A Review of Seven AI Drug Discovery Stocks

In the dot-com days, everyone launched a website and became a “dot com company.” Today’s equivalent is peppering your investor deck with mentions of generative AI and large language models (LLMs). For retail investors, this means contending with financial pundits who now spin every company with a chatbot as “AI-enabled.” We’re fortunate enough to have been following AI stocks before there were AI stocks, so we can more easily distinguish between hype or substance. One area that’s been a focus for AI investors for a while now is AI drug discovery.

The thesis speaks for itself. If AI algorithms can best the world’s most elite Go player, then turn around and master protein folding shortly afterwards, it’s realistic to think we might use these same algos to make the broken drug discovery process more efficient. So, a handful of companies started building AI platforms that digest large amounts of big data and spit out drug candidates that have better safety and efficacy characteristics than what might have been produced using traditional methods. Today, we’re going to take a brief look at each while considering key metrics such as:

• Size – bigger is better as raising capital becomes easier and the likelihood of trading on a major exchange (where there’s liquidity) increases.
• Pipeline progress – Developers often receive revenues upfront from pharma companies when they start collaborating, but the bulk of potential lies in downstream milestones and royalties.
• Runway – cash on hand divided by some estimated burn rate number tells us how soon a firm might have to give away more equity or take on more debt.
• # of top-10 pharma partners – large companies allow firms to then attract others using social proof. It’s a double-edged sword though if a major player abandons a platform.
• Total drug pipeline – breadth of potential winners.
• Drugs in clinical development – progress of potential winners.

It’s important to conceptualize what a success story will look like for a leading AI drug discovery firm. First, it won’t go bust. The smaller a firm becomes, the fewer options exist to raise capital, the more of the company needs to be given away to raise money to stay in business. That’s where size comes into play along with runway. While we don’t invest in companies with a market cap of less than $1 billion, we’ll still discuss the four names that fall below this threshold starting with Lantern Pharma (LTRN).

Keeping the Lantern On

Choosing to bring a product to market with a large pharma partner means you can share the costs and potentially receive milestone payments that help supplement that giant pile of cash you’re lighting on fire. Lantern’s press release a few weeks ago talks about their successful IND Application for drug candidate LP-184 which is “the first of Lantern’s drug candidates to be developed entirely internally.” That’s not a good thing when you’re a $56 million company with $56 million in cash and no large pharma partner to share the costs with. Some might say that with a market cap of $56 million, and the same amount in cash on the books, you’re getting a free drug development company. What you’re actually getting is a funding headache.

Developing a candidate from discovery through clinical trials is an expense that now exceeds $2 billion per therapy on average. Small companies that choose to develop candidates internally without resources are sending a signal to the market that their platform really isn’t that good. And if the strategy is to “keep a bigger piece of the pie,” there will be no pie unless they can raise the money needed to bring their drug to market. This same criticism holds true for all AI drug discovery companies developing products in-house without a partner. Those are pipe dreams and need to be valued as such.

Market research firm Frost & Sullivan produced the below graphic (brushed up by Recursion) for which Lantern is suspiciously absent.

Big Trouble at Little BioXcel

Another firm not seen on the above chart is BioXcel (BTAI) which we covered way back in 2018 in a piece titled BioXcel Therapeutics’ AI Drug Development IPO.

In the case of BioXcel, we see a convoluted offering that is difficult to make sense of which begs the question of why anyone would feel the need to make sense of it in the first place. 

Credit: Nanalyze, Feb 2018

Our piece highlighted a company with a convoluted corporate structure, a value proposition that was difficult to decipher, and a lack of large pharma partners. Less than 24 hours ago, the WSJ published a piece on a fraud investigation which was picked up on by multiple news sources.

When you’re investing in a domain with exceptional amounts of risk, red flags like this means you immediately move on to the next candidate on your list.

BenevolentAI Takes a Break

In December of 2021 Amsterdam-listed Odyssey “bought” BenevolentAI (BAI.AS) in a 1.5 billion euro ($1.64 billion USD) deal that was essentially Europe’s version of a SPAC debut. That deal finally closed several years later, and the latest news doesn’t seem so promising. An article by FierceBiotech tells us that “BenevolentAI is cutting back its drug development operation in the wake of a midphase flop, laying off up to 180 staff, reducing its lab footprint, pausing some programs and dropping its lead candidate.” The messaging that accompanied this news doesn’t seem to bode well for the remainder of their pipeline (some of which is being developed with Astrazeneca, a top-10 pharma company).

In May 2023, BenevolentAI paused all its TargetID deployment programmes with future optionality to re-initiate these on a case-by-case prioritised basis, resource permitting.”  If there is any change to this we can update and wrap up in communications later in the year.

BenevolentAI, May 2023

Your lead candidate is always your best foot forward. When something happens to a drug developer’s lead candidate, it sends a bad signal to the markets, and they’re suddenly pivoting into some turnaround story. We like to keep things very simple. If your lead candidate bombs, or if a major pharma partner bails and suddenly it’s being developed in-house (while management tries to spin this as a positive), these are red flags which mean we immediately move on to the next option.

Exscientia’s Runway

Can Exscientia Stock Win the AI Drug Discovery Race? That was the title of last year’s article which talked about their broad pipeline of 30 projects, a third of which are backed by big pharma names like Bristol Myers Squibb (BMS), Sanofi, and Bayer. Several months ago, they announced their “sixth AI-designed compound reaching the clinical development stage.” All boxes appear to be checked (aside from our market cap rule), which means Exscientia only needs to focus on surviving long enough to see their platform start churning out some success stories. There’s always the option of raising more cash, but this will either dilute existing shareholders or incur debt. Let’s see how our remaining companies compare in terms of runway.

The above doesn’t take into consideration unpredictable revenue streams that may push runways beyond what we’re calculating. When it comes time to raise cash, partners who are seeing progress may be willing to pony up some money to support the efforts, so the terms at which any of these companies raise money will telegraph the extent to which they’re creating client success stories. Having active projects with multiple partners creates a potential pipeline of success stories. That’s something Relay Therapeutics seems to be missing.

Relay Therapeutics’ Breadth

Doesn’t have the same breadth and depth of pipeline and partnerships that other firms like Recursion or AbCellera have. Would prefer to focus on names with more irons in the fire.

Nanalyze, Aug 2022

The above note was pulled from our tech stock catalog and summarizes what we learned in our piece last August titled Relay Therapeutics Stock: Pure Play on AI Drug Discovery. The only notable partner (aside from some undisclosed preclinical programs) is Genentech which controls data disclosures for a compound with a status of “early clinical.” Relay’s lead candidate, RLY-4008, is in Phase 1/2 development with completion slated for Fall of next year. Great results from that study will help the company raise the cash needed to survive given the nearly $1 billion in their coffers is only expected to last “into 2025.” With three active clinical trials and one disclosed big pharma partner, we’re just not seeing the breadth of projects that might lead to consistent revenue streams in the future.

Revisiting Recursion

Some weeks back we published a piece titled Recursion Stock: Do We Still Like What’s on Offer? which questioned whether the company can start realizing success stories to pay the bills before having to raise again. With $473 million in cash at the end of Q1-2023 (and only about $1 million in debt), it reported about $65 million in losses for the most recent quarter which means they’ll be out of runway in less than two years. This doesn’t take into account revenues they might start to realize from an internal drug pipeline that initiated five clinical trials in 2022, including three Phase 2 programs. Both Insilico and Exscientia claim to have “the first AI-designed drug to enter human clinical trials,” but the first developer to have an AI-discovered drug approved will then have the first piece of evidence that – hopefully – points to the platform bringing a drug to market quicker and/or more efficiently. AbCellera’s already done that.

AbCellera’s Pipeline

AbCellera (ABCL) helped Eli Lilly discover bamlanivimab from a single blood sample obtained from a convalescent patient, and with their partners, advanced into clinical testing 90 days after initiation of the program. That drug has now been used to treat over one million COVID patients, but that success story took place under an exceptional set of circumstances. Now, we’d like to see a similar success story that has progressed through the traditional hoops that need to be jumped through before a drug can be brought to market. Our recent video on AbCellera noted that just “28% of clinical pipeline will demonstrate the promise of their AI-powered antibody discovery platform,” while 43% of the pipeline relates to their specialized rodent platform. Like all the other names discussed today, AbCellera needs to start realizing more success stories.

Follow the Leader

Uncertainty equals risk, and it’s why we’d like to see some success stories come out of AI drug discovery before investing in this niche (more on this in a bit). The milestones towards a proper AI drug discovery platform ought to look something like this:

• Initial AI success story is first AI-created approved drug which gets the media hype machine going.
• AI drug discovery stocks rise together as a group demonstrating price action is hype.
• Things eventually settle.
• The next several AI-created drugs get approved with less fanfare.
• By now, one or more pipelines should be showing steady progression with revenues starting to resemble some sort of trend.

Investing at around the third bullet point is ideal, but only in candidates with enough breadth in their pipelines to keep the money machine going.

Honorable Mention: Schrodinger

That brings us to a firm that’s definitely not an “AI stock,” but that’s being pigeonholed as one by eager pundits hoping to find “the next NVIIDA.” We’re nervous talking about Schrodinger (SDGR) in the same breath as artificial intelligence, because the last time we did their legal team chastised us. Schrodinger’s technology platform uses computational modeling, a topic we’ve covered extensively over the years. Our last piece on the company published in March – Schrödinger Stock: Drug Discovery Platform Making Money – talks more about their business model. Since then, shares have risen +108% compared to a Nasdaq return of +24%. That isn’t a good thing, because nothing leads us to believe the intrinsic value jumped that quickly. We’re left concluding that this may be related to certain market hype we’re seeing around “you know what.”

Conclusion

AI drug discovery companies have been flying under the “everything that says AI is a play on generative AI” radar which has ascribing hype status to a number of names out there such as C3, Soundhound, and Palantir (the latter of which boasts one of the largest cheerleading expert communities around). It still needs to be proven that AI can bring drugs to market faster and more efficiently than traditional methods, and the first success story will just be the beginning of a much larger opportunity – an AI-powered platform that churns out drugs for the Americans as fast as they’ll take them.

Tech investing is extremely risky. Minimize your risk with our stock research, investment tools, and portfolios, and find out which tech stocks you should avoid. Become a Nanalyze Premium member and find out today!