Ginkgo Bioworks

NYSE-listed Boston lab-automation company that abandoned its "cell-programming foundry" model and now sells autonomous labs, AI training data (Datapoints) and onshore R&D services.

1. Core Product / Service

Founded in 2008 out of MIT as a synthetic-biology foundry — engineering strains for customers in exchange for fees plus downstream equity or royalties — Ginkgo has largely retired that model; non-cash (equity-for-services) revenue fell to $1.3M in H1 2026 from $9.9M a year earlier [1]. It now sells three things on top of its Boston Nebula facility (100+ robots built from Reconfigurable Automation Carts, RACs) [1][6]:

1. Automation / autonomous labs — RAC hardware and software sold to government, universities and pharma: a $47M, 97-instrument autonomous lab for EMSL at Pacific Northwest National Laboratory; NSF-funded cloud-lab nodes at Caltech, Northwestern and the University of Maryland plus an MIT Media Lab node (Q2 2026) [6]; and, in September 2026, an autonomous lab at Novo Nordisk's new Waltham R&D site [7].
2. Datapoints — standardized, large-scale assay data for training AI models: antibody developability (10+ biophysical assays), Cell Painting, DRUG-seq, iPSC screens, and ADME-One (launched May 2026; $199 per panel, 17 customers in six weeks per the company) [1][4][8]. Terms state the customer owns the data with no royalty [4]. An Antibody Developability Consortium with Apheris (founding members AbbVie, argenx, Lundbeck, Takeda) pools data under federated learning [9].
3. Onshore R&D services — pitched as a US alternative to Chinese CROs amid Biosecure-driven onshoring [1].

Wet lab: fully owned. Everything runs in Ginkgo's own automated Boston labs; it is itself the wet-lab supplier other AI companies outsource to. The flip side is a heavy fixed base — non-current operating lease liabilities of ~$404M versus ~$302M of cash and securities at mid-2026 [1].

Business model: who pays. Government and defense became the largest customer group (45% of Q2 2026 revenue vs 21% a year earlier), while pharma/biotech fell to 20% from 44%; automation contracts are recognized largely on delivery and installation, Datapoints and services per project or data point [1][6].

2. Target Users & Pain Points

• National labs, universities, research funders building autonomous "self-driving" labs.
• AI-bio model builders and pharma that need large, consistent training datasets they own outright.
• Pharma ADME/developability teams seeking low-cost US-based assays.

3. Competitive Landscape

4. Unique Observations

• A shovel that failed, then a new shovel: the foundry was a non-standard service — every customer project different — so robot utilization never scaled, the same failure mode that closed earlier cloud labs (Transcriptic/Strateos) [1][5].
• Selling the lab instead of using it: autonomous-lab contracts are project-based system integration funded by grants; scale is real but margins look closer to integrators than to software [1].
• Data ownership as a wedge: "customer owns the data, no royalty" plus federated consortia is a deliberate contrast to platform companies that keep the data [4][9].

5. Financials / Funding

• Pre-IPO venture: ~$798M (Crunchbase estimate) [10].
• SPAC (Sept 2021) with Soaring Eagle at ~$15B pre-money equity value; $1.633B gross proceeds incl. a $775M PIPE [10].
• Revenue: $478M (2022, incl. COVID testing) → $170M (2025); Q2 2026 $20.2M (−48% YoY), net loss $57.3M [1][6].
• Cash burn guidance 2026: $125–150M; cash + securities ~$302M (June 30, 2026) [1][6].
• 1:40 reverse split (Aug 2024); biosecurity unit spun into Tower Biosecurity for ~20% stake (Apr 2026) [1].

6. People & Relationships

• Co-founders: Jason Kelly (CEO), Reshma Shetty, Barry Canton, Austin Che, with MIT's Tom Knight [1].
• Customers/partners: PNNL/EMSL, NSF cloud-lab network, Novo Nordisk, AbDev consortium members, Inductive Bio and Tangible Scientific (lab-in-the-loop) [2][3][7][9].
• Peers: recursion-pharmaceuticals, schrodinger, insitro.

Sources

• [1] Local equity research note on Ginkgo, SEC 10-K/10-Q based (2026-10-01)
• [2] Local AI-bio platform comparison (2026-10-01)
• [3] Local company profiles: Ginkgo / Axolabs / Aurigene (2026-09-29)
• [4] Local comparison: Adaptyv / Twist / Ginkgo Datapoints (2026-10-04)
• [5] Local AI-for-Bio "shovels vs miners" summary (2026-10-03)
• [6] Ginkgo IR, Q2 2026 results (2026-10-14)
• [7] Ginkgo, "Ginkgo Bioworks to Build an Autonomous Lab at Novo's New U.S. R&D Site in Waltham" (2026-10-14)
• [8] Ginkgo Datapoints website (2026-10-14)
• [9] BioSpace, Antibody Developability Consortium founding members (2026-10-14)
• [10] SEC 8-K Ex.99.1 (Soaring Eagle closing) and GEN/Crunchbase (2026-10-14)