Inceptive

Inceptive is an AI biology company headquartered in Berkeley, California with research operations in Berlin, founded in 2021 by Jakob Uszkoreit and Rhiju Das. Uszkoreit is a former senior researcher at Google Brain and a co-author of "Attention Is All You Need" (June 2017), the paper that introduced the Transformer architecture and that has become the most-cited AI paper of the past decade. Das is a Stanford School of Medicine biochemistry professor whose lab developed Eterna, the citizen-science RNA-design game whose crowdsourced data has shaped one of the principal RNA-folding training corpora. Inceptive applies transformer-based deep learning to the design of "biological software" — primarily RNA molecules engineered for therapeutic effect, including next-generation mRNA vaccines, RNA-based gene therapies, and protein-replacement constructs. As of April 2026, Inceptive is one of the small set of AI-biology startups founded specifically to translate the AI-foundation-model paradigm into a regulated drug-development pipeline rather than a research-tool product.

At a glance

• Founded: 2021 by Jakob Uszkoreit and Rhiju Das. Berkeley headquarters with a research site in Berlin.
• Status: Private. Series A closed in November 2023 at $100 million.
• Funding: Approximately $120 million in cumulative private capital across seed and Series A. Andreessen Horowitz led the Series A; backers also include the Nvidia-affiliated NVentures, Obvious Ventures, and Greylock.
• CEO: Jakob Uszkoreit, Co-Founder and Chief Executive Officer. Former Google Brain Distinguished Scientist; one of the eight authors of the Transformer paper. PhD computer science (TU Berlin).
• Other notable leadership: Rhiju Das, Co-Founder and Stanford biochemistry professor. Stephan Eismann, Head of Computational Biology. The senior team has recruited from Google Brain, DeepMind, Moderna, and Stanford.
• Open weights: No. Inceptive's models are closed and the company has not published model architectures or weights.
• Flagship outputs: AI-designed mRNA constructs delivered into preclinical pharma partnerships. Initial public partnership with Eli Lilly, announced October 2023.

Origins

Inceptive's founding thesis is that mRNA medicines, demonstrated at scale by the COVID-19 vaccines, were just the first commercial moment for what Uszkoreit calls "biological software" — molecules whose function is determined less by chemistry than by a sequence of instructions that can be written, debugged, and version-controlled. The 2020 mRNA-vaccine wave validated the broader category but exposed how primitive the design tooling was: vaccine-grade mRNA design relied on rules of thumb and iteration rather than the kind of direct sequence-to-function modeling that deep learning had recently transformed for protein structure (AlphaFold 2, July 2021).

Uszkoreit had spent 13 years at Google, joining as a software engineer and rising to a senior research scientist role within Google Brain. His co-authorship on the 2017 Transformer paper, alongside Ashish Vaswani, Noam Shazeer, Niki Parmar, Llion Jones, Aidan Gomez (later founder of Cohere), Łukasz Kaiser, and Illia Polosukhin, made him part of the cohort whose post-Google careers have produced a string of frontier-AI startups. Uszkoreit's specific research interest at Google had moved progressively toward applying deep learning to biological sequences. He left Google in 2021 to co-found Inceptive with Das, whom he had met through scientific collaboration on RNA modeling.

Das brings the wet-lab and biology-domain depth. His Stanford lab has run Eterna since 2010, a multiplayer game whose players have submitted RNA designs that get synthesized and tested in vitro, generating one of the largest publicly available RNA-design datasets. The 2020 launch of OpenVaccine (a community Kaggle competition for mRNA-vaccine stability prediction) gave Das direct experience with the gap between academic RNA modeling and the demands of clinical-stage drug development.

Inceptive emerged from stealth in March 2022 with a $20 million seed round led by Andreessen Horowitz. The November 2023 Series A added $100 million at a reported $300 million post-money valuation. The October 2023 partnership with Eli Lilly — under which Inceptive designs RNA molecules for Lilly therapeutic targets — was the company's first publicly announced pharma deal and validated the contract-research-services revenue model that runs alongside in-house pipeline development.

Mission and strategy

Inceptive's stated mission is to design RNA medicines using deep learning. The strategy combines three threads. First, in-house experimental infrastructure to generate proprietary training data — the company operates wet labs that synthesize and test the RNA molecules its models propose, closing the loop between in-silico design and in-vitro measurement at much higher throughput than typical academic labs. Second, foundation-model research that applies transformer architectures to RNA sequences, with the resulting models conditioned on protein-coding intent, secondary structure, and stability targets. Third, partnership revenue from pharmaceutical companies that fund design work against specific therapeutic targets, providing both capital and access to clinical-development infrastructure that a stand-alone biotech would take years to build.

The competitive premise is that the bottleneck in RNA medicine is not delivery (LNPs, the lipid nanoparticle delivery system pioneered by Moderna and BioNTech, are workable) or chemistry (RNA synthesis is industrialized) but the design of sequences whose translated proteins fold correctly, do not provoke unwanted immune responses, and remain stable in storage. AI design at scale, the bet goes, can compress what currently takes months of human iteration into days.

Models and products

• Internal RNA design platform. Closed-weights transformer-based models that take a target protein sequence (and additional design constraints such as half-life, codon-usage requirements, and storage temperature) and produce candidate RNA sequences ranked by predicted performance. The platform is not externally productized.
• Wet-lab generation pipeline. Automated synthesis-and-test infrastructure at the Berkeley site that produces and measures the molecules the models propose, generating the proprietary feedback dataset.
• Pharma partnership engagements. Active partnership with Eli Lilly (announced October 2023). Industry coverage has reported additional partnerships with European pharma companies but specific names have not been confirmed publicly.
• Selected academic publications. Inceptive researchers have published at NeurIPS, ICLR, and biology venues including RECOMB, mostly on RNA structure modeling and on benchmark methodology for RNA-design evaluation.

Distribution channels are exclusively business-to-business: pharma partnerships fund the contract-design work, and any clinical-stage candidates would move through the partner company's regulatory pipeline rather than Inceptive's directly.

Benchmarks and standing

Inceptive does not publish horizontal benchmarks because no widely accepted RNA-design benchmark exists at the level of detail the company's commercial work requires. The OpenVaccine Kaggle competition that Das co-organized in 2020 set a baseline for mRNA-stability prediction but did not generalize to therapeutic design. Inceptive's standing is measured through pharma-partnership signings and, ultimately, through whether designed candidates progress through the partner's preclinical and clinical pipelines.

Industry coverage has consistently grouped Inceptive with Isomorphic Labs (the Alphabet-spun drug-discovery company), EvolutionaryScale (the protein-language-model company spun out of Meta AI), Generate Biomedicines, and Insilico Medicine as the principal AI-foundation-model-native biotechs. Inceptive's distinguishing feature within that group is the explicit RNA-modality focus and the founder's transformer-research lineage.

Leadership

As of April 2026, Inceptive's senior leadership includes:

• Jakob Uszkoreit, Co-Founder and Chief Executive Officer. Former Distinguished Scientist at Google Brain (13 years at Google). Co-author of the Transformer paper.
• Rhiju Das, Co-Founder. Stanford School of Medicine professor of biochemistry. Founder of Eterna and OpenVaccine.
• Stephan Eismann, Head of Computational Biology. Former Stanford postdoc in protein-structure modeling.
• Senior research and operations leadership recruited from Google Brain, DeepMind, Moderna, BioNTech, and Stanford.

The senior team is small relative to the capital raised; staff count was reported in the 30 to 50 range as of late 2024.

Funding and backers

• Seed (March 2022): $20 million led by Andreessen Horowitz.
• Series A (November 2023): $100 million led by Andreessen Horowitz with NVentures (Nvidia), Obvious Ventures, Greylock, and existing seed investors.
• Cumulative capital approximately $120 million as of April 2026. The company has not announced a Series B.

Industry position

Inceptive occupies a niche position at the intersection of foundation-model AI and RNA therapeutics. The Uszkoreit research credibility and the Andreessen Horowitz-led capital base have given the company an unusually visible profile relative to its commercial maturity, with industry coverage consistently treating Inceptive as a bellwether for whether AI-foundation-model methods will produce a generation of regulated-medicine companies in the next decade. The company has not yet announced a candidate in clinical trials.

The structural risk is the long timeline. RNA candidates designed in 2023 to 2025 will not produce clinical readouts before 2027 to 2028 at the earliest, and the company's commercial trajectory depends on partner-pipeline outcomes that Inceptive does not directly control.

Competitive landscape

• Isomorphic Labs. Alphabet-affiliated drug-discovery company with a broader small-molecule, antibody, and protein scope. Different research origin (Google DeepMind) but overlapping AI-biology positioning.
• EvolutionaryScale. Spun out of Meta AI in 2024 by the FAIR ESM-protein-language-model team. Protein-modality focus rather than RNA, but a closely comparable founder-and-investor profile.
• Generate Biomedicines, Recursion Pharmaceuticals, Insilico Medicine, BenevolentAI. Earlier-cohort AI-biology companies. Generate is the closest peer on protein design; Recursion focuses on phenotypic-screening drug repurposing; Insilico is a generative-chemistry small-molecule company.
• Moderna, BioNTech. RNA-medicine incumbents with internal AI design teams. Both are potential customers and potential competitors at the platform level. Moderna's Drug Discovery and Drug Design platforms are the most direct internal-tooling comparator.
• Eli Lilly. Inceptive partner. Partners are not direct competitors but represent the principal alternative path: large pharma can build internal AI-design teams rather than buying contracted design.

Outlook

• The Eli Lilly partnership progression and any disclosure of designed candidates entering preclinical IND-enabling work.
• A potential Series B and the valuation step-up from the November 2023 round.
• Public disclosure of model architectures, training data, or partnership candidates — the company has been deliberately quiet on technical detail to date.
• The first publicly announced AI-designed RNA molecule from any of the AI-biology cohort to reach clinical trials, which would establish whether the design-throughput-as-bottleneck premise translates into commercially differentiated assets.
• Continued recruitment from Google Brain, DeepMind, and the broader AI-foundation-model cohort.

Sources

• Inceptive official site. Company reference.
• Jakob Uszkoreit LinkedIn. Co-Founder reference.
• Rhiju Das Stanford profile. Co-Founder reference.
• Eli Lilly partnership announcement. October 2023 partnership.
• Andreessen Horowitz Inceptive announcement. Lead investor reference.
• "Attention Is All You Need" paper. Founder's research lineage.