DARPA Bio-Attribution Challenge (2026)

DARPA Bio-Attribution Challenge (2026)

The DARPA Bio-Attribution Challenge was a two-round federal prize competition for computational tools that could identify biological signals in very large datasets and help determine their likely origin. It was not a grant solicitation or a wet-lab research award. Participants supplied software for evaluation in a secure, government-controlled environment; the challenge did not involve handling actual pathogens or biological materials.

This page is now a historical reference. DARPA’s official challenge page says that the program is complete, that its content is retained for reference, and that the page is no longer maintained. DARPA’s final news report also says that the second round concluded and that the competition awarded its full prize pool. There is no announced follow-on cycle on the official page, so readers should not treat this entry as an open application opportunity.

Key details

What the completed challenge tested

The competition focused on software and analytical performance rather than treatment, procurement, or general infrastructure spending. DARPA described two linked tasks:

1. Detect and characterize pathogens within complex environmental samples.
2. Determine the origin of engineered pathogens by identifying physical, chemical, or design signatures.

The problem was framed as one of finding a meaningful signal inside biological data at a scale that ordinary analysis systems cannot process quickly. DARPA described tools capable of working with petabyte-scale datasets in near real time. The intended use was to help investigators distinguish whether a biological event was natural, accidental, or intentional, while keeping the evaluation data controlled and non-sensitive.

The official page emphasizes that participants worked with deliberately curated data developed by Lawrence Livermore National Laboratory. Those data were designed to mimic complex scenarios without disclosing sensitive information. Software ran in a secure government-controlled setting, and no participant needed to receive or manipulate real biological material.

Competition structure and final result

Round 1, called Detection, lasted two months and concentrated on accurate identification and characterization of pathogens in complex environmental samples. Round 2, called Attribution, lasted one and a half months and asked qualifying teams to determine the origin of engineered pathogens using distinctive physical, chemical, or design signatures.

The final DARPA report describes the second round as the conclusive “Determination - Attribution” phase. Qualifying teams processed datasets ranging from 600 to 800 terabytes within a strict 24-hour window. Competitors combined deep sequence screening with geospatial and temporal metadata, while receiving very limited information about the samples. That design tested not only classification quality, but also throughput, software efficiency, and the ability to reason from sparse contextual signals.

The final event awarded the remaining $90,000 to the top three Round 2 teams: $50,000 for first place, $30,000 for second place, and $10,000 for third place. DARPA’s report identifies Team Crits-Christoph and Hakim, Team Kannadasan, and Aclid in those positions. Round 1 had the same $50,000, $30,000, and $10,000 prize structure. DARPA also recorded non-monetary recognition for categories such as Best in Show, Fastest Analysis, Most Innovative Method, Most Data-Efficient, Highest Precision, and Highest Accuracy.

DARPA’s official challenge page identifies 2026-06-30 as the awards ceremony date and describes the program as complete. The archive deadline in this entry uses that final published event date rather than the stale, unverified date previously present in the file. It is a closed-cycle date, not a claim that applications remain open.

Eligibility and application status

There is no current application route. The official DARPA page now labels the program complete and says the page is no longer maintained. No next round, open registration period, or replacement deadline is announced there. A new reader cannot join the 2026 event by following the archived URL, and should not send a proposal or prepare a new submission on the assumption that the competition is still running.

The official launch material described the intended participant pool broadly: DARPA encouraged individuals and teams with expertise in bioinformatics, data science, high-performance computing, information theory, and machine learning. That description is useful for understanding the challenge, but it is not a current eligibility notice. The archive does not present a new applicant checklist, an active registration form, or a future-cycle rule set. Any team considering a later DARPA competition must read that later competition’s own rules rather than carry these historical assumptions forward.

Who the challenge suited

While closed, the event was suited to teams with a strong software and systems core. Its design favored groups that could:

• build reliable bioinformatics or sequence-analysis pipelines;
• work with large distributed datasets and strict runtime limits;
• combine machine learning, information theory, and biological reasoning;
• produce repeatable software that behaves consistently under controlled evaluation; and
• reason about provenance signals instead of stopping at pathogen detection.

It was a poor fit for teams that needed wet-lab equipment, physical samples, general operating support, or an unrestricted research grant. It also demanded more than a promising model idea: a system had to process difficult inputs, respect the evaluation environment, and return usable results under time pressure. These fit observations are useful when assessing similar computational challenges, but they do not create a new route into this completed event.

Application steps: none for the closed cycle

There are no current application or submission steps to follow. The official source does not announce a new round, publish a new deadline, or provide a live registration path for a future competition. The correct action for a reader who wants to pursue this area is to monitor DARPA’s current challenge and Biological Technologies Office pages for a separately announced opportunity. A future announcement would need its own eligibility terms, dates, technical instructions, and contact route.

For historical context, the 2026 event was a virtual, software-evaluation competition rather than a conventional proposal process. Teams needed the ability to develop and run computational methods against DARPA’s controlled datasets, but the archived page does not present a current checklist of forms or deliverables. It would be misleading to turn general engineering observations into application requirements for a program that has ended.

Why the archive still matters

The completed challenge is relevant to teams studying large-scale bioinformatics, biological threat attribution, anomaly detection, and high-performance data processing. Its evaluation model shows the type of operational problem DARPA wanted to expose: sequence screening had to work at an extreme data scale, with limited contextual information and a hard time boundary. The final report says competitors combined sequence analysis with geospatial and temporal metadata, which makes the event a useful case study in multimodal evidence handling.

The secure setup also matters. By using curated data from Lawrence Livermore National Laboratory and avoiding actual pathogens, the event let teams test demanding algorithms without creating a need for biological sample handling. That boundary distinguishes this opportunity from laboratory grants, clinical studies, and programs that supply organisms or experimental materials.

The award structure provides another useful signal. The $180,000 pool was divided evenly between the two rounds, with three monetary ranks in each round. DARPA separately recognized speed, precision, accuracy, data efficiency, innovation, and overall performance. This combination rewarded systems that were not merely accurate in a narrow test, but also practical to run and reproducible under controlled conditions.

Decision summary

Do not apply: the 2026 DARPA Bio-Attribution Challenge is complete, its official page is an archive, and no next cycle is announced there. The page remains worth reading for its description of the two-round technical problem, the computational-only evaluation model, the prize structure, and the final results. Teams seeking an active opportunity should use a current DARPA announcement rather than this historical entry.

Official links and next steps

• Official DARPA challenge page: https://www.darpa.mil/research/challenges/bio-attribution
• DARPA final results: https://www.darpa.mil/news/2026/bio-attribution-challenge-round-2
• DARPA launch announcement: https://www.darpa.mil/news/2026/bio-attribution-challenge