Takeda's $4B AI-Designed TYK2 Inhibitor Zasocitinib Beats BMS's Sotyktu in Phase III Head-to-Head Trial, Nears 2027 Launch

An AI-Discovered Drug Succeeds in Phase III Trials!

2023Year, Takeda with40Acquisition with a down payment of hundreds of millions of US dollarsNimbusofTYK2Drug,and renamed tozasocitinib(TAK-279)。

Last night, Takeda announced thisTYK2Drugthe latestIIIPhase clinical information. Data show that the drug head-to-headHeadEfficacy Directly Outperforms BMS’s Same-Target Drug Sotyktu!

Top-line results show that,Zasocitib at the primary endpoint, i.e., PASI 100 response rate at Week 16as high as 33.4%, whereas it was 14% among patients treated with Sotyktu.

That is to say, inNo.16ZhouIn terms of the proportion of patients with complete clearance of skin lesions，The efficacy of Zasocitinib is more than 2.5 times that of Sotyktu.。

The study also shows that,Zasocitinib was superior to the active control across all key secondary endpoints, including PASI 90 response and a static Physician’s Global Assessment (sPGA) score of 0 at Week 16. Zasocitinib was generally well tolerated, with no new safety signals identified.

Takeda plans to2027First Half of the YearYear，PlacezasocitinibOfficially launched in the market.

For Takeda, the $4 billion gamble is finally paying off.

BIC Drug Targeting TYK2

TYK2YesJAKA member of the kinase family that plays a central role in the signaling of multiple key inflammatory cytokines。

2022Year9Month, Bristol-Myers SquibbofTYK2Allosteric inhibitor deucravatinib (deucravacitinib, with the trade nameSotyktu) marketing application for the treatment of plaque psoriasis approvedFDAApproved.

Thus,SotyktuBecomeThe world's first approved and marketedTYK2Inhibitor。

BMSInitially forSotyktuwith peak sales forecasts reaching as high as40hundred million US dollars, however, the reality is thatIts2025global annual total sales amounted to only2.91$100 million，The U.S. market also experienced a decline.。

The reason is that the competition in the psoriasis market is extremely intense; despite the advantages of oral administration, the efficacy and current maincompared to mainstream biologicsA significant gap remains, making it difficult to challenge the latter’s status as the gold standard.。

In contrast,Zasocitinibis aCalculate the predicted sumDesignedSmallMoleculeDrug，PossessingBICpotential.

This drug was initiallybyNimbus TherapeuticsandAICo-developed by Schrödinger Pharmaceuticals, thereby obtaining targetedTYK2highly selective drugs.

ZasocitinibAs aHighly Selective OralTYK2Inhibitor, can maintainIL-23and other core immune signaling pathways driving the disease24Hourly inhibition.

Based on in vitro data, compared to otherJAKCompared with enzymes,zasocitinibYesTYK2with over -fold higher selectivity10010,000-fold, which may not affectJAK1、JAK2andJAK3In the context of signal transduction, maximizeTYK2Inhibition.

2022years, based on goodIIbPhase data,Takeda60Acquisition for Hundreds of Millions of US DollarsThisTYK2Inhibitor, Prepaid40hundred million U.S. dollars.

Transaction stipulates that, subsequently based on post-launch sales revenue of the drug,NimbusHave the opportunity to obtain again20billion-dollar revenue share.

Computation-driven,Distinct from Traditional R&D Approaches

2009Year,Atlas VenturePartnerBruce Booth, combined with SchrödingerCEO Ramy Farid, co-foundedNimbus Therapeutics。

Since its inception, the company has transcended the traditional framework of drug development.

NimbusLarge-scale adoption of the Schrödinger computational platform, leveraging physics-based principles to guide drug development decisions and bypassing the traditional trial-and-error approach to drug discovery.

The core of this model is dual deep collaboration: first, the coordinated efforts of teams from both companies; second, the deep integration of medicinal chemistry and computational chemistry talent. In the traditional R&D framework of the industry, these two roles have long been siloed, whereasNimbusIntegrating these elements to create a proprietary competitive advantage.

The core R&D philosophy of the team is computation-driven drug design. The entire R&D process relies heavily onCalculationTechnical modeling and prediction, rather than the trial-and-error approach of traditional models.

Preclinical compound synthesis and testing, guided primarily by modeling results, represent a significant innovation in drug development paradigms.

AtTYK2Inhibitorin the research and development, the team extensively utilizedFEP+WaitSimulation technology,TargetingTYK2A specific site on the proteinJH2DomainConduct targeted therapy。

The results of this technologyRapidly predict the effects of molecules,The result was that a candidate molecule with super-high potency (at the picomolar level) was identified in just about three months.

Finally, the teamwhile optimizing its drug-like properties (such as stability and absorption), ultimately leading to the successful identification of a drug candidatezasocitinib。

PreviouslyMost pharmaceutical companies are unable to implement computational drug discovery and development at scale, whileNimbusThis has been achieved, with the full implementation of a computational R&D philosophy and empowering teams to lead R&D decisions through technology, which holds revolutionary significance in the field of drug development.

What other late-stageAIDrugs Worth Noting?

Today, AIThe designed drugs are increasingly entering the pivotal clinical validation phase, becoming2025-2026One of the most watched tracks in the pharmaceutical industry this year.

According to incomplete statistics,Currently, there are more than10AmountAIComputer-Aided Drug DesignIIPhase and Later-Stage Clinical Phases, covering multiple fields including oncology, autoimmune diseases, respiratory diseases, and weight management.

Here, we highlight several drugs with the most rapid progress that are worth attention:

Generate Biomedicines  GB-0895（III)

2025Year12Month,Generate BiomedicinesAnnouncing the Launch of Two GlobalIIIPhase Clinical Trial——SOLAIRIA1andSOLAIRIA2，This is the world's first completelyAIDesign and EntryIII-stage antibody drugs.

GB-0895is a targetedTSLPlong-acting monoclonal antibody，For the treatment of patients with severe asthma。IPhase data demonstrate a significantly prolonged half-life, supporting semi-annual dosing.，Current standard therapies typically require monthly or even more frequent injections.IIIPhase trial plans to enroll approximately1600patients with severe asthma, covering North America, Europe, Latin America, and the Asia-Pacific region.

Insilico Medicine Rentosertib (ISM001-055)

Rentosertibis the world's first to utilize generative AI throughout the entire process, from target discovery to molecular designAIdriven candidate drugs.

2025Year6Month, itsIIaPhase clinical results published in “Nature Medicine》, showing that compared with the placebo group,RentosertibTreatment of Patients' Forced Vital Capacity (FVC）FVCImproved by an average of compared to baseline98.4mL, while patients in the placebo groupFVCMeaningful decrease from baseline20.3milliliters, demonstratingRentosertibTreatmentIPF’s immense potential. Next, the company plans to launch in ChinaIIIPhase, launched in the United StatesIIbPhase Trial.

Jitai Tech MTS-004（III(completed on schedule)

2025Year10Month, Nimbus Therapeutics'MTS-004CompletedIIIPhase clinical trial, becomingChina's First CompletionIIIphaseAIFormulation-Enabling Drugs.

This drug is used for the treatment ofPseudobulbar Affect (PBA）。MTS-004ofAIThe application focuses on formulation optimization rather than molecular design. ThroughAINano Delivery PlatformNanoForge, reducing the preclinical formulation optimization cycle from the industry average1-2Years shortened to3to be completed within months.Currently, this drug has been1The down payment of hundreds of millions of yuan was successfully licensed out.

Medicilon  MDR-001（IIIPhase)

MindRank Therapeutics' Self-Developed Oral Small MoleculeThe GLP-1 receptor agonist MDR-001 officially launched its Phase III clinical trial (the MOBILE study) in China in December 2025. This isChina Takes the LeadbyAI-assisted designed small-molecule drug entering Phase III confirmatory clinical trials.

PreviouslyIIbPhase data show that treatment24Weight loss after weeks10.3%, and no drug-related serious adverse events occurred throughout the trial period, with a discontinuation rate due to adverse events of only0.8%, significantly outperforming competing products with the same target.IIIPhase trial plans to enroll approximately in China750overweight or obese subjects, conducted52Weekly Assessment, Estimated2028Year-end or2029Listed in [Year].

Recommended Reading