{"slug": "codebreaker-labs-files-first-form-d-leaves-offering-amount-undisclosed", "title": "Codebreaker Labs files first Form D, leaves offering amount undisclosed", "summary": "Codebreaker Labs, a Boulder startup co-founded by Ryan Gill, filed its first SEC Form D on August 26, signaling a fundraising effort for its plan to manufacture experimental data for genomics AI, though the filing does not disclose the offering amount. The company, operating as Codebreaker Tx, Inc., aims to test genetic variants in living cells to provide the experimental data that AI models need, a capital-intensive endeavor requiring high-throughput DNA synthesis, genome editing, and computing infrastructure.", "body_md": "# Codebreaker Labs files first Form D, leaves offering amount undisclosed\n\n**Codebreaker Labs, Ryan Gill's Boulder startup, wants to turn engineered genetic variants into experimental training data for biology AI.**\n\nBy [RuntimeWire Staff](/author/runtimewire-staff)\n· Published\n\nPrimary source: [SEC](https://www.sec.gov/Archives/edgar/data/2148535/000214853526000001/0002148535-26-000001-index.htm)\n\n## Why it matters\n\nCodebreaker's filing exposes a capital-intensive bet on the missing input for genomics AI: experimental data showing what genetic variants actually do in living cells.\n\nRyan T. Gill's [Codebreaker Labs](https://codebreakerlabs.io/?ref=runtimewire) filed its first [SEC Form D](https://www.sec.gov/Archives/edgar/data/2148535/000214853526000001/0002148535-26-000001-index.htm?ref=runtimewire) on August 26, making public a financing effort behind the Boulder startup's plan to manufacture the experimental data that genomics AI models need.\n\nThe filing records an exempt securities offering by Codebreaker Tx, Inc., Codebreaker Labs' Delaware corporate entity. The indexed record does not quantify the offering or identify a valuation or investors, so the filing establishes that Codebreaker is fundraising without supporting a round-size headline. A Form D is a notice of an exempt offering, rather than proof that a financing has closed.\n\nGill co-founded Codebreaker with Tanya Warnecke-Gill, the startup's chief technology officer, and Ryan Layer, its chief data advisor. Their thesis comes from a recurring problem across clinical genetics and drug discovery: sequencing has made genetic variants easier to find than to interpret. Codebreaker is building laboratory infrastructure to test those variants directly in living cells and record what changes.\n\nThat is an expensive proposition. It requires high-throughput DNA synthesis, genome editing, cell libraries, single-cell assays and enough computing infrastructure to turn the resulting measurements into a useful dataset. The Form D is the first public indication that Gill and his co-founders are bringing outside financing into that effort.\n\n### The founders behind Codebreaker\n\nGill has spent more than 25 years in genome engineering, according to [Codebreaker's biography](https://codebreakerlabs.io/?ref=runtimewire). He previously co-founded Inscripta, serving as its founding CEO and later chief science officer, and led Artisan Biosciences. His [academic CV](https://www.colorado.edu/chbe/sites/default/files/attached-files/curriculumvitae_chbe_web_rtg.pdf?ref=runtimewire) traces an earlier path through chemical engineering at Johns Hopkins University and the University of Maryland, followed by postdoctoral work at MIT and a faculty career at the University of Colorado Boulder.\n\n[Codebreaker says Warnecke-Gill](https://codebreakerlabs.io/?ref=runtimewire) also held senior technical roles at Inscripta and Artisan Biosciences. Codebreaker credits her with more than 35 patents and puts her in charge of the synthetic biology system that designs, builds and tests its variant libraries. The pairing gives Codebreaker a founding team that has already spent years turning genome-editing methods into products and protected intellectual property.\n\nLayer supplies the computational half of Codebreaker. A professor in computer science and the BioFrontiers Institute at the University of Colorado Boulder, he has developed tools for structural-variant analysis and querying large genomic datasets. His [university biography](https://www.colorado.edu/cs/ryan-layer?ref=runtimewire) identifies BITS, LUMPY, GQT and GIGGLE among that work. At Codebreaker, Layer leads the models intended to learn from the experimental data generated by Gill and Warnecke-Gill's platform.\n\nThe founders describe Codebreaker as the product of both scientific and clinical frustration. Vast sequencing datasets can reveal that a variant exists, while leaving patients and clinicians without a clear answer about what the variant does. Codebreaker's proposed answer is to engineer the change, place it in a relevant cell and observe the resulting behavior.\n\n### Building the data before building the model\n\nCodebreaker says its [causal genomics platform](https://codebreakerlabs.io/our-platform?ref=runtimewire) can introduce thousands of variants into living human cells and profile the effects through single-cell assays. The workflow starts with variants and genes selected from public and proprietary databases. Codebreaker then designs editing libraries, introduces them into relevant cell types including primary cells, and measures changes using single-cell omics.\n\nThe output is what Codebreaker calls its Codex: a growing collection of experimentally measured links between genetic changes and cellular effects. Codebreaker says it can program tens of thousands of variants in parallel and generate tens of millions of phenotypic fingerprints. Those scale figures remain Codebreaker's claims; Codebreaker has not published comparative performance data in the materials supporting the financing disclosure.\n\nCodebreaker began showing the platform outside the company before disclosing the financing. On December 19, 2025, [Parse Biosciences announced a collaboration](https://www.parsebiosciences.com/news/parse-biosciences-and-codebreaker-labs-partner-to-apply-whole-transcriptome-single-cell-profiling-and-causal-genomics-at-scale/?ref=runtimewire) combining Codebreaker's engineered variant libraries with Parse's Evercode whole-transcriptome single-cell technology. The companies said the work was designed to test thousands of variants in parallel and measure their effects at single-cell resolution.\n\nThe partnership points to Codebreaker's likely role in the biology AI stack. Codebreaker does not need to build the dominant genomics model to create a valuable business. It can sell or license experimental datasets, run studies for drug developers and clinical researchers, or use the Codex to improve its own interpretation models. Each path depends on the same asset: biological measurements that are difficult and costly to reproduce.\n\n### AI raises the value of a well-designed experiment\n\nSequence models have made variant-effect prediction a serious AI category. Google DeepMind's [AlphaGenome](https://deepmind.google/blog/alphagenome-ai-for-better-understanding-the-genome/?ref=runtimewire) processes long DNA sequences and predicts molecular properties associated with gene regulation. DeepMind also acknowledges limits around distant regulatory elements, cell-specific patterns and the path from molecular outcomes to complex disease.\n\nThose limits create the opening Codebreaker is pursuing. Models can infer from existing datasets, while experimental interventions can supply labels showing what happened after a specific variant was introduced into a specific cellular context. Better labels may improve model training, evaluation and target selection, provided the experiments represent the biology customers care about.\n\nCodebreaker's hardest task will be proving that measurements in engineered cells generalize across patients, tissues, disease states and experimental conditions. A 2025 [Nature Genetics review of causal machine learning in single-cell genomics](https://pubmed.ncbi.nlm.nih.gov/40164735/?ref=runtimewire) identified generalization, interpretability and cellular dynamics as core open problems. Scale alone will not settle them. Codebreaker will need to show that its Codex captures effects that remain meaningful outside the original assay.\n\nGill, Warnecke-Gill and Layer have built their careers around the tools needed to run that test. The Form D indicates they are preparing to fund it at startup scale. The financing details remain private, but the intended use is visible: Codebreaker is betting that the scarce input for the next generation of genomics AI will come from cells in the lab, rather than another pass over the same observational data.", "url": "https://wpnews.pro/news/codebreaker-labs-files-first-form-d-leaves-offering-amount-undisclosed", "canonical_source": "https://runtimewire.com/article/codebreaker-labs-first-form-d-causal-genomics-financing", "published_at": "2026-08-26 19:35:35+00:00", "updated_at": "2026-08-26 19:44:33.444006+00:00", "lang": "en", "topics": ["artificial-intelligence", "ai-startups", "ai-infrastructure"], "entities": ["Codebreaker Labs", "Ryan Gill", "Codebreaker Tx, Inc.", "Tanya Warnecke-Gill", "Ryan Layer", "Inscripta", "Artisan Biosciences", "University of Colorado Boulder"], "alternates": {"html": "https://wpnews.pro/news/codebreaker-labs-files-first-form-d-leaves-offering-amount-undisclosed", "markdown": "https://wpnews.pro/news/codebreaker-labs-files-first-form-d-leaves-offering-amount-undisclosed.md", "text": "https://wpnews.pro/news/codebreaker-labs-files-first-form-d-leaves-offering-amount-undisclosed.txt", "jsonld": "https://wpnews.pro/news/codebreaker-labs-files-first-form-d-leaves-offering-amount-undisclosed.jsonld"}}