The Company that Built TikTok’s Algorithm Is Now Designing Drugs for Diseases Pharma Called Undruggable

Why It Matters

If AI can reliably produce oral drugs for targets previously considered undruggable, it could dramatically lower treatment costs and reshape R&D pipelines. Success would validate large‑tech AI infrastructure as a competitive advantage in pharma.

Key Takeaways

• Anew Labs unveiled AI‑designed IL‑17 inhibitor at immunology conference.
• AnewOmni trains on 5 million complexes, designs across molecular scales.
• Oral small‑molecule could replace costly injectable antibody therapies.
• ByteDance joins AI drug race against Isomorphic Labs, Anthropic, Insilico.

Pulse Analysis

The convergence of massive data‑driven AI models and drug discovery is reaching a tipping point, and ByteDance’s entry underscores the trend. Anew Labs’ AnewOmni model leverages a training set of over five million biomolecular complexes, enabling it to generate candidates ranging from traditional small chemicals to peptide‑like scaffolds. This cross‑scale capability addresses a known limitation of earlier generative tools, which often excelled at one molecular size but faltered elsewhere. By automating the search through an astronomical chemical space, the platform promises to accelerate early‑stage hit identification and reduce reliance on costly high‑throughput screening.

Targeting IL‑17 illustrates the clinical promise of this approach. IL‑17 drives several autoimmune disorders, yet its shallow protein‑protein interface has resisted conventional small‑molecule inhibition, forcing the market toward injectable antibodies that generate billions in revenue but cost tens of thousands of dollars per patient annually. An oral inhibitor that matches antibody efficacy could disrupt that market, offering cheaper manufacturing, easier dosing, and broader patient access. Moreover, the pan‑spectrum design—blocking multiple IL‑17 isoforms—could improve therapeutic breadth and reduce resistance, a key advantage for chronic inflammatory diseases.

ByteDance is not alone; Isomorphic Labs, backed by Eli Lilly and Novartis, and AI‑focused firms like Anthropic and Insilico Medicine are racing to commercialize similar platforms. Yet the sector still grapples with a 90 % clinical failure rate, and AI‑generated hits must survive the translational gap from test tube to human body. Anew Labs’ four pipeline candidates remain preclinical, and the path to regulatory approval will demand years of investment. Nonetheless, the combination of a $300 billion parent company, world‑class AI infrastructure, and seasoned biotech advisors positions ByteDance to test whether algorithmic design can move beyond proof‑of‑concept and become a mainstream drug‑development engine.