fbpx

Schrodinger Stock: Digging into The Business Model

Checking in with your stocks about once a year seems like the minimum cadence for a long-term tech investor. In reality, you’ll be checking in more often when dramatic stock price movements happen, or corporate events take place. Today, we’re going to check in on a company we haven’t written about in several years – Schrodinger Inc (SDGR).

The last time we wrote about the firm was in a piece titled A Computational Chemistry IPO From Schrodinger which looked at the S-1 filing document which accompanied their initial public offering (IPO) which took place in February 2020. Simply put, Schrodinger has developed a physics-based computational platform that uses artificial intelligence to help drug developers better predict which novel molecules will successfully pass the FDA drug approval gauntlet. The business model derives revenues from two main sources:

  • Licensing software to customers – largely on-premise
  • Capturing some downstream upside through royalties and milestone payments – large rewards if drugs are successful

Let’s start by looking at how overall revenue growth has been progressing.

Schrodinger’s Revenue Growth

In looking at Schrodinger’s latest revenue projections, steady growth across the business appears promising. The below information was current as of Q3-2021 earnings with 15% projected year-over-year growth for 2021 based on the lower end of their guidance.

2021 Schrodinger Financial Outlook
Credit: Schrodinger

Let’s see if we can better understand their business model by first examining the software business.

Schrodinger’s Software Business

Here’s what we said in our earlier piece:

In looking at the S-1 filing for Schrodinger, they largely install their software on location for their clients – on-premise – which means the revenue is recorded differently. In other words, it’s not a traditional Software-as-aService (SaaS) business model that’s easy to understand. Still, the principles of selling software remain the same regardless of how the revenue is being recognized.

Credit: Nanalyze

The SaaS-related metrics provided by Schrodinger are found in their 10-K annual filing, but they’re named differently. Annual recurring revenue (ARR) is called annual contract value (ACV). That’s because Schrodinger largely sells licenses with one-year terms where cancelation terms are rather loose.

Our customers have no obligation to renew their product licenses or subscriptions for our software solutions after the license term expires, which is typically after one year, and many of our contracts may be terminated or reduced in scope either immediately or upon notice.

Credit: Schrodinger

Because there is no recurring element to the transaction, Schrodinger reports these revenues as “point in time.”

Revenues chart from Schrodinger
Credit: Schrodinger

The software point in time revenue nearly matches the percentage of revenues Schrodinger receives from selling on-premise software.

Revenues chart from Schrodinger
Credit: Schrodinger

The next three categories – hosted software, software maintenance, and professional services – represent the “over time” revenues. Using the lower range of Schrodinger’s guidance, software revenues will barely clear double-digit growth this year with a 10% year-over-year increase. As for the “professional services” segment having declining growth, that seems a bit concerning. The company defines professional services as “training, technical support, installation, or assisting customers with modeling.” Those are the sorts of activities you engage in when onboarding new customers, so we’ll look forward to seeing the below chart updated for 2021:

Bar chart of the number of Schrodinger's customers with an annual contract value of at least $1,000 during the year
Credit: Schrodinger

Analysts seemed to pick up on the slowing software growth in the last earnings call when the CEO assured investors that “customers require significantly more licenses than what they have access to now” with the ideal penetration representing “somewhere in the range per company of around $30 million to $40 million.” He also talked about how the company expects to maintain its “long historic extremely high retention rate” by achieving “very, very high 90s” going forward.

Expectations and reality are often different, so gross retention rate (number of clients who don’t cancel) will be a key metric to watch going forward along with net retention rate (existing clients who spend more). The latter metric can be deduced from charts in the 10-K which show the number of clients who contribute $100,000 or more (153 in 2020) and the number of large clients which should always be growing.

Bar chart showing Schrodinger's customers with annual contract value >$1 million
Credit: Schrodinger

Despite the slow growth, the software business seems healthy if they can keep growing all those numbers and maintain the low cancelation rate. Where the big upside seems to be is in the drug discovery segment.

Schrodinger’s Drug Discovery Revenues

Capturing additional value downstream is the trendy new business model being put to use by firms such as Ginkgo Bioworks, Twist Bioscience, and AbCellera. We find it desirable under two conditions:

  • Downstream revenue streams aren’t “lumpy” as that creates volatility
  • Present-day valuations aren’t richly priced in anticipation of future rewards

Regarding the first bullet point, Schrodinger mentioned this in their most recent earnings call stating, “revenue in this segment can be highly variable period-to-period as the decisions on program timelines are largely in the hands of our collaborators.” In fact, they lowered the upper range of the drug discovery revenue guidance from $32 million to $24 million for that very reason. However, for the past four years, drug discovery revenues have been consistently growing.

Bar chart showing Schrodinger's annual drug discovery revenues - Credit: Schrodinger
Schrodinger’s annual drug discovery revenues – Credit: Schrodinger

So, the question remains. Is the current valuation of Schrodinger giving too much weighting to future drug discovery revenues? We can answer that question by using our simple valuation ratio.

  • Market capitalization / annualized revenues
    1,870 / 119.6 = 16

Contrast that to Ginkgo Bioworks which currently has a ratio of 29 (that’s with pandemic and related revenues included). Sure, it’s apples to oranges, but the basic idea holds true. Even with drug discovery revenues removed, Schrodinger wouldn’t be excessively overvalued.

Schrodinger and Bristol Myers Squibb Company

The largest drug discovery partnership Schrodinger has is a multi-target collaboration with Bristol Myers Squibb (BMS). Last quarter, nearly 79% of drug discovery revenues were attributed to $4.4 million recognized from their collaboration with BMS.

Under the agreement, Schrodinger will spend – at their own cost – up to four years developing novel molecules for BMS to commercialize. In exchange, they received “an initial upfront fee payment of $55 million” and a promise of future riches – “up to $2.7 billion in total milestones across all potential targets.” Of course, that’s assuming BMS can successfully commercialize multiple molecules that Schrodinger develops, and that needs to happen to demonstrate the platform actually works as advertised. The icing on the cake? A tiered percentage royalty on annual global net sales for Schrodinger if BMS hits pay dirt with a successful blockbuster.

Schrodinger’s Recent Acquisition

The reason Schrodinger came across our radar again was news of their recent acquisition. On January 14, 2022, they acquired XTAL BioStructures for $6 million in cash in order to offer customers “access to protein structures that have been computationally validated and are ready for structure-based virtual screening and lead optimization.” Our recent article on Gandeeva Therapeutics and Cryogenic Electron Microscopy was well-timed because in that piece we learned about how structural biology works.

One tool used to examine proteins in great detail is x-ray crystallography, something XTAL specializes in. Schrodinger will use this tool to “scale up production of high-resolution structures” which serve as starting points for their computational platform. Providing more detailed input will certainly result in more accurate outputs.

Buying Schrodinger Stock

We’re presently long Schrodinger with a position that’s -40% below our cost basis with capital still remaining to allocate. In other words, there’s no reason we shouldn’t be adding shares right now, except that we’d rather wait until the year-end results come out.

Given the high expectations Wall Street analysts set for companies these days, sometimes a generally positive year-end will coincide with shares taking a big fat dump. Our plan is to be prepared in advance so that we might add shares if an opportunity arises early next month (year-end results are expected to be announced around March 2nd).

Conclusion

Schrodinger’s basic metrics appeared healthy up until last year, so next month’s updated numbers should be given a quick checkup. The company doesn’t seem to be overly valuing the promise of future revenues coming from the drug discovery business, and that’s great because the software business isn’t growing very fast. If no red flags are raised during the year-end results, we may add some more SDGR stock to our portfolio. If we do, Nanalyze Premium subscribers will be sent an alert.

Share

Leave a Reply

Your email address will not be published.

  1. I was reading the Shroedinger article, which says, “We’re presently long Shroedinger with a position that’s -40% below cost.” So, do you mean “40% above cost?”
    “-40% below cost” is a double negative. So, I don’t know whether to read this literally, as in “40% above cost,” or just assume you all are really bad writers. Either way, it’s hard to read (cringe) and hard to interpret (WTF?)

    1. Thank you for that scathing critique. It’s how we write negatives here – our own unique style. We’re aware it’s confusing to some people, and we’ve always been okay with that. Millions have read our articles and nobody ever raised this before.

      You raise a good point though. We’ll run this by our editorial review board and see if it’s something we want to change going forward or not. Thank you Heather!

  2. Schrodinger is one of those companies that I simply can not understand what is doing, but which I really would like to understand.
    I’ve looked at dozens of companies that help drug developers in one way or another using artificial intelligence, but I simply can not understand what it is you can do with Schrodinger’s services / products?
    To me, medicine and its effect is a matter of chemistry and biology or biochemistry. I simply can not understand what what it means to be a “physics-based computational platform?

    1. Good comment Thomas. The company is trying to move away from using AI to label what they do which makes the whole thing even more confusing. Check out this piece we did, particularly, the diagram titled “Schrodinger’s FEP+ platform.”

      https://nanalyze.com/2020/01/computational-chemistry-ipo-schrodinger/

      We’ll try and paraphrase using our limited understanding.

      Let’s say you have 1 billion molecules that could represent a potential therapeutic. Take a random sample of 1,000 and then feed them to (don’t say machine learning algorithms) the “physics-based platform” which then analyzes them for whatever properties you’re trying to optimize for. Use the output to then rank the 1 billion molecules. (We would imagine that this is where iterations happen.) Once you have a ranking that isn’t changing much based on re-running iterations, take the top-ten molecules and advance them. Kind of makes sense.

      This is where the proof needs to be in the pudding. If companies pay to access the platform – and MOST importantly – keep paying to access it and even increase spend, then that means it’s adding value. That appears to be the case. Of course, the biggest proof of traction will be having success with the BMS relationship.

  3. In drug discovery, the first calculations you do attempt to determine how tightly each drug candidate binds to the target protein you’re tying to deactivate. That’s physics, not chemistry. Schrodinger offers a suite of software that’s able to calculate the binding energy at different levels of approximation. You start with quick methods that you can apply to thousands or millions of potential drugs, and as you eliminate candidates, you start using the more expensive, but more accurate methods on the remaining candidates. The quick calcluations may not allow for the fact that proteins’ own structure often changes when binding a drug, and that the interaction of water with both the protein and the drug often affects the binding energy significantly. The FEP+ methods produce very accurate binding energies for your top candidates by including the physics of those interactions. In the later stages of this process, chemists can use these calculations to guide them in fine-tuning the candidate.

    None of the above is AI – it’s just using known physics intelligently to guide drug discovery. AI methods can complement this approach. They might also be better able to address questions that go beyond binding energies, like how quickly the drug candidates will break down in the body, and how the interact with other drugs.

    I hope this makes it a little more clear how drug companies use Schrodinger’s software.

    1. We love it when a subject matter expert chimes in and sorts everything out. Well explained, thank you so much for that!

  4. I’ve seen interesting point: Schrödinger with its valuable technology and projects is likely to become a prime acquisition target in the near or medium term. It’s market cap is only $2.2B, so a large biotech company could easily buy Schrödinger offering a good premium: eg 150%-300%.

    1. We’ve always found that speculating on acquisitions never comes to fruition as one expects. We wouldn’t have thought this would be a target, but we typically never try and think about these things because it’s speculating. 🙂