Professional players of Texas hold ’em, the most popular variant of poker, go to great lengths to play a balanced, unexploitable strategy. They mix value bets and bluffs in every single hand. The bluffs are not usually to make money. They are there to stop other players adjusting profitably to the game. If another player over-folds to avoid losing to your value bets, only then do your bluffs go from break-even to hugely profitable. Against strong opponents, a balanced strategy is essential.
National strategy likewise should strive for balance, particularly from the perspective of a weaker country. Since it’s almost impossible for a mid-sized open economy to be self-sufficient, better make sure you don’t depend on just one key partner, especially if you have minimal leverage over them and reason to doubt their long-term goodwill. Britain and Europe’s national strategies are both hopelessly dependent on America, while cutting themselves off from any alternatives. The recent Democratic primary election results in New York, where the socialist candidates, endorsed by New York mayor Zohran Mamdani, swept the board against the mainstream, are just another sign of how ruinous this strategy might prove.
Access to the most advanced American models is now explicitly controlled by the White House
The problem is particularly bad for Britain. By 2060 the world will probably be dominated by three great superpowers: America, China and India. If nothing changes, all three will be either indifferent or actively hostile to British interests. America’s socialist left, with its Third-World-ist sympathies, has no love for a former colonial superpower. The MAGA right likes Britain in theory but not in practice. It regards our elites as compromised Eurocrat sellouts in disguise. Even a future Reform government might struggle in the face of this antipathy, especially from the younger generations of American right-wingers. Donald Trump could well be the last intuitively Anglophile president, and even this has bought us precious few favours.
Britain’s relationship with India will always be complicated by the legacy of Empire. Modern-day Indian nationalism, championed by Narendra Modi, is built on the rejection of the colonial legacy in all its forms. It will never be easy to have a good relationship with a superpower where ‘Britain bad’ is a core part of the ideology of both people and elites.
This leaves us with China, which is rapidly becoming the global leader in technological innovation, from biotechnology to cars to robotics to AI. China is perhaps the only superpower whose elites do not have a fundamental aversion to Britain on historical-ideological grounds: the Opium Wars were a long time ago, and the ‘century of humiliation’ was something that happened to a different regime that, in its own historical understanding, the Chinese Communist party rescued China from. Like Saruman gazing through the palantir at the armies of Mordor, it is hard not to conclude that against the power rising in the East there is no victory, and it would be wise to extend the hand of friendship.
The advent of artificial intelligence makes a reconciliation with China mission-critical for both Britain and Europe. AI will be the bedrock of the future economy. Its importance could outstrip transformational technologies of the past such as the steam turbine, electricity, semiconductors or even the internet itself. If the continent wants alternatives to the frontier models released by the major American labs, there is no alternative but to look to China, whose elite universities churn out technical talent Europe can never hope to match. Since the launch of ChatGPT sparked the generative AI boom, Europe has managed to produce only Mistral, a French start-up founded by a team of ex-Meta and DeepMind researchers, backed by the French government and ASML, the Dutch lithography company. Mistral is valued at around $14 billion, while OpenAI and Anthropic close in on trillion-dollar valuations.
More importantly, Mistral’s models are feeble compared with what Chinese companies have managed to achieve with similarly onerous computing constraints. China has at least four or five first- or second-tier start-ups (DeepSeek, Moonshot AI, Zhipu AI, MiniMax and StepFun) plus more richly funded labs at its big tech companies, most notably ByteDance and Alibaba. Byte-Dance is the world leader in image and video generation, while DeepSeek, Moonshot and Zhipu AI are second only to OpenAI, Anthropic and Google in producing top-tier models for coding and research.
Access to the most advanced American models is now explicitly controlled by the White House. Anthropic’s top-tier ‘Fable’ model and GPT 5.6 have been shut off to foreign nationals. This sudden development provoked a great deal of comment about the possibility of Europe building its own frontier model, since getting stuck with old models would lead to economic catastrophe. Unfortunately, the money is not there. Europe lacks America’s ultra-deep private capital markets to fund such an endeavour and too many of its governments are broke. Those that are not broke, such as Germany, are instead pouring money into rearmament. Crucially, Europe lacks the necessary talent. American tech companies have been strip-mining Europe for talent for more than ten years.
But through a deep strategic partnership with China, Europe can still build a balanced strategy and avoid perpetual vassalage to the US. China needs more from us than America and Europe has great but desperately underused bargaining power. The underfunded Chinese AI start-ups need money and Huawei needs ASML’s cutting-edge EUV machines to manufacture top-of-the-line chips. Everyone needs more data to train bigger and better models, both to transform white-collar work but also for the coming wave of humanoid robots. In a continent crippled by high labour costs and inflexible labour markets, humanoid robots could offer European industrial competitiveness a much-needed boost. Policy-makers will find the coming robot wave much easier to accept if it is genuinely a Sino-European co-creation rather than just another external shock.
Comments