Earnings Call - EXSCIENTIA PLC

Financial Performance:

• Completed an oversubscribed $200 million financing in October, bringing pro forma cash and equivalents to $749 million as of September 30.
• Extended cash runway into 2029 through financing and capital concentration.
• Lowered 2026 full-year cash operating expense guidance to under $375 million (down from $390 million), representing a nearly 40% reduction from 2024 pro forma expenses.
• Further reduced 2027 cash operating expense guidance to under $360 million.
• Total cash inflows from partners now exceed $535 million, including over $100 million in milestones earned in the last two years across nine milestones.
• Sanofi option triggered a $10 million payment, bringing total proceeds from Sanofi to $144 million.

Guidance and Future Outlook:

• Target initiation of registrational study for REC481 (FAP) in the second half of 2027, pending FDA alignment.
• Sanofi partner candidate is expected to enter Phase 1 in 2027.
• Preliminary data from REC1245 endometrial cancer expansion cohort expected in the second half of 2027.
• Plan to advance additional targets with Roche Genentech and explore new tumor types for REC1245 expansion.
• Commitment to lowering OPEX year-over-year while aggressively investing in highest-conviction pipeline and AI engine programs.

Business Segments and Product Lines:

• REC481 (FAP): Updated Phase 2 data shows 87% of evaluable adults had reduced total polyp burden after 12 weeks; 92% of upper GI patients showed reduction. Durability observed with 85% of patients maintaining burden reduction off-treatment for 12 weeks. Significant reduction in high-risk polyps (>5mm).
• REC1245 (Oncology): First-in-class RBM39 degrader; observed 3 partial responses (2 confirmed in endometrial cancer, 1 unconfirmed in hepatocellular carcinoma) in heavily pretreated patients. Selected 45mg daily dose for expansion cohorts based on tolerability and activity.
• Partnerships:
  • Sanofi: Optioned a potential first-in-class oral dual-target INI candidate for inflammatory disease.
  • Roche Genentech: Advanced first unexplored neuroscience target from proprietary maps; developed novel approach for an "undruggable" oncology pathway.
• AI Platform: Generating proprietary data at scale (50 petabytes), synthesizing 80-90% fewer compounds than industry benchmarks for first-in-class molecules, and achieving trial site screening and enrollment rates 30-60% faster than industry averages.

Market and Competitive Landscape:

• FAP market addresses an estimated 37,000 post-colectomy patients in the US/EU5 with a $15 billion total addressable market; no approved therapies currently exist.
• REC1245 targets over 100,000 patients annually in hard-to-treat tumors (endometrial, HCC, ovarian, pediatric), offering a distinct mechanism of action compared to emerging ADCs (e.g., TROP2).
• Management emphasizes "repeatability" across the AI engine as a key differentiator against competitors relying on single-target or single-dataset approaches.

Risks and Challenges:

• Clinical: REC481 safety profile includes grade 3 adverse events (anemia, CPK elevations, skin reactions) in a subset of patients; dosing regimen (chronic vs. intermittent) requires confirmation in Phase 3.
• Regulatory: Polyp burden is not a traditional regulatory endpoint; Phase 3 must demonstrate clinical benefit in hard endpoints (cancer prevention, delay of surgery).
• Operational: Reliance on AI platform predictions requires continuous validation; "undruggable" target validation is complex.
• Financial: Despite strong cash position, continued efficiency is required to reach milestones; partnership milestone realization depends on partner decision-making timelines.

Management Commentary and Tone:

• Tone is confident and forward-looking, emphasizing the "repeatability" of their AI-driven drug discovery engine across biology, design, and clinical evidence.
• Management highlighted the successful integration of new Board Chair Fred Hassan (former Pfizer/Amgen leader) to guide the company through clinical proof-to-approval stages.
• Emphasis on data-driven capital allocation, prioritizing wholly-owned programs (REC481, REC1245) while maintaining high-conviction partnerships.
• Executive team expressed strong conviction that the closed-loop AI system (data-to-model-to-experiment) creates a sustainable moat.

Analyst Questions and Answers:

• Impact on Standard of Care (REC481): Management and KOLs confirmed the drug could reduce surveillance frequency and delay/avoid major surgeries (duodenectomy/Whipple) by reducing polyp burden and size, particularly in the upper GI tract.
• Patient Retention: KOLs and patients indicated a high willingness to tolerate side effects if the drug delays morbid surgery or reduces monitoring burden.
• AI Differentiation: Management cited the combination of proprietary data quality, the closed-loop learning system, and bilingual talent (science + AI) as primary competitive advantages over generic AI labs.
• REC1245 vs. ADCs: The distinct RBM39 mechanism allows for potential synergy or use in patients who have progressed on ADCs; expansion cohorts will validate position in the line of therapy.
• Capital Allocation: Company will fully fund high-conviction assets (REC481, REC1245) while maintaining discipline, using the extended runway to reach critical milestones without dilution.

Other Key Points:

• Board Update: Fred Hassan appointed as non-executive Chair; Rob Hirschberg remains on the board.
• Patient Insights: Patient panel highlighted the financial and emotional burden of lifelong surveillance and major surgeries, underscoring the unmet need for oral, disease-modifying therapies.
• Drug Holiday Potential: Clinical data suggests durability off-treatment, opening possibilities for intermittent dosing strategies to improve patient quality of life.
• Takeda License: REC481 is a MEK1-2 inhibitor originally licensed from Takeda, selected by Recursion's AI engine for its ability to reverse APC-deficient cell changes.
• Data Factory: The proprietary "scientific data factory" enables rapid hypothesis testing and model refinement, reducing time from target identification to candidate selection.