Cutting through the hype of AI drug discovery

In the world of AI and biotechnology, everyone seems to be shouting, “drug discovery is the next big thing!” 

Do yourself a favor, and don’t get swept away by the hype train just yet. Teasing out those who are tackling critical challenges with plausible paths to market from ones that may be suited for therapeutic-focused funds (and from others that simply do not make sense) is what we do best at Creative Ventures, so let’s get into it. 

Unraveling the hype of AI drug discovery

Will AI drug discovery be the cure-all for the pharmaceutical industry’s woes? Some people sure think so. After all, faster, cheaper, better drugs sound like a dream come true. 

If only it were that simple. 

While generative AI does show promise in speeding up candidate identification, it’s no magic wand. In fact, an overwhelming proportion (~80%) of large pharmaceutical companies’ drugs were developed externally and later acquired. This is because beyond candidate identification, in vitro biochemical assays, animal safety tests, and a series of clinical trials are required to move the drugs through the development and approval cycle. It’s neither a simple nor inexpensive task. 

In a way, large pharmaceutical companies are taking full advantage of their large balance sheets and have the luxury of waiting it out; acquiring assets that are or close to FDA approval. They are also in the best position to manufacture, market, and conduct follow-up outcome studies if needed.

Only the most existing companies can compete

Given how most drugs are discovered and developed today, it’s more or less implied that AI drug discovery companies need to also be developing their own assets beyond the “discovery” and bringing them through clinical trials. A handful of companies may be able to successfully close a collaborative development partnership, though they make up only ~5% of pharmaceutical companies’ assets.

This is why we generally passed on traditional AI drug discovery opportunities. Their go-to-market strategy has a similar risk profile as that of therapeutic companies; something we believe makes the most sense for therapeutic-focused funds.

So, what type of drug discovery companies might have the winning tickets for collaborative development deals with pharmaceutical companies? My bet will be on mechanism-based discovery—the like of intelligence derived from the understanding of how potential drugs and their targets interact. One of the most relevant well-known platforms is Alphafold—an AI system for protein structure prediction, though the current status still seems too static to derive accurate prediction.

Finding opportunities in drug discovery and beyond

The excitement of identifying potential drug candidates can overshadow the challenges that come after. Some AI-generated molecules may pose difficulties in synthesis and purification, impacting their viability for further development. 

As beautiful as the candidates may work in silico, they still need to be physically made and tested to demonstrate their efficacies. We have come a long way in chemical synthesis for chemical API-based drugs, but biologics APIs (e.g. antibodies) are still quite troublesome to make. Traditionally, the process involves DNA synthesis, gene expression in relevant host cells, and protein purification–many factors could go wrong along the way. 

Given how biologic drugs accounted for 93% of U.S. net drug spending growth between 2014-2017, technologies that enable cost-effective, reliable, and efficient production of biologic drug candidates are the necessity and overlooked opportunities in the AI drug discovery realm.

Look past the latest trending topics

Buzzwords come and go. It’s essential to remain cautious of the hype and focus on solutions with solid scientific foundations and actual results that speak to their promises. More importantly, we believe that the most promising investment opportunities need to solve actual problems and provide economic incentives for customers to buy. 

AI for drug discovery hype might unwind in a few months or a few years, but providing cost-effective and high-quality care for human health is a long-term macro trend that’s here to stay. 

Leave a Reply