AI Investor Forum: Gus Tai on AI in Healthcare (Part 1)

Gus Tai, Investor, Board Member and Retired General Partner at Trinity Ventures, discusses AI in Healthcare startups. Fascinating, comprehensive discussion with concrete pointers.

Sramana Mitra: Today we are going to welcome back Gus Tai, my friend and colleague. For several years, Gus has been a very active veteran investor in the industry. He has been a partner at Trinity Ventures and has also done a tremendous amount of angel investing in very important companies. Gus and I have shared a lot of brainstorming for several years. We are bringing some of that brainstorming to you in light of all that is going on in the world right now, especially AI driving such major changes and opportunities. Gus, welcome back.

Gus Tai: Thank you, Sramana. It’s always lovely to see you.

Sramana Mitra: So, Gus, we discussed education thoroughly last time in the context of AI in education and startup opportunities in particular. I thought we would focus on healthcare today and deal with different aspects of healthcare that are being touched by AI at the moment.

So, the starting point where it would be good to brainstorm is one that you have done some work in – drug discovery. Now, in a recent round roundtable, we had Mo Islam from Threshold Ventures. They are investors in Atomwise, which is a drug discovery platform. They are selling the platform for drug discovery to pharmaceutical companies. I asked him, “Are you selling only to large pharmas, or are smaller startups are to access such a platform?”

As you know, in One Million by One Million, we emphasize hugely on bootstrapping. If you’ve to build the entire AI stack to even start doing drug discovery work, that is not a bootstrapped venture. I’m a little bit concerned that everybody who’s in this ecosystem – the scientists, people who have real knowledge of specific diseases, specific biology, and life sciences conditions that could inform drug discovery work, will not have access to a platform to springboard off of. How do we think about this, Gus?

Gus Tai: There’re tremendous opportunities to apply AI to healthcare and to drug discovery. Whether those opportunities are easily accessible to people who want to bootstrap, I think that’s a question mark. The reason is that the drug discovery market is most vibrant here in the US where there’s an FDA process of approval of drugs, and it requires a lot of capital to test the drugs. The expense to discover the drugs is actually declining, but then the testing process is rigorous and takes a long time and a lot of capital. The way that the pharmaceutical industry is designed is that you have these organizations that are doing the R&D, and once they get further along, you have these aggregators. They do more than aggregation.

I’m being simple, but I just want to talk about the business concept of the aggregator distributor, like a Pfizer. If they’re not developing their own drugs, they will partner or acquire these drugs because they have quite hefty capability and skill at marketing and distributing drugs.

This whole process is designed around finding, discovering, and distributing multi-billion dollar drugs. It’s jumbo in size.

One of my very last formal investments as an institutional general partner was in a company called Cyrus Biotech that was founded by Nobel Prize winning David Baker, who had innovated a lot of techniques on how to apply AI for drug discovery.

That company had to decide whether to be a platform or whether to use the technology for individual drugs. There are different business models. They had a platform business, and the platform business can be much more capital efficient. But there haven’t been as many platforms of that facilitate drug discovery that have become large and jumbo in size for the venture industry.

So perhaps there could be some new opportunities for entrepreneurs, but I’m not sure they would be as large as what a median-sized venture firm would want to see in terms of size.

Sramana Mitra: So, somebody who’s a scientist, let’s say somebody who has a life sciences background to do drug discovery work and wants to use AI to build a company in a specific area of drug discovery, what are we saying? Are we saying that this group of scientists have to then pair up with some computer scientists who can build them a platform?

Is that the path that they have to take?

Gus Tai: So there were traditional paths, and then there are, of course, always going to be novel, innovative paths. Even those traditional paths, which are using AI but not generative AI, tend to be more capital intensive. It would be like the industry of developing fusion energy.

It’s hard to see what would be the opportunities that would be very capital efficient. But to your question, if you are someone who is investigating either discovering new drugs of discovering techniques that would facilitate discovering a class of drugs, they would need to be partnered with people who are skillful with using the AI algorithms appropriate for that area. Then that group of people could look for sort of like NRE funding or development funding and develop modules.

That could be an interesting IT type of business that wouldn’t be very capital intensive initially.

This segment is part 1 in the series : AI Investor Forum: Gus Tai on AI in Healthcare
1 2 3 4 5 6 7 8