When our stars aligned…

Is AI in Drug Discovery the marriage saving child of Tech and Biotech VCs?

You know Tech and Biotech VCs have always had a faulty marriage. One backs concepts that can escalate in popularity almost overnight, the other places bets on extensive pipelines that may not yield returns for decades. This fundamental difference in approach sees Tech investors spreading their resources widely, almost compulsively, across numerous startups, driven by a sense of urgency and a fear of missing out. In contrast, Biotech investors often engage with a deep, almost matrimonial commitment to their choices, nurturing them over many years. One multiplies fiction, the other multiplies hope. One says left, the other…But there’s one thing both of them agree upon, and that is- AI!

Investing in a drug discovery startup for an early stage VC is like being forced to choose at sixteen whether you want to marry your high school sweetheart or not.

The average time to bring a molecule from discovery through to launch is 10-12 years. While the average cost of the R&D process is ~$1Bn+ per drug. Out of a total of 10,000 molecules initially screened, only 5 ever make it to clinical trials. The development of new drugs is a sport that is subject to the principles of extreme first-mover advantage, complex R&D processes, long development periods and substantial funding.

All these warnings always deterred our Tech VCs to play their odds in the drug discovery affair. But there seems to be a new crush in the town suddenly: AI!

It promises to make targets more “druggable”, testing more “efficient” and finally drugs “safer”.

It’s zipping through drug discovery tasks, turning tough targets into "druggable" prom dates, making testing procedures as efficient as a fast-food drive-thru, and ensuring that the drugs are safer than a padded room. With AI on the scene, it’s like suddenly finding a cheat code in the toughest video game of the pharma world. Now, every tech investor wants a piece of that action, hoping AI will swipe right on their venture capital profile.

Okay no more analogies. AI significantly impacts the drug discovery process in three critical ways:

1. Target Identification: AI algorithms can analyze vast datasets, from genomic information to disease pathways, to identify and validate new therapeutic targets much quicker than traditional methods. This capability not only speeds up the initial stages of drug discovery but also enhances the precision of targeting mechanisms, potentially leading to more effective treatments.
2. Molecule Design: Through the use of advanced machine learning models, AI can predict the behavior of molecules and their interactions with biological targets. This allows for the rapid design of more effective and selective drug candidates. AI models can generate multiple iterations in silico, reducing the need for extensive lab testing and thus saving time and resources.
3. Efficiency in Testing: AI streamlines the drug testing process by improving the efficiency of simulations and analyses. This includes predicting how drugs metabolize in the body and their potential side effects, which can accelerate preclinical trials and make the steps toward clinical trials more seamless and less risky.

Let’s be honest, the?number of possible molecules that might make drugs, according to the rules of organic chemistry, is 10^33 and no human is going to read millions of biology papers!

The above chart shows the potential of AI in drug discovery process as we move from Target identification to the clinical stages (it decreases as we move more towards the human trials)

From mining data (millions of pages of biomedical literature), speeding up discovery for hard targets, building AI imagination models (which ‘imagine’ molecules with specific properties for better testing), to in-silico testing (computational clinical trials)- AI is improving the accuracy, predictability and speed of drug discovery.

While the speed of human clinical trials can’t be improved, these AI drug startups stack up odds of a higher ROI in their favour by cutting down the time (avg ~40%) and cost involved (avg ~35%). So for every billion dollar molecule released in the market today- the time would have shrunk to 5-6 years and the costs <$600mn.

Challenger models like AlphaFold, CADD, GENTRL etc are defining the space.

But,

AI still isn’t a magic crystal ball yet—it can't always pick the perfect molecule for a target on the first try, or even most of the time.

“If somebody tells you they can perfectly predict which drug molecule can get through the gut … they probably also have land to sell you on Mars.”

-Adityo Prakash, CEO of Verseon

And may be the market has recently realised this:/

And this is where the entire fuss seems to be:)

• Majority of bets have been taken by Tech VCs at ultra-premium valuations that neither the market nor the business models of these companies could support. Tech VCs think AI had been the only missing link of turning this process into a high RoI business. Churn out ten molecules instead of one in the same time with 40% lesser costs The expectation is to keep generating assets with lifelong revenue streams, but this is something that rarely materialises. You can churn 100s of molecules but the funnel for human trials still remains tight, slow and extremely expensive.
• And business models? They haven’t had any up until now xD. Where biotech investors know how to price the “early stage assets”, pricing the platform vs pricing the successful experiments has remained the usual conundrum. Pricing seems to be the biggest uncommon grounds for Tech vs Biotech investors. And thus, biotech startups that rarely got any attention from the tech world are raising funding at 150% premium valuation from the existing norms. This doesn’t seem much like a handshake.

All in all it seems, the Tech VCs are too optimistic about the potential of AI in drug discovery and Biotech VCs are too skeptical of AI really moving the needle for them.

But one thing is very clear, it’s been quite some time that these investors have kept waving to each other from afar. If advanced tech companies are on one end and unique biology companies on the other end, AI is the bridge that’s bringing these two worlds together. They're closer than ever, but there's still a decent stretch of path to pave before they're truly connected.