Global Teams Race to Design Proteins With AI
The open contest sets seven nanobody and minibinder targets, with at least 7,000 designs slated for wet lab validation after registration closes.
Updated on Nov. 4, 2026 in Artificial Intelligence
The Changjiang Cup 2026 Global AI Protein Design Challenge announced its competition topics and complete rules on Sunday, with worldwide registration open until the end of November 2026. The competition is a contest, not a shipped product, and its validation stage is still to come.
Why it matters
The challenge is aimed at gathering global AI-for-science innovation through an open contest, with seven protein design targets focused on de novo design of VHH nanobodies or minibinders. Its scale of registrants suggests AI-driven protein design is becoming a shared competitive field across universities and pharmaceutical companies.
The challenge sets seven protein design targets and requires de novo design of VHH nanobodies or minibinders, with no fewer than 7,000 selected candidate proteins to undergo wet lab validation. The source does not report any measured performance results yet, so the competition's outcomes remain unknown.
The players
Shanghai Jiao Tong University
A major Chinese research university that initiates and hosts the challenge.
Changping Laboratory
A Beijing-based research laboratory co-hosting the challenge.
Lin Gang Laboratory
A Shanghai-based research laboratory co-hosting the challenge and linked to the municipal science authorities.
Shanghai Municipal Commission of Economy and Informatization
A Shanghai municipal authority guiding the challenge alongside the city's science commission.
Shanghai Municipal Science and Technology Commission
A Shanghai municipal authority guiding the challenge in science and technology policy.
The details
Protein design means creating new amino acid sequences that fold into a desired three-dimensional shape and perform a function. Nanobodies (VHH) are small antibody fragments derived from camelid antibodies, and minibinders are compact engineered proteins that bind a target molecule. Because a 100-amino-acid protein has 20 to the power of 100 possible sequence combinations, AI models are used to search this space, and the organizers will then test the most promising designs in a physical laboratory.
Timeline
Competition topics and complete rules were announced on Sunday, November 1, 2026.
Worldwide registration remains open until the end of November 2026.
The Tech Race
The challenge sits within the growing field of AI-driven protein design, a contest format open to universities and companies worldwide. Its registrant base of nearly all the global top 50 universities and over 50 pharmaceutical firms places it in direct competition with other AI-for-science programs.
Universities, research institutions, companies, and interdisciplinary teams anywhere in the world can register until the end of November 2026. Participation requires an entry in the open contest, and the source does not report any prizes, fees, or downstream product availability.
The takeaway
AI-driven protein design is moving toward open, large-scale competitions that test computational predictions against physical lab results. Watch for the organizing committee's announcement of which candidate proteins are selected for wet lab validation after registration closes.
What happens next
Worldwide registration closes at the end of November 2026, after which the organizing committee will begin wet lab validation of selected candidate proteins.
Further reading
For more on how AI is being applied to scientific discovery, see Artificial Intelligence.
Source note: This article includes information reported by China Daily.
