A Company Betting Its Physics-Based AI Will Fuel a Quantum Leap in Drug Discovery

Episode Summary

Executive Summary: Tom Miller says Iambic Therapeutics uses physics-informed and quantum-mechanics-based AI to design better drug candidates faster by combining predictive models with rapid in-house design-make-test cycles. The company has advanced multiple oncology programs, raised $100 million in Series B funding, and aims to validate its platform through upcoming Phase 1 trials and early human efficacy signals.

Main Topics: Iambic’s AI/quantum mechanics drug design platform (Priority: 5/5): Miller explains that Iambic combines physics-based modeling and AI to search chemical space more efficiently and predict novel molecules beyond conventional data-driven methods. Closed-loop discovery and human-in-the-loop execution (Priority: 5/5): The company emphasizes weekly design-make-test cycles using in-house chemistry and biology, with scientists remaining central to decision-making rather than being replaced by automation. Platform validation through speed and candidate quality (Priority: 5/5): Iambic argues that the strongest proof of its approach is repeatedly delivering development candidates to IND in about 24 months, including multiple programs reaching clinic quickly. Pipeline strategy and lead oncology programs (Priority: 5/5): The transcript highlights IAMH1, a HER2-targeted tyrosine kinase inhibitor, and IAMC1, a selective CDK2-4 inhibitor, both designed to address resistance and improve therapeutic windows. Business evolution and financing (Priority: 4/5): Iambic has shifted from a broader technology orientation toward a clinical-stage biotech identity, supported by a $100 million Series B that funds near-term clinical milestones through 2026. Partnering, capital allocation, and portfolio focus (Priority: 4/5): Miller discusses the need to balance platform opportunity with limited capital, using selective partnerships and focusing resources on the most promising lead assets. Investor expectations for AI in biotech (Priority: 3/5): He notes that investor enthusiasm for AI has matured, with more focus on whether the technology truly translates into better medicines rather than novelty alone.

Key Arguments: Physics-informed AI is more data-efficient than purely data-driven models, allowing better prediction in novel chemical space with fewer experimental examples. OrbNet and related tools use molecular orbitals and quantum mechanics to improve prediction of conformational and interaction energies. Iambic’s differentiator is not only model quality but the rapid closed-loop system that converts AI designs into weekly experimental data. Human expertise remains essential; AI is meant to augment medicinal chemists and biologists rather than replace them. The platform is flexible enough to support multiple mechanisms of action, including covalent, non-covalent, allosteric, and protein-protein interaction modulation. A key measure of success is speed to IND and repeated delivery of clinical candidates across challenging targets. IAMH1 is positioned as a potentially best-in-class HER2 inhibitor because of brain penetrance, pan-mutant activity, and high selectivity over EGFR. IAMC1 aims to overcome resistance and toxicity limitations of approved CDK4/6 inhibitors by more selectively targeting CDK2/4 while sparing homologous off-targets. The company’s funding and pipeline strategy are designed to support clinical entry while preserving flexibility for additional platform-enabled programs and collaborations. Investor attention has shifted from AI hype to proof that AI can create differentiated therapeutics and clinical value.

Data Points: Series B financing: $100 million - Completed in October to support the platform and advance multiple candidates into clinical development. Time from program launch to IND submission: ~24 months - Miller says Iambic can deliver high-quality development candidates to clinic in about two years. Company team size: ~60 people - Roughly half are AI/computational scientists and half are drug development/discovery experts. Internal data-generation pace: hundreds or thousands of molecules per week - Through in-house plate-based chemistry and biology in rapid design-make-test cycles. First IND submission: This quarter - Miller says the company had its first IND submission imminent at the time of the interview. Second IND submission: Very soon after the first - He indicates a second IND is expected shortly after the first. Cash runway: Through 2026 - The untranched Series B is expected to fund operations until 2026. Lead program target: HER2 cancers - IAMH1 is described as a tyrosine kinase inhibitor for HER2 cancers. Selectivity claim: Over 1000-fold selectivity versus EGFR - Miller cites this as a differentiator for IAMH1. Approved HER2-specific inhibitor count: 1 approved specific HER2 inhibitor (tucatinib) - Used to frame unmet need despite progress in HER2 cancer treatment.

Pivotal Quotes: "We think about it in terms of three aspects of differentiation." — Tom Miller: Explaining why Iambic believes its platform stands out from other AI drug discovery companies. "AI design isn't enough." — Tom Miller: He emphasizes that speed, wet-lab execution, and repeated closed-loop learning are essential to the platform. "The most important thing is really to validate the platform on the basis of delivering differentiated development candidates to clinic at the fastest possible pace." — Tom Miller: On what ultimately proves Iambic’s approach is working.

Implications: Iambic is betting that physics-informed AI can become a repeatable engine for clinical-stage drug discovery. If upcoming trials show efficacy, it could strengthen investor confidence in AI-native biotech platforms and accelerate broader adoption of closed-loop, human-guided discovery models.