The United States’ decision to intervene in Venezuela and bring a foreign leader before its courts reflects a wider shift in the character of the international order. Western power is no longer exercised solely through deterrence or diplomacy, but through increasingly direct forms of intervention.
Yet behind this display of strength lies a more troubling reality. In classified Pentagon war games simulating a conflict with China, US forces lose every time.
These simulations are not abstract. As Russia probes Nato’s vulnerabilities and China signals its intent to seize Taiwan within years, they reveal not a lack of Western spending power, but a dangerous dependence on systems that take too long to build and cannot be produced at scale. The imbalance is stark. China now possesses roughly 600 hypersonic weapons while the United States has yet to deploy a single operational missile.
As Nato embarks on its largest rearmament cycle in a generation, the risk is that it repeats the very weaknesses these war games expose. This is not simply a contest of spending, but of fundamentally different approaches to building capability. The decisive question is whether the West can innovate and deploy quickly enough to sustain a prolonged conflict.
Here the signs are troubling. Historian Chris Miller’s recent warning to the US Senate that America’s advantage in AI talent is ‘deteriorating dangerously’ reflects broader anxieties that our ability to generate the latest technologies critical to military power is beginning to weaken.
Miller is right to sound the alarm, after all, China is now producing more AI researchers than the US. Yet Miller’s diagnosis misses a fundamental part of this story. China’s true advantage lies not in innovation, but in scale, and scale alone doesn’t win modern wars.
During a recent visit, what struck me was how Chinese cities function as living factories, with producers clustered across entire supply chains. Shenzhen hosts more than 1,200 robotics companies, supported by state infrastructure and local governments spending the equivalent of £110 billion on business parks last year.
This capacity matters enormously in a war of attrition. In Ukraine, thousands of artillery rounds and drones are consumed daily. Meanwhile, Britain’s stockpile could be exhausted within just eight days.
But these same top-down structures that enable scale also constrain experimentation and risk-taking. China’s 2023 AI rules require public-facing or opinion-shaping models to pass state security assessments and comply with ‘core socialist values’, limiting how even advanced systems can be deployed.
By contrast, the West’s historic military and technological advantage has always rested on openness. During the Cold War, many of the breakthroughs that reshaped civilian life and military capability, such as the transistor, emerged from independent labs. That openness defined the rise of the global software ecosystem, built in public on technologies like Linux and Python rather than instruments of the state.
The West led the software era because it allowed people to try things. Today, almost 90% of modern software contains open-source components and open-source startups are 20% faster at reaching Series A funding (the first round of venture capital funding). While Silicon Valley mythology can be overplayed, the ecosystem that enabled this innovation hasn’t vanished: Y Combinator’s median founder age has fallen to just 24.
Recent interventions may project strength, but they do not resolve this underlying weakness. The danger, highlighted by the Pentagon’s war games, is that Western governments respond to China’s scale by abandoning this advantage and doubling down on central planning. Unfortunately, that is already becoming reality. Across advanced economies, industrial interventions have risen sharply, from just over 200 globally in 2017 to more than 1,500 by 2022.
The Trump administration’s move to take a federal multibillion-dollar stake in Intel effectively amounts to treating it as a national champion. Biden, in rolling out hundreds of billions in subsidies through the CHIPS Act and the Inflation Reduction Act, layered programmes with detailed rules and restrictions on supply-chain sourcing, domestic content and more.
The problem isn’t just in the state slowing innovation down, but also in the potential waste of taxpayers’ money. Germany’s defence surge illustrates this exact problem. Berlin now plans to spend €649 billion within five years, despite its defence industry warning that it cannot absorb spending at anything like this scale. As a result, much of the investment is already flowing abroad rather than backing domestic production. Money cannot compensate for a lack of capacity or innovation, and mimicking China’s state-led mobilisation risks replicating its inefficiencies without any of the benefits.
Procurement systems that favour large, established contractors over speed or adaptability reinforce this weakness. The number of major US defence contractors has fallen from more than 50 in the early 1990s to just five today, concentrating power in firms whose incentives reward long timelines. No wonder programmes like the F-35 took roughly two decades to reach production.
Modern wars are not just won with the most advanced missile system, but by sustaining production and adaptation over time. This is precisely where the Pentagon’s assessments show we’re falling behind. I am involved, as an investor and strategic adviser, with several US-based companies focused on scalable, cost-effective approaches to defence manufacturing. That is the kind of capability the current system struggles to generate.
As such, the West is at an inflection point. The drift toward central planning reflects a deeper misunderstanding of what has made liberal societies competitive. The race ahead is not to replicate China’s command-and-control structure, but to build the future industries, from AI-enabled systems to advanced manufacturing, that will define military and economic power over the next 50 years.
That means backing younger founders earlier and more boldly. Institutions like Y Combinator demonstrate how powerful this model can be: by taking risks the state never could, they move faster than any government ministry and identify transformative ideas long before they become obvious.
This next phase of military competition won’t be decided by who spends the most, but by who can turn ideas into usable capability the fastest and sustain it. Deterrence depends on convincing countries like China and Russia not only that we can fight, but that we can endure – a reality Nato ignores at its peril.
