I separate the AI investment bubble from the longer-term AI transformation. Company outcomes and investment returns are not the same question as whether AI will reshape businesses and economies; I focus on who may capture the value created by that transformation.
I expect that many companies will fail, valuations will reset and billions may be lost; AI itself isn't going anywhere and will fundamentally reshape businesses and economies.
I frame the central question as: Who will capture the value created by this revolution?
I see three interesting shifts in the AI race.
I see the race as no longer just about AI models; it's about compute, capital, energy and infrastructure.
I expect autonomous agents and multi-agent systems will increase demand for compute, networking, storage and power.
I see infrastructure could be the biggest bottleneck; data centres need electricity, cooling, connectivity and specialised engineering.
I believe India doesn't necessarily need to build the world's largest AI models to benefit; there could be significant opportunities in the ecosystem powering the revolution.
- I am looking at enabling infrastructure such as transformers, switchgear, fibre connectivity, cables, cooling, grid infrastructure and energy storage.
I have been researching and investing in Indian companies around the picks and shovels of AI infrastructure.
My takeaway is that a technology can transform the world even when many companies investing in it fail to make money; sometimes the biggest opportunity is in building everything the revolution needs to scale.
My investment lens is to investigate the infrastructure and suppliers that let AI scale, rather than focusing only on companies building AI technology. This is a thesis to research, not a claim that any particular company or component will succeed; I have not named companies or specified a screening method.