We’re taking a look at the trends that matter most for leaders navigating the next 12–36 months.
We organize trends into three categories based on their maturity and impact timeline:
Game Changers — Trends reshaping entire industries right now. These affect how businesses operate, how value is created, and how competitive advantage is built.
Foundational Breakthroughs — Scientific and engineering advances unlocking new possibilities. These create strategic optionality for the decade ahead.
Weak Signals — Early indicators of transformations that will dominate the late 2020s. These require positioning now, even if mainstream adoption is years away.
Today, we’re diving into soverign stacks.
For decades, technology supply chains were optimized for efficiency. Design here, fabricate there, assemble somewhere else, ship everywhere. That model is being rewritten because efficiency is no longer the only thing governments are optimizing for.
Why now
This isn’t theoretical anymore. In just the past few months, escalating export controls, chip bans, and abrupt platform restrictions have made clear that shared global technology platforms can no longer be taken for granted, and governments are moving fast to reduce that exposure. Three pressures are converging behind the shift. Supply chain fragility, exposed during the pandemic, made clear how concentrated chip production was and how vulnerable every downstream industry was as a result. National security concerns followed close behind, as AI capabilities came to be seen less as a commercial advantage and more as a strategic asset. And economic competitiveness rounded it out: countries that can train and run their own AI models, on their own chips, are positioning to export that capability to others.
Countries are now building independent technology stacks from the ground up, from semiconductor fabs to the AI training infrastructure that runs on top of them. Since 2021, government incentives in the US have helped catalyze nearly $400 billion in announced semiconductor investments, with major production facilities breaking ground in Arizona, Ohio, and across Europe under parallel initiatives. The scale of capital being committed to domestic chip production hasn’t been seen in a generation.
The result is a layered build-out. Fabs at the bottom. Cloud and compute infrastructure in the middle. Sovereign AI models on top. Each layer reinforces the case for the others.
Recent movements
A few signals show how far this has already progressed.
China continues to invest heavily in domestic AI alternatives. DeepSeek is reportedly developing its own AI chip, aimed at inference rather than training, in a move that would reduce its dependence on Nvidia and Huawei silicon, a notable shift for a company built on squeezing more out of other people’s hardware. Alongside this, Harmony OS represents a parallel effort to reduce dependence on Western operating systems at the device level.
India is pursuing a similar strategy through BharatGen, a national generative AI initiative backed by the country’s Department of Science and Technology. The program has trained models on thousands of GPUs and introduced a 17-billion-parameter multilingual model built specifically for India’s languages and governance needs. More recently, BharatGen was named the anchor for India’s participation in Project Tapestry, a global AI Alliance consortium that lets nations fine-tune a shared base model on their own sovereign data while retaining full ownership of the result, an explicit hedge against depending indefinitely on AI systems trained on foreign contexts and governed elsewhere.
Europe and South Korea are moving at similar speed. The European Commission’s new technology sovereignty package pairs a “Chips Act 2.0” with a Cloud and AI Development Act aimed at building out domestic compute and reducing reliance on non-European providers. South Korea, meanwhile, has laid out more than $1 trillion in combined public and private investment behind what its government calls a “triple axis” strategy spanning semiconductors, physical AI, and data centers, with Samsung and SK Hynix anchoring new fabs at home.
And in the US, the shift is as much institutional as it is industrial. The agency formerly known as the US AI Safety Institute has been reformed into the Center for AI Standards and Innovation, with a mandate that includes representing US interests internationally and pushing back against what it considers burdensome foreign regulation of American AI technology. CAISI has since expanded its pre-deployment testing agreements to Google DeepMind, Microsoft, and xAI, alongside existing partners OpenAI and Anthropic, and has now completed more than 40 model evaluations, some in classified environments. The renaming reflects a broader pattern: countries aren’t just building their own technical stacks, they’re building the standards and institutions to defend them.
Why it matters
Technology, geopolitics, and economic leverage are converging into a single conversation. A chip is no longer just a chip, and a foundation model is no longer just a product. Each is also a statement about whose infrastructure, whose standards, and whose rules a country plans to operate under.
For global businesses, this means navigating an increasingly nationalized landscape, where the device manufacturer, cloud provider, the AI model creator, and even the operating system a customer uses may be shaped less by what’s best and more by where it was built. A strategy that assumes a single global technology stack is a blind spot.
Andrew Bolwell 