Behind the apparent simplicity of AI, a global industrial and geopolitical battle is being waged. US leadership, the Chinese offensive, critical dependencies and European decisions: what are the geopolitical trends redefining the uses of data and AI? Insights from Aurélien Duthoit, sector economist at Coface, to help business decision-makers secure their growth in an era of major technological transformation.
In the world of tech (and far beyond!), the spotlight is now firmly on data and artificial intelligence (or almost exclusively so): developments in capabilities, uses, and expected or proven effects on business productivity and the labour market. Given the huge sums it mobilises and its potential, artificial intelligence (AI) is also becoming a matter of economic power, industrial power and, increasingly, geopolitical power.
Acting as a genuine catalyst for the acceleration of new technologies, geopolitics is redefining the uses of data and AI. Behind an instant and almost intangible technology that we generally access via a simple chat window lie, in fact, hundreds of billions dollars invested in infrastructure, increasingly complex financing structures and value chains that are particularly concentrated around a select few major players.
AI is no longer just a matter of innovation or productivity. It has become an instrument of economic, industrial and geopolitical power,
states Aurélien Duthoit, sector economist at Coface.
1 – Is US hegemony as solid as it seems?
Following the US company Nvidia, which has the world’s largest market capitalisation ($2,000 billion), Anthropic and OpenAI have already announced their IPOs for the coming months. This is a major development, given that Anthropic’s stock market entrance is unanimously being described as the biggest IPO of all time!
This represents yet another competitive advantage for US companies in the sector, which currently account for 70% of the total valuation of the global AI financial ecosystem. This is in addition to the US monopoly on infrastructure (servers, data centres): indeed, 50% of current capacity is based in the United States, and 70% of capacity currently under construction is also located in the United States.
Whether in terms of financial valuation or infrastructure, US leadership is extremely strong. And when we look at the flows, everything suggests that this trend will only become more pronounced.
explains Aurélien Duthoit.
This leading position, however, masks a major industrial dependency: the US remains heavily reliant on imports of key machinery and components needed to power these AI-related technologies. This is an exponential trend, which has quadrupled in recent years. The paradox is this: how can a country dominate a technology to such an extent whilst remaining dependent on other countries for such critical components?
At the heart of the new technological geopolitics, a new paradigm is therefore emerging: a country can concentrate value, software and capital without controlling all the components necessary for their operation. For companies, this fragmentation turns the choice of a technology supplier into a strategic decision.
In return for their hegemony in the sector, we are seeing a strategic weakness in the US supply chains. The further down we go into the physical layer of this economy (not just microchips), the more apparent US vulnerability becomes.
The Americans are now investing significantly to reduce these dependencies, but the pace is not as fast as expected to truly rebalance the power dynamics.
observes Aurélien Duthoit.
2 – China-US Tech War: who will really win the battle of the models?
Eighteen months ago, the ‘DeepSeek moment’ served as a reminder to all players in the Tech sector that competition is not only about a race for investment to develop the most powerful language model. The real challenge also lies in the balance between performance, costs and deployment conditions. This is all the more true as China is fighting back and, moreover, pursuing a different strategy. Several of its models are ‘open’: they can be deployed outside the developers’ premises, within the IT environment chosen by the company, including locally or with a European hosting provider. This approach contrasts with closed offerings that bundle the model, service and infrastructure.
By delivering high-performance models that are more accessible, Chinese players are not necessarily seeking to win the race on their own. They can also prevent their American competitors from turning their technological lead into a global rent and a lever of influence.
The current issue is no longer a question of who has the best technological models, but who offers the best quality-price ratio on the market.
For China, the geopolitical interest lies in reducing the ability of US players to turn their technological lead into a global rent. They are also reducing the geopolitical leverage the US could wield over the rest of the world.
Aurélien Duthoit.
3 – Europe: not sovereign, but not defenceless?
In this new geopolitical-technological paradigm, what role is there for Europe? The major digital markets are moving towards oligopoly: language models due to network effects, and the cloud due to economies of scale. For any company adopting AI, the risk of supplier concentration is therefore very real. When it comes to components and models, Europe is lagging behind, a situation stemming directly from more general weaknesses (regulation, integration, vision) already highlighted in the Draghi report.
And yet, when faced with the major powers of China and the US, Europe is by no means defenceless. It occupies key positions in the value chain that it could capitalise on more effectively:
- ASML for photolithography,
- Zeiss for optical inspection,
- Spea for electrical testing,
- Siltronic for silicon wafers.
Europe can also capitalise on the size of its market – the second largest in the world – which could become a real bargaining chip.
For Europe, as for its companies, four questions remain to be addressed:
- What am I dependent on?
- How critical is it?
- What alternatives are available?
- And what is the cost of switching from one to another?
One imperative remains: not to fall behind in adopting new technologies, whose productivity gains could restore margins and competitiveness.
Data & AI: move from experimentation to impact
> What if managing risk actually enabled you to seize more opportunities? Head to your local Coface website to download the complete study and access the Open Advantage, five key strategies to unlock your growth
