Investors are increasingly hearing how artificial intelligence may make drug development more efficient. It's far from science fiction.
Pfizer (PFE) leaned on advanced computational modeling to develop an antiviral drug that could be taken orally, had robust activity against SARS-CoV-2, and had an acceptable safety profile. The drug, now available under the brand name Paxlovid, was discovered in only four months – less than half the time typically required.
The pandemic star is far from the only company deploying artificial intelligence to increase the efficiency of drug development. Although many technology-enabled drug developers are small today, the emerging crop could set new standards for the global industry. Could a novel technology platform being developed by Schrodinger (SDGR) distinguish itself from the field?
Massive Screening Capabilities
Pfizer used souped-up statistical analysis to comb through mind-bogglingly large datasets when developing Paxlovid. Essentially, the industry leader was trying to find the right key (an antiviral drug) for a specific lock (a molecular process allowing viruses to replicate). The only problem is there were billions of possible keys, including dozens that fit the lock. Trying to find one that had acceptable drug-like properties – and the unique characteristic of being orally administered, unlike remdesivir from Gilead Sciences (GILD) – was not a trivial process.
Schrodinger is one of several companies that have built end-to-end platforms for this exact problem. The company's software tools allow customers to simulate physical interactions between chemical compounds (potential keys) and molecular targets (the lock), enabling more efficient screening of chemical space (billions of potential keys).
The $1.8 billion company counted each of the top 20 pharmaceutical companies by revenue as software customers in 2021, with the group accounting for 37% of software revenue. The technology-enabled drug developer also collaborates with certain drug developers directly, which creates the opportunity to earn milestone and royalty payments down the road. Additionally, the physics-based modeling software is used for materials science applications, albeit with more limited revenue generation to date.
Potential Speed Bumps Ahead?
In addition to licensing software, Schrodinger has begun developing a wholly-owned pipeline of drug candidates. While this creates the opportunity for additional upside, investors must acknowledge it also creates new headaches. Existing customers may feel uncomfortable licensing the company's software due to competitive reasons, especially considering Google Cloud now handles the computational load. Alphabet's (GOOGL) health care aspirations and data privacy concerns may make customers wary.
Even if competitive threats are cast aside, developing drugs is associated with significant costs and risks that may reduce the attractiveness of the industry-leading software platform. For example, Schrodinger outlined plans to quadruple drug discovery revenue from $25 million in 2021 to $100 million in 2023. That doesn't include potential upside from licensing wholly-owned programs to external companies. However, the drug discovery business has never been profitable. Schrondinger spent $1.85 to generate every $1 in drug discovery revenue in 2021.
Investors need to see larger-scale operations translate into improved margins in 2023 and beyond. If the company delivers, then it could earn a durably higher market valuation. The risks of being wrong – especially against a backdrop of tightening financial conditions – could lead to more pain for shareholders.
Don't Ignore Schrodinger's Fundamentals
Investors cannot forget that drug developers have fundamentals, too. Precommercial companies may not have revenue or earnings to analyze, but can be evaluated against metrics such as the number of pipeline programs, the length of the cash runway, and the number and maturity of partnerships.
Schrodinger held $525 million in cash, had 14 active programs, and listed three drug development collaborations at the end of March 2022. It reported an operating loss of $28 million in Q1.
The ability to earn high-margin software revenue sets the company apart from the competitive landscape. Schrondinger generated $113 million in software revenue at a gross margin of 76% in 2021. Unfortunately, it lost money on drug discovery revenue and grew total operating expenses 42% last year, resulting in an annual operating loss of $111 million. Considering the business has guided for full-year 2022 software revenue of only $131 million and increasing investments in drug development efforts, fundamentals are bound to deteriorate.
How should investors view the technology-enabled drug developer? It all comes down to execution. There's a real risk Schrodinger fails to deliver on its ambitious goals, or that doubling down on drug development turns out to be a bad idea. That would not pan out well for shareholders while analysts are increasingly worried about tightening financial conditions.
Shares are currently valued at about 14x software revenue, which is a little pricey given the relatively slow growth rate. While the drug discovery part of the business should be factored into the valuation, investors might want to acknowledge the company could struggle to earn a durably higher valuation without significant de-risking -- or exiting hands-on drug development altogether.
