Investors

CIPHER^TM: Solving the Undruggable Challenge with Next-Generation Explainable AI

Cipherica’s development of a next-generation Explainable AI (XAI) platform represents a shift from “black-box” machine learning toward transparent, AI-driven drug discovery, specifically aiming to unlock therapeutic targets previously considered untreatable. By providing interpretable AI decisions, the platform aims to address the high failure rate in drugging complex, “undruggable” targets associated with neurodegenerative diseases and aggressive cancers.

The Current Market Is Ripe for Next-Generation Generative AI in Molecular Design

The pharmaceutical industry stands at an inflection point where traditional generative AI approaches largely built on SMILES strings, 2D graphs, or ligand-based models are hitting structural limits, while demand for innovative small-molecule solutions has never been higher. Global spending on drug discovery technologies is projected to reach approximately $77.6 billion in 2026, nearly doubling by 2032, driven by AI-native platforms that promise to compress timelines and costs. Within this, the dedicated generative AI segment for drug discovery has grown from roughly $250 – 260 million in 2024 – 2025 to an expected $330 million in 2026, with forecasts showing a robust 26 – 30% CAGR through the mid-2030s, reaching $2.7 – 2.8 billion by 2034 – 2035. Broader AI in drug discovery is expanding even faster in some estimates, from $2.35 billion in 2025 toward $13 – 16 billion by 2033 – 2034.

This growth is fueled by three converging forces:

• Persistent pipeline attrition and cost crisis: Traditional discovery still fails ~90% of candidates, with “undruggable” targets (intrinsically disordered proteins like amyloid-β, Tau, α-synuclein, mutant p53, and TDP-43, or flat/featureless surfaces) representing a massive unmet market estimated in the hundreds of billions annually across neurodegeneration, oncology, and rare diseases. Current generative models struggle here because they rely on low-dimensional representations that ignore full topological, dynamical, and 3D conformational realities, leading to poor synthetic accessibility, hallucinated structures, and weak binding predictions in disordered regions.
• Maturation of compute and data: Post-AlphaFold advances have solved many structured proteins, but IDPs and dynamic ensembles remain unsolved. Pharma now has petabytes of proprietary multimodal data (cryo-EM, X-Ray, NMR, MD simulations, phenotypic screens) ready for physics-aware generative engines, yet legacy SMILES/graph models cannot fully exploit it without losing explainability or thermodynamic fidelity.
• Investor and Big Pharma validation: In 2025 alone, 114 AI/ML drug-discovery deals totaled $43.4 billion, shifting from pilots to platform-scale integrations. Pharma giants are actively licensing foundation models, generative platforms, and AI-designed candidates rather than building everything in-house.

The market is therefore primed for technologies that move beyond statistical pattern-matching to rigorously structure-preserving, explainable generative engines capable of handling multi-scale dynamics and producing thermodynamically stable, synthesizable molecules for precisely these “undruggable” challenges.

Proven Deal Structures for Licensing This Technology to Biotech and Pharma

Licensing in this space has evolved rapidly in 2025 – 2026 into flexible, risk-sharing models that reward platform performance while protecting IP. Deals are no longer one-off asset transfers; they blend platform access, co-generation of candidates, and milestone-driven economics. Here are the primary structures observed in recent transactions, tailored to how a platform like CIPHER^TM (with its categorical mappings, SKI-powered rewrites under statistical-mechanical sampling, and DFT-validated energies) would be positioned:

1. Platform Access + Subscription License (Foundation-Model Style): Pharma receives non-exclusive or field-limited access to the core generative engine (e.g., API or on-prem deployment) for internal use or joint projects.
   • Typical terms: $50 – 150 million upfront (or multi-year subscription fees), annual maintenance/usage fees, plus data-generation add-ons.
   • Examples: GSK’s 5-year licensing of Noetik’s foundation models ($50 million upfront + subscriptions); Lilly’s repeated expansions with Insilico’s Pharma.AI.
   • CIPHER^TM fit: License the functorial + statistical-mechanical core for target-specific campaigns; pharma nominates IDPs, and the engine runs Metropolis-Hastings sampling internally. Retain ownership of the platform IP while granting rights to outputs.
2. Target-Specific Collaboration + Option to License Candidates: Pharma nominates 3–10 “undruggable” targets; the licensor runs the full generative pipeline (functorial pattern detection → SKI rewrites → statistical-mechanical annealing → DFT refinement) and delivers optimized small-molecule candidates. Pharma has an exclusive option to license selected hits/lead series.
   • Typical terms: $50 – 120 million upfront + research funding, $100 million–$6 billion in development/regulatory/commercial milestones (heavily back-loaded), tiered royalties (mid-single to low-double digits on net sales), and sometimes equity stakes.
   • Examples: XtalPi – DoveTree ($51 million upfront + $5.9 billion milestones + royalties), Monte Rosa-Novartis (molecular glues, $120 million upfront + $5.4 billion), AstraZeneca – CSPC (AI-driven small molecules, $110 million upfront + $5.2 billion), Insilico-Lilly expansions (Insilico’s AI-enabled discovery platform, $115 million upfront and up to $2.75 billion total plus royalties)
   • CIPHER™ advantage: Emphasize XAI traceability (morphisms + rewrite sequences + thermodynamic averages) and superior performance on IDPs – features that de-risk downstream development and accelerate IND filing.
3. Multi-Target Portfolio or Co-Development Alliance Broader strategic partnership covering an entire therapeutic area (e.g., neurodegeneration or oncology). Includes joint steering committees, shared wet-lab validation, and rights to out-license co-generated assets.
   • Typical terms: Larger upfront ($100 – 300 million), shared R&D costs, milestone pools exceeding $1–6 billion, royalties, and sometimes co-commercialization options in select geographies.
   • Examples: Servier-Iktos (>$1.5 billion generative AI + robotics); Lilly–Nimbus (obesity, $55 million + $1.3 billion).

These deal structures minimize upfront risk for pharma while providing the technology provider with immediate capital, validation, and downstream upside—exactly the model that powered 2025’s $43 billion deal wave. With CIPHER^TM ’s unique combination of categorical rigor, combinatorial generation under thermodynamic control, and inherent explainability, licensing deals can be positioned not as another incremental AI tool, but as the first platform explicitly engineered to deliver the small-molecule cures that have eluded every prior approach. The market timing is optimal: Big Pharma is actively seeking differentiated generative engines, and the financial templates are already battle-tested.

Investment Opportunity

We are actively seeking strategic investors who share our vision of transforming early-stage drug discovery. If you would like to learn more about our technology, team, pipeline progress, or investment opportunity, we invite you to contact us to schedule a dedicated presentation.

For investor inquiries: Please email [email protected]

We look forward to discussing how Cipherica can deliver substantial therapeutic and commercial impact in areas with massive unmet medical need.