate together. AI talent cannot be created by universities alone. It cannot be retained by public grants alone. It cannot be scaled by Big Tech alone.

Europe needs a coordinated talent pipeline: education, research, startup creation, corporate adoption, mobility, immigration, funding and procurement.

European companies should not wait for policymakers to solve the talent gap. They need to act now.

AI should become a board-level talent priority. Companies should map which AI skills they need, which they already have and which they should build internally. They should identify where they depend too heavily on external vendors and where internal capability is strategically necessary.

They should also rethink hiring. Not every company needs PhD-level AI researchers. Many need practical AI operators: people who can redesign workflows, evaluate tools, manage data, connect APIs, build automations, govern risks and train teams.

The most competitive companies will combine three layers of talent: senior technical expertise, business-side AI literacy and operational adoption capability. Without that combination, AI remains a demo. With it, AI becomes transformation.

The AI talent war is not only about recruiting famous researchers. It is about organisational capability.

The companies that create real value from AI will not necessarily be those with the most tools. They will be those with teams capable of asking better questions, redesigning processes, integrating systems and measuring outcomes.

This is especially relevant in Europe. The EU has deep industrial sectors, strong universities, regulated markets and sophisticated public institutions. Its advantage may come from applying AI in complex, high-trust environments where domain knowledge matters as much as model performance.

That requires a different kind of AI talent: not only model builders, but translators between technology and business. People who understand compliance, customer experience, operations, cybersecurity and sector-specific constraints.

Europe’s AI opportunity is not only to build models. It is to build deployable intelligence for the real economy.

Noam Shazeer’s move to OpenAI is a timely reminder that AI power is concentrating around people as much as platforms. Capital matters. Compute matters. Regulation matters. But talent is the multiplier that turns all of them into advantage.

Europe has a credible regulatory framework. It is investing in infrastructure. It has strong research institutions and industrial depth. But if it wants to become a serious AI power, it must treat talent as strategic infrastructure.

That means competing for senior builders. It means creating conditions for AI startups to scale in Europe. It means helping companies move from AI curiosity to AI capability. It means connecting regulation with execution, and education with deployment.

The AI talent war is not a side story. It is the core story.

Because the future of AI will not only be built by the companies with the biggest models. It will be built by the teams with the people capable of turning intelligence into real-world advantage.

Europe’s next challenge is clear: it cannot regulate its way into AI leadership. It has to build the people, companies and ecosystems that make leadership possible.