Ginkgo Earnings: Why Biotech Platforms are Hard $DNA
Thanks to all who are reading this week - criticism is welcome.
Ginkgo Biowork’s most recent earnings call was a sobering moment for the company, and illustrated the challenge of building a horizontal biotech platform. Part of its woes stem from its Biosecurity business segment, which thrived during Covid-19 ( Covid testing was a substantial portion of revenue) and has taken a serious hit since then. If management’s forecasts are correct, overall revenue will be down 60% from FY22 highs, with Biosecurity revenue down 85%. There were a few other noteworthy, largely positive, highlights: Gingko expanded its agreement with Novo Nordisk, will cut costs by 100mm this fiscal year, and will replace its own lab automation system with Zymergen’s, a company it acquired in late ‘22.
More important, however, was Ginkgo’s announcement that they would be changing how customer contracts are structured. Historically, pricing for Cell Engineering customers had two components:
- Service Fees, paid to Ginkgo for designing synthetic DNA sequences and then testing those DNA sequences within cells.
- Downstream Value Fees – as mentioned in my initial note on Ginkgo, its Cell Engineering services are meant for initial R&D. Should a synthetic biology product the company helped design succeed, Ginkgo doesn’t then manufacture that product as it scales. In other words, it’s like AWS if startups left AWS once they found product-market fit. To solve for this problem, Ginkgo typically negotiates with its customers for equity and/or milestone/royalty payments.
Going forward, management has decided to nix most downstream value agreements and give its customers ownership of IP generated over the course of a partnership with Ginkgo. In my mind the latter is bigger deal than the former, but it’s worth understanding the rationale behind both:
The biopharma industry is used to downstream value share as a form of payment for services. Given how lucrative patent protected drugs can be, it makes sense to share the upside with a partner if that partner increases the chances of discovering an effective therapeutic. But biopharma isn’t the only industry Ginkgo serves. If you watch the company’s YC demo, the pitch is geared towards companies trying to replace natural ingredients with synthetic ones. It takes three years before a vanilla orchid will produce vanilla beans; altering a yeast cell to produce vanilla can significantly cut down on both cost and complexity for those companies that use vanilla in ice cream or perfume. Unfortunately, a vanilla supplier for Haagen-Dazs has neither patent protection nor brand power, so operates a much lower margin business, and is understandably more reticent to agree to something like royalty fees. These market dynamics have meant that Ginkgo gets into protracted sales negotiations with Industrial Biotech customers, a non-ideal position to be in when you’re supposed to be a high-growth company but revenue is down sharply YoY. It’s difficult to tell how significant getting rid of most downstream value agreements will be. On the one hand, management has said that biopharma contracts will still ‘frequently’ involve this component, which will ensure the company benefits from the innovation, and very high margins, it enables in that vertical. On the other hand, Ginkgo serves industrial biotech customers because it believes there will be real innovation in the space. Charging only service fees means the company benefits from that innovation in a much more limited way, and becomes permanently, rather than temporarily, indexed to the vicissitudes of the industrial biotech venture market and the propensity of established players to spend on experimental R&D.
More notably, Ginkgo’s decided to give customers IP reuse rights, a decision that will affect the competitive advantage of its entire business. A key part of Ginkgo’s pitch to investors and customers is the strength of its Codebase, which consists of genetic sequence and experimental data it’s collected from its work with customers over the years. The more data Ginkgo collects, the better its chances are at designing functional synthetic biology products for its customers, the more sense it makes for companies to partner with Ginkgo rather than reinventing the wheel, with much less data, in house. Understandably, while customers do indeed benefit from Ginkgo having all this data, they’re not all that thrilled at the thought of their own data being used to help potential competitors. In biopharma IP is everything, and another company trying to develop a similar therapeutic to what I’m developing poses a significant threat. The same can be said for other segments Ginkgo serves: if I’m an industrial biotech customer, neither myself nor my investors are likely all that pleased that competitors using Ginkgo are benefiting from the data contributed over the course of our partnership. While this customer reaction makes sense, Ginkgo ceding IP rights threatens an essential part of its value proposition. If it can’t use previous customers’ data, it’s harder to argue that the company’s in all that much of a better position to design and test synthetic DNA sequences. That’s not to say there’s no advantage at all to using the company. Using Ginkgo is significantly cheaper than building a lab internally, and a company that’s done synthetic biology for years likely has better design instincts than a startup or pharma division with less experience. But there’s also a reason that management continually harps on its data advantage throughout annual reports, and reason to think lack of IP ownership will significantly hamper Ginkgo’s prospects over time:
Ginkgo’s recent lab-data-as-a-service offering is also a mixed bag. Historically, the company’s scientists work with its cell-engineering customers to design sequences that will be inserted into host cells and tested in its Foundry. Some companies are reticent to collaborate with Gingko’s scientists on this, and would prefer to instead design their own sequences, but then use the company’s Foundry to generate much more lab test data than they could plausibly generate in-house. Management believes this service will unlock additional portions of customer R&D budgets, and it also means the company’s own scientist headcount won’t need to scale up to the same degree. These are both real benefits, but the downside is that lab-data-as-a-service is its primary offering for AI companies:
“We think we can really be the picks and shovels to all the folks that are inventing amazing new AI models in biotechnology. What we are hearing again and again, there's many new startups getting funded, large big funding rounds. Most of the large biopharmas have now a person in charge of AI strategy. And what we hear from these people again and again is that data is the missing piece for building new and better models in biology……we had huge large English language data sets and things like that to train AI models for English language or videos or images. In biology, the missing piece is actually the data. And so our lab data as a service is exactly the right offering for the -- we could generate large multimodal data sets. And we expect to do business here with customers wanting to access both our automation scale and our expertise, like I said earlier, in conducting large pooled assays. That type of assay generation is particularly important and both of those are available right now on a fee-for-service basis. You own the IP, there's no royalties or milestones for any AI company that's tuning in. We'd love to do that work for you, and you can get that data much faster than anyone else.” (Jason Kelly, Q1’24 Earnings Call)
I said above that it’s unclear whether Ginkgo getting rid of downstream value fees for many of its traditional Cell Engineering customers will end up meaningfully affecting its long-term value. That is not the case when it comes to removing these fees for AI customers. In the short term, no royalty or milestone fees is the right play. It’s unclear how long the AI bull market will last, and the biotech startups leveraging AI are a bright spot in an otherwise challenging funding environment. Negotiating downstream value agreements would lengthen the sales process and make it harder to benefit from this hype cycle. Without downstream value agreements, however, Ginkgo doesn’t own call options on its AI biopharma customers enjoying massive success down the road. Assuming at least some of the therapeutics developed over the next decade are designed with AI, this is a substantial loss for the company.
This feeds into a broader point about how incredibly hard it is to remain a horizontal biotech services company.1 Without downstream value agreements, Gingko becomes a pure services provider, rather than a company that can benefit from the innovation it enables for its customers. To build an advantage in designing and testing synthetic DNA sequences, Ginkgo needs to be able to learn from its previous work. But previous customers recognize that their design and experimental data could be used to help competitors, and so don’t want Ginkgo learning from that data. Without that data, it’s harder for Ginkgo to claim that it has a true advantage when it comes to DNA sequence design. That doesn’t mean the company won’t do well if synthetic biology takes off, but it does means that it will do well in a much more limited way. At this point in time, the most logical step is perhaps one that many biotech services platforms before Ginkgo have taken: to develop therapeutics in-house. While the company won’t hold onto a lot of future customers’ IP, it does have incredibly valuable IP from previous years; Ginkgo scaling up its own AI talent would give likely give it a real chance to discover an exciting drug prospect. That doesn’t come without accompanying costs, though. Management’s repeatedly emphasized on earnings call that its customers don’t need to worry about Ginkgo using their data to develop competing innovations. Going against this might increase the chances of blockbuster success, but it also comes with the price of losing the trust of its customers, should it make sense to keep them.
Disclaimer: The information in this post is not intended to be and does not constitute investment or financial advice. You should not make any decision based on the information presented without conducting independent due diligence.