I had a good chat with Alex Fink at the end of Feb. Alex is Principal at Empirical Ventures, which has a particular focus on healthcare, and he also invests through a family office, so he's sitting across both specialist science bets and broader market opportunities at the same time. It's why I wanted to get his view on where things are heading this year.

The AI moat problem

Everyone is chasing AI deals right now. But Alex said something I think a lot of people are thinking and maybe just beginning to say out loud.

There is a danger with AI companies that because AI has gotten so advanced, the defensible moat around whatever you're creating just isn't as much there as it probably once was. You really need to look for specialist, niche things where people have extra experience or expertise that other people don't have.

He pointed to one of Empirical's portfolio companies, Lucida Medical, as a concrete example. They've built a tool that reads MRIs as an additional check for doctors diagnosing prostate cancer. It's already picked up cancers that doctors missed. The reason it's hard to replicate isn't the technology in isolation, it's the expert in-house knowledge of what is and isn't cancerous on an MRI. Empirical currently has three AI companies in the portfolio, and this is the kind of specificity Alex is looking for across all of them.

2026 is more about compute than AI

When I asked Alex how this year compared to last, he went straight to compute rather than AI.

I think this year has really been the year of advanced compute, more than just AI. There's been some really interesting deals in the quantum space, in the photonic space, and around chip design, chip engineering, new novel chip manufacturers, and the infrastructure and hardware around that. There's a lot more going into hardware than previously.

On the macro

We talked through the current environment. Equities off their highs, gold up, bonds flat, rates starting to move.

Sometimes when there's a lot of volatility, quite counter-intuitively, investing in startups brings some stability to your portfolio, because you know that you're not marked to market every day. It's an investment that lasts a couple of years. You can ride out cyclical waves.

He also talked about how the post-2021 reset has changed the quality of what's coming through. Leaner teams, more focus on profitability, better entry valuations. "The wins now are more about landing clients or doing something great, rather than expanding into five new territories and getting this many users."

Robotics

When I asked what was on his watchlist for the rest of the year, Alex focussed on robotics.

Hardware is hard, and hard to scale. But you really need to look for something that has a massive need, and something that will be game changing on the frontier science. If you look at Chinese New Year, Chinese showcasing their robots was the highlight of the ceremony. Elon Musk is going to unveil his Optimus this year. I think there'll be lots of startups coming out that can maybe optimise the robots, or do something that can be an add-on into them.

Alex called it a potential GPT moment for robotics. Given what's already happening at the hardware layer, that's not an unreasonable call.