As Hormuz rattles global markets, these are the AI startups worth watching
Warren Buffett once observed that “only when the tide goes out do you discover who’s been swimming naked.” It is an overused line, but for good reason. Periods of market stress reveal what rising markets allow investors to ignore.
That is why the current geopolitical moment matters. Escalating tensions in the Middle East, and the possibility of confrontation involving the United States, are not just political developments. They have direct implications for global markets. The Strait of Hormuz remains one of the world’s most important energy chokepoints, and even limited disruption feeds quickly into volatility across commodities, inflation expectations and growth forecasts.
There is no need to assume the most extreme outcomes. But the fact that markets are even discussing them tells you something about the mood. Capital is becoming more cautious and more selective. If that shift continues, parts of the AI market may struggle.
Over the past two years, AI has attracted extraordinary levels of investment, driven by a belief in its transformative potential. That belief is justified. Long before the arrival of ChatGPT, I spoke about the transformative potential of this technology. But alongside that conviction, there has also been a tendency towards indiscriminate allocation, where in some cases the label has mattered more than the underlying business.
Even senior figures within the technology sector have begun to acknowledge this risk. Their remarks reflect a broader reassessment already underway. None of this undermines the long-term case for AI, but it does suggest that parts of the market have run ahead of themselves. In a tighter environment, that becomes harder to sustain. As Buffett’s analogy suggests, these are the moments when weaker propositions are exposed.
This is not a moment for fear, but for judgment. The question for investors is not whether AI will matter, but which parts of the ecosystem will still matter if conditions become less forgiving. History tends to favour companies solving problems that persist regardless of sentiment, not those benefiting from it. That is how I approach AI investing today.
At Kohli Ventures, the focus is not on backing AI as a broad theme. It is on identifying businesses addressing structural constraints within the emerging AI economy, where demand is likely to endure.
The companies easing AI’s hardest bottlenecks
Much of the durable value in AI sits below the surface, in infrastructure.
Voltai is a good example. While attention has focused on applications, progress is still constrained by hardware. Designing advanced chips remains complex, slow and expensive. Voltai uses AI to help engineers move faster, identify problems earlier and make better use of scarce expertise.
Its traction with major semiconductor, electronics and automotive groups reflects the importance of that constraint. In a more selective market, companies solving these kinds of foundational problems tend to hold up better.
A similar issue exists in engineering design. Innovation is often limited not by imagination, but by the time and cost required to test ideas. Godela is trying to change that, using AI-driven physics simulation to allow faster, cheaper experimentation. Shorter development cycles translate into better decisions and lower costs, which matters more, not less, when conditions tighten.
Building trust into autonomous systems
If infrastructure is one layer of value, trust is another. Despite rapid progress, many AI systems still struggle with reliability. They repeat mistakes and fail under pressure, and until that improves, adoption at scale will remain constrained.
Judgment Labs is focused on that gap. Rather than building more agents, it is building the systems needed to evaluate and monitor them, with the aim of making AI more reliable and therefore more useful. As AI becomes more embedded in decision-making, that kind of capability becomes essential.
The same logic applies to autonomous systems more broadly. Before they can be deployed at scale, they need environments where they can be tested and refined safely. Archer AI is building in that space, applying simulation techniques already used in autonomous driving to a wider range of systems. It is the kind of infrastructure that becomes indispensable as the technology matures.
Why broader access matters
There is also a third dimension to long-term value: accessibility. The future of AI will not be determined by hyperscalers alone, but by whether the technology becomes useful to a much wider group of builders.
Vibecode is an example of that shift. It allows users to describe what they want to build and turn that into a working application from their phone. Early traction suggests there is real demand for tools that lower the barrier to creation. AI becomes easier to justify as an investment when it is widely used, not just heavily funded.
What outlasts the hype
What unites these companies is not visibility, but utility. If markets become more volatile, the most inflated parts of the AI trade may come under pressure. That would not be unusual. Periods of rapid innovation are often followed by adjustment, and what tends to endure are businesses solving real problems, particularly those tied to infrastructure, efficiency and trust.
The situation around Hormuz may or may not escalate further, but it is a reminder that conditions can change quickly. When they do, markets become less forgiving, and the difference between narrative and substance becomes harder to ignore. For investors, that is the only distinction that matters.
Tej Kohli is a philanthropist, technologist and investor, and the founder of Kohli Ventures, through which he backs high-growth companies in sectors including AI, robotics, biotech and esports.
