Better Living through Computational Chemistry

Episode Summary

Executive Summary: Nimbus Therapeutics CEO Abbas Kazemi explains how the company’s computational chemistry roots evolved into an AI-augmented drug discovery engine focused on precision, target quality, and fast go/no-go decisions. He also details Nimbus’s hub-and-spoke structure, which returns capital to investors, enables flexible partnering, and supports a private-company model built around asset-specific deals rather than an IPO exit.

Main Topics: Nimbus’s origins in computational chemistry (Priority: 5/5): Kazemi describes Nimbus’s 2009 founding around physics-based, structure-driven drug design, using tools like X-ray crystallography, cryo-EM, molecular docking, molecular dynamics, free-energy calculations, and WaterMap to improve target understanding and molecule design. How AI is now augmenting discovery (Priority: 5/5): Nimbus has expanded beyond classical computational chemistry to generative AI, machine-learning ADMET prediction, retrosynthesis planning, and competitive-intelligence tooling, but Kazemi stresses AI is a force multiplier rather than a substitute for judgment. Target selection and the value of failing fast (Priority: 5/5): A central thesis is that success depends first on choosing the right biological target; Nimbus uses a rigorous, democratized internal process to assess clinical, biological, and commercial attractiveness and will stop programs early when the odds are poor. Case studies: TIC2 and ACC validate the platform (Priority: 5/5): Kazemi points to the Takeda TIC2 deal and earlier Gilead/NASH-related ACC deal as proof that Nimbus can engineer differentiated molecules and create major partnering value from hard targets. Hub-and-spoke business model and investor liquidity (Priority: 4/5): Nimbus separates assets into LLC/subcorporate structures so each program can be partnered or sold independently, returning capital to investors while preserving upside and simplifying transaction structure. Current pipeline and financing runway (Priority: 4/5): The company’s latest $210 million private round funds its Warner helicase program in oncology and SIC2 program in immunology, with runway extending into 2027 and continued appetite for BD partnerships. AI’s future in biotech (Priority: 4/5): Kazemi predicts AI will become embedded across drug discovery and development, but the winners will be teams that ask better questions, design better experiments, and maintain high standards on data quality and target choice.

Key Arguments: Computational chemistry was foundational to Nimbus because it lets the company predict protein-ligand interactions and design molecules more rationally than brute-force screening. AI has already improved speed, cost, and precision in tasks like patent landscape review, ADMET prediction, retrosynthesis planning, and program triage. Nimbus does not view AI as a magic solution for novel target discovery; its real advantage is improving validated targets and helping kill weak programs faster. High-quality target selection is the most important determinant of success; tools are only useful if the biological hypothesis is sound. A democratized internal review process, where scientists pitch and debate targets, has sharpened Nimbus’s decision-making and improved hit quality. The company’s culture enables “failing fast” because employees act like owners and are empowered to pause weak programs without ego. The Takeda TIC2 partnership validated Nimbus’s precision-design approach by producing a molecule with exceptional selectivity and strong clinical differentiation. The Gilead ACC/NASH deal validated the platform by showing Nimbus could identify new disease relevance for a difficult target and create early BD value. The hub-and-spoke structure creates tax efficiency, capital recycling, and easier asset-level M&A or licensing deals. Nimbus believes the IPO is not the only exit; the company is designed to monetize assets via business development while staying private and capital efficient. AI will become ubiquitous in biotech, but it will augment rather than replace human medicinal chemistry and strategic judgment.

Data Points: Nimbus founding year: 2009 - Company was founded around computational chemistry and a capital-efficient business model. Takeda upfront payment: $4 billion - 2023 acquisition of Nimbus’s experimental TIC2 inhibitor. Takeda additional milestones: up to $2 billion - Potential milestone payments in the TIC2 deal. TIC2 total potential deal value: $6 billion - Upfront plus milestones in Takeda partnership. Gilead upfront payment: $400 million - 2016 deal for ACC program in NASH/liver disease. Gilead additional milestones: up to $800 million - Potential milestone payments in ACC deal. Total deals mentioned: north of $8.5 billion - Combined M&A and licensing deals across Nimbus’s history. Capital returned to investors: $4.1 billion - Returned through Nimbus’s asset/LLC structure and deal proceeds. Private financing round: $210 million - 2023 round co-led by GV and others to fund lead programs. Takeda TIC2 selectivity: 1.5 million full selectivity over other JAKs - Kazemi cited this as a key differentiator for Nimbus’s molecule. Bristol comparator selectivity: 75X - Ducravisitinib (BMS/TIC2 comparator) selectivity cited for contrast. Years Nimbus has operated: 16 years - Used to describe the company’s data generation, private-company longevity, and evolution. Kazemi tenure at Nimbus: 11 years - He references his experience leading target selection and BD. Program review group size: 15 to 20 scientists - Internal target-selection pitches are reviewed by a broad scientific group. Target pitch length: 20 minutes - Each scientist presents target rationale and development considerations. Runway: through next year and out into 2027 - Kazemi says the recent financing provides enough runway for milestones and development. Near-term IND target: early next year - For the SIC2 program moving toward IND. First-in-human timing: end of next year - Expected for the SIC2 program. Target portfolio examples: Warner helicase and SIC2 - Two lead programs funded by the latest capital raise.

Pivotal Quotes: "The first sin in drug discovery, as I have always believed, my predecessors at Nimbus have always believed, our investors believe, is target selection." — Abbas Kazemi: Explaining Nimbus’s philosophy that tools matter only if the biological target is right. "We celebrate all of our wins and we celebrate all of our failures. We celebrate when a project team is asked to make the decision to pause a program." — Abbas Kazemi: Describing Nimbus’s culture around discipline and failing fast. "AI helps us design smarter." — Abbas Kazemi: Summarizing how AI fits into Nimbus’s precision-focused discovery process.

Implications: Nimbus presents AI as a practical enhancer of disciplined drug discovery, not a hype-driven revolution. The model suggests biotech winners will combine strong target selection, rapid program pruning, and asset-level deal structures that recycle capital and reward precision.