For the past two years, the dominant narrative about AI has been one of boundless possibility. Larger models, trillion-token training runs, and record-breaking capex (capital expenditure) cycles have reinforced a sense of uninterrupted acceleration. But technological change is rarely so straightforward, and this time is no exception. As AI moves from experimentation to real-world applications, the limits imposed by the physical world, capital markets, and political systems clearly matter more than its theoretical potential.The most immediate constraint is electricity. Nowhere is this more evident than in the US, where data-center power demand is expected to rise from roughly 35 gigawatts to 78GW by 2035. Northern Virginia, the world’s largest cloud-infrastructure cluster, has already effectively exhausted its available grid capacity. Utilities in Arizona, Georgia, and Ohio warn that new substations may take almost a decade to build. A single campus can require 300-500MW, enough to power an entire city. Silicon can be manufactured quickly; high-voltage transmission cannot.Markets are responding with the speed and ambition one would expect. Hyperscalers (the major tech firms building advanced AI models on the back of ever-greater computing capacity) have become among the world’s largest buyers of long-dated renewable energy. Private solar and wind farms are being built expressly to serve cloud facilities, and some firms are exploring next-generation small modular reactors as a way to bypass slower municipal infrastructure.These efforts will eventually expand the frontier of what is possible, but they do not eliminate the constraint so much as redirect it. The next wave of AI capacity will likely be concentrated not in Northern Virginia or Dublin, but in regions where land, power, and water remain abundant: the American Midwest, Scandinavia, parts of the Middle East, and western China. The geography of AI is being written by physics, not preference.Silicon is the next constraint, and here the story is becoming more complicated. While Nvidia once appeared to be the universal substrate beneath all AI development globally, that era is ending. In a significant milestone, Google trained its latest large language model, Gemini 3, entirely on its own Tensor Processing Units – and Amazon’s Trainium2, Microsoft’s Maia, and Meta’s MTIA chips are all being developed for similar purposes. Similarly, in China, Huawei’s Ascend platform has become the strategic backbone for domestic model training in the face of US export controls.Some of this shift reflects natural technological maturation. As workloads increase, specialised accelerators become more efficient than the general-purpose GPUs originally adapted for AI. But the timing is not accidental. Scarcity, geopolitical friction, and cost pressures have pushed hyperscalers to assume a role once reserved for semiconductor firms. Given that departing from Nvidia’s CUDA ecosystem carries enormous organisational costs, the growing willingness to incur it signals how severe the constraint has become. What will follow is a more fragmented hardware landscape, and with it, a more fragmented AI ecosystem. Once architectures diverge at the silicon level, they rarely reconverge.The third constraint, capital, operates in a more subtle way. Hyperscaler investment plans for 2026 now exceed $518bn, a figure that has risen by nearly two-thirds just in the past year. We are already witnessing the largest private-sector infrastructure buildout in modern history. Meta, Microsoft, and Google revise their capex guidance so frequently that analysts struggle to keep pace.Yet it is still early days for economic returns. Baidu recently reported CN¥2.6bn ($369mn) in AI-application-related revenue, driven largely by enterprise contracts and infrastructure subscriptions, and Tencent says it has lifted profitability through AI-enhanced efficiencies across its mature businesses. But in the US, most companies still bury their AI earnings within broader cloud categories.The gap between AI adoption and monetisation is wide but familiar. In past technological waves, infrastructure spending routinely preceded productivity gains by years. The constraint comes not from weak investor sentiment, but from the strategic pressure enthusiasm creates: different firms pursue different conceptions of value because their business models and cost structures demand it.Many sectors simply cannot adopt AI at the pace that new models are being released. Large banks, for example, remain bound by security and compliance frameworks that require air-gapped, on-site, fully auditable software deployments. Such rules instantly cut them off from the most advanced frontier models, which rely on cloud-side orchestration and rapid iterations through new versions. Health-care systems face similar limits, and governments even more so. The problem is not AI’s theoretical capabilities, but the difficulty of incorporating such tools into legacy systems built for a different era.Taken together, these forces suggest a future very different from the one implied by the standard media narrative. AI is not converging toward a single universal frontier. Diverse regional and institutional architectures are being shaped by different limits – from power shortages in the US to land and cooling constraints in Singapore and Japan, “geopolitical” scarcity in China (where Western export controls limit access to advanced chips and cloud hardware), regulatory friction in Europe, and organisational rigidities across the corporate world. Technology may be global, but implementation is local.Fortunately, real-world constraints are not the enemy of progress. Often, they form the scaffolding around which new systems take shape. The fibre-optic glut of the late 1990s, initially derided as wasteful overshoot, later underpinned the rise of streaming, social media, and cloud computing.Today’s constraints will play a similar role. Power scarcity is already shifting the geography of AI. Silicon fragmentation is creating new national and corporate ecosystems. Capital asymmetries are pushing firms into different strategic equilibria. Institutional limits are shaping the first real use cases.The next decade of AI will belong not to the systems with the greatest theoretical capability, but to the ecosystems most adept at turning real-world limits into design advantages. Possibility defines the horizon, but constraint will determine the route the world ultimately takes. - Project Syndicate(Jeffrey Wu is Director at MindWorks Capital.)

Few would deny that there has been a shift away from multilateral co-operation in recent years. As the world becomes more multipolar, geopolitical tensions are hampering efforts to devise common solutions to shared problems, and rising nationalism and fiscal crises within many traditional donor countries are threatening the institutions on which multilateralism depends.As a realist, I recognise that today’s world is more dangerous than the one we inhabited not so long ago. But I am also confident that possibilities for long-term global collaboration remain. I have seen firsthand that multilateral co-operation often delivers results that otherwise would not be attained. My confidence stems from my experience as the chair of Gavi, the Vaccine Alliance. As my five-year tenure draws to a close, I find myself reflecting on what has underpinned Gavi’s success over the past 25 years and what this experience can teach us about adapting multilateralism for a rapidly changing world.The first lesson may sound simple, but it is too often forgotten: Always be mission-driven. Gavi exists to save lives and protect health by expanding access to vaccines in lower-income countries. It is this clarity of purpose that has helped halve child mortality in 78 countries and protect every one of us against the threat of infectious diseases. Nor is there any secret to our success. We have done it by uniting a multitude of public and private stakeholders, many with divergent interests, behind a common purpose.Gavi has always been a coalition of the willing, bringing together national governments, United Nations agencies, philanthropies, vaccine manufacturers, innovators, development banks, research institutions, and civil society. With its diverse skill set, expertise, and political clout, it has protected over half the world’s children against preventable diseases in any given year, as well as providing the world with core competencies during crises like the Covid-19 pandemic, when we led the global vaccine response.In a more multipolar world, similar approaches will be needed to drive progress in other areas where the provision of public goods (conflict resolution, education, health security, equitable access to AI) is too important to be held hostage by adversarial politics and sectional interests.That leads me to the second key lesson: Be mission-driven, but country-led. Gavi was founded in the spirit of partnership, not paternalism. Promoting national self-reliance has always been at the heart of its mission. Countries pay more toward the cost of their vaccine programs as their national incomes rise, up to the point where they can fully sustain their own immunisation services. Some countries have even transitioned from recipients to donors.This responsiveness to country needs has made us relentlessly focused on innovation. In 2024, Gavi embraced the historic introduction of malaria vaccines because we recognised how unjust it was that so many countries, particularly in Africa, had to wait so long for such a breakthrough. The same year, Gavi also launched a financial innovation, the First Response Fund, to provide surge financing for the procurement of mpox vaccines, saving precious time that otherwise would have been lost raising additional funding.Today, Gavi is directing the same innovative zeal toward the future rollout of vaccines against tuberculosis, the world’s deadliest infectious disease. It is also advancing a new initiative, the African Vaccine Manufacturing Accelerator, with strong backing from the European Union and other donors, to support the African Union’s ambitions for regional high-value manufacturing. I predict we will see a far greater role for, and collaboration between, regional economic and political blocs as the key drivers of multilateralism in the years ahead.Every coalition needs strong governance and leadership. That is why Nelson Mandela was chosen as Gavi’s first chair. But ensuring that the interests of every stakeholder remain aligned is no simple task, and this insight was not lost on me when I was approached for the role in 2020. I was honoured, and I could not help noticing that the Gavi board had 28 seats, the same number of member states whose interests I sought to align when I was president of the European Commission.Throughout my tenure at Gavi, I have been guided by the enduring wisdom of Jean Monnet, a leading postwar advocate of European unity: “Nothing is possible without people, but nothing lasts without institutions.” Gavi is truly a unique institution. Not only is it a broad, inclusive alliance of national and international, as well as public and private, entities; it is also an international organisation that has managed to avoid paralysis and inertia, unlike some major intergovernmental bodies. It has done so by maintaining a laser focus on protecting children – even in war zones where the only respite from fighting came from the need to vaccinate populations.Countries will always have reasons to disagree, but if anything can elevate the cause of peace above extreme national interest or ideology, it is the protection of children. Throughout my life, I have been at the heart of many seismic changes, from the Carnation Revolution in my native Portugal in the 1970s to the effort to advance peace, reconciliation, and democracy in Europe (for which I had the great honour of receiving the Nobel Peace Prize on behalf of the EU). In each case, historic changes needed a catalyst, which is exactly the role that Gavi has played in promoting public health.As we enter a more multipolar world, I would urge everyone to recognise the need for more mission-driven public-private partnerships like Gavi. There simply is no better way to address the challenges of our age. — Project Syndicate • Jose Manuel Barroso, a former president of the European Commission and former prime minister of Portugal, is Chair of the Board of Gavi, the Vaccine Alliance.