Business

Know the Business — Ginkgo Bioworks (DNA)

Bottom line. Ginkgo is a cash-rich, revenue-shrinking biological R&D-services platform priced close to its net cash. The historical "platform with royalty upside" thesis has been retired by management; the new thesis is a tools-and-automation pivot (RACs, Datapoints, Cloud Lab) financed out of an $422M cash pile burning at $170–180M per year. The decisive question is not revenue growth but whether the tools pivot finds product-market fit before the cash optionality runs out — and whether $606M of future lease commitments end up being a cash liability the market is currently ignoring.

1. How This Business Actually Works

Ginkgo sells biological R&D capacity to customers who want to engineer cells, proteins, enzymes or strains but do not want to build (or expand) the lab themselves. Two things are now happening at once: the legacy business (multi-year "Solutions" programs with downstream royalties) is shrinking by design, and a new tools business (selling lab automation hardware, AI-ready datasets, and a "cloud lab" service) is being scaled in its place. The pivot — formalized in 2024 with the explicit removal of downstream value share from "certain program types" — is from a CRO-with-call-options model to a picks-and-shovels model.

The economic engine in one sentence. Ginkgo wants to convert experimentation into a metered service the way AWS did with compute — the unit of sale moves from a multi-year scientist-team to a per-experiment data package or a rental for an automated lab. The platform "improves with scale" only if customers actually buy the metered units; in the old Solutions model, scale was Ginkgo's own scientist headcount, and that has now shrunk to 485 (from ~1,400 at peak).

Where margin actually comes from. Gross margin looks healthy at 72% for FY25 because reported "cost of revenue" only captures Biosecurity COGS plus a thin Datapoints/automation COGS — the much bigger labor and facility costs sit in R&D and G&A. The true contribution structure is:

The bottleneck of the business is no longer scientific capability — Ginkgo has the technology — it is customer adoption of a new buying motion. Pharma scientists are accustomed to either keeping work in-house or contracting it out as a project. Buying "an autonomous lab as a service" or "a Datapoints package" is a different procurement, a different budget line, and often a different signoff. That re-education is what the Boston frontier lab and the OpenAI/PNNL announcements are designed to accelerate.

Bargaining power lives with the customer. Three of Ginkgo's most public partnerships (Motif, BiomEdit, related-party termination) ended in non-cash deferred-revenue releases totaling roughly $57M across FY24 and FY25 — i.e., the customer walked, Ginkgo kept the cash already paid, but no new economics. The 2024 decision to remove downstream value share from many program types was effectively an admission that customers would not pay the platform's stated royalty price.

2. The Playing Field

The right peer set is five publicly-traded TechBio / synbio platforms named in Ginkgo's own 10-K Competition section — none of them is profitable, and the market is paying for very different things in each name.

What this peer set reveals. The market is paying very different prices for very different things. Twist trades at 9× revenue on a real tools-company trajectory (20% YoY growth, shrinking FCF burn, mid-50s gross margins). Recursion and AbCellera trade at 15–16× revenue on drug-pipeline option value, not revenue — gross margins there are accounting illusions (ABCL reports 100% GM because scientist costs sit below the gross-profit line; RXRX has 5% GM because pass-through pharma collaboration costs sit in COGS). Ginkgo at 3.6× EV/revenue sits with the cheapest names (Codexis at 3.8×) — the market is not paying for the option book or the tools pivot. That is the central valuation tension.

What "good" looks like here is Schrödinger — FCF-positive in FY25 ($12M) on $256M of revenue at 56% GM, validating that a TechBio platform can reach breakeven without giving up pipeline option value. Ginkgo's $125–150M FY26 burn target (versus $179M FY25) implies a similar trajectory in roughly two years if revenue stops decaying. Twist shows the alternative path — accept the tools-company identity and grind margins higher quarter by quarter.

3. Is This Business Cyclical?

Yes, on three independent cycles — and the company is currently on the wrong side of all three at once, which is why the income statement looks the way it does.

The two cycles that hurt most were synchronized: a step-down in biosecurity (post-pandemic normalization) collided with a multi-year biotech-funding drought that thinned out small-customer demand for Solutions. Cell engineering revenue per active program has compressed materially — from roughly $1.0M per active program at the FY22 peak to roughly $0.75M in FY25 (133/176 ≈ 0.76, using the FY24 active-program count since management stopped publishing the metric in Q4 2024). That is the cycle's fingerprint at the unit level.

What is not cyclical: the lease overhang ($606M of post-2026 commitments) and the convertible-note-style obligations to Google Cloud and Twist. Those are structural and persist regardless of where the biotech cycle goes.

4. The Metrics That Actually Matter

GAAP earnings tell you almost nothing about Ginkgo right now — net loss is dominated by stock-based compensation, lease impairments, mark-to-market on warrants and equity investments, and one-time deferred-revenue releases. The metrics that determine value sit elsewhere.

Burn was cut 55% in FY25, but the cash balance has fallen 73% from its 2021 IPO peak to $423M, with $171M of operating outflow in FY25 alone. Management guides $125–150M for FY26, which puts year-end 2026 cash near $280–300M before any ATM issuance, biosecurity divestiture proceeds, or new financing — and before honoring the $47M PNNL surety bond restricted until ~2029. The optionality window for the pivot is roughly 2026–2028; not infinite, not imminently in crisis.

5. What Is This Business Worth?

The right valuation lens for Ginkgo today is liquidation value plus optionality, not multiples of revenue or EBITDA. Revenue is shrinking; EBITDA is negative; the company has $1.8B of federal NOLs that will never be used by Ginkgo alone. What you are buying is a balance sheet plus a portfolio of real options whose values depend on independent outcomes.

What would justify a premium. Three things, in this order: (1) hard evidence that RAC deployments are happening outside Boston at scale — preferably one named top-10 pharma anchor; (2) Datapoints revenue disclosed separately and growing >50% YoY with positive contribution margin; (3) FY26 cash burn coming in at the low end of guidance ($125M or below) with no new ATM issuance beyond modest amounts. Each of these tests the autonomous-lab thesis without requiring a return to FY22-level revenue.

What would justify a deeper discount. (1) Failure to sublease the new Drydock space and persistent $40M+ excess-lease drag; (2) RAC sales pipeline failing to convert beyond DOE/national-lab one-offs; (3) Biosecurity divestiture closing for trivial proceeds or being delayed; (4) restart of major SBC issuance to retain talent during the pivot; (5) cash burn re-accelerating in 2027 as restructuring cuts are lapped and growth investments are required.

The reason this is not a sum-of-the-parts exercise: there are no listed subsidiaries, no separately-valued asset stakes large enough to swing the thesis, and the two reporting segments are converging on one (post-Biosecurity-divestiture). The right mental model is a single-engine business with a balance sheet floor and three independent options layered on top.

6. What I'd Tell a Young Analyst

Stop modeling revenue. Model cash and adoption. Ginkgo's revenue line is contaminated by deferred-revenue releases from terminated contracts, non-cash consideration, and one-off government programs. The numbers that determine whether this stock is worth $9 or $25 or $3 in eighteen months are quarterly cash burn, RAC deployments at third-party sites, and whether Datapoints turns into a disclosed recurring revenue line. If the next earnings deck does not break out tools revenue separately, that is itself information.

Take the company's own KPI suspension seriously. Management stopped reporting New Programs, Active Programs, and Cumulative Programs in Q4 FY24 — the same window in which those metrics were inflecting down. This was disclosed transparently, but the analytical work is to insist that the replacement disclosures (RACs deployed, Datapoints customers, Cloud Lab pilots) be at least as informative. Until they are, assume the worst about the trend in whatever is not being reported.

Watch the lease, not the lab. The $606M of post-2026 lease commitments and $54M annual excess-lease drag is the single biggest hidden liability and the one most likely to surprise. Boston biotech lab vacancies are at multi-year highs; subleasing the new Drydock build-out at favorable economics is not assured. Every quarterly call should be a chance to update the sublease pipeline.

The thesis is binary, the price is asymmetric. Trading at roughly 1× cash + securities, downside is bounded by the balance sheet (assuming the lease overhang doesn't reprice it), while upside requires the autonomous-lab pivot to find one validating anchor customer. The market that called Ginkgo a 2021 SPAC bust may be missing that the company is quietly turning into Twist circa 2019 — a sub-scale tools business on a path. Or it may be right, in which case the cash is the floor. The work each quarter is to update which of those two stories the new disclosures support.

The thing the market is most likely underestimating: that the OpenAI/PNNL announcements are real validation of a category (autonomous labs) that did not exist as a category eighteen months ago, in a year when the US government is actively funding it ($100M NSF AI cloud-lab initiative; BIOSECURE Act tailwinds). The thing the market is most likely overestimating: management's ability to operate that pivot inside a cost structure built for a $500M-revenue services business, with $606M of lease obligations and an ATM that has only printed $18M in twelve months.