Axios Pro Rata: "Disappointingly… | Axios Pro Rata Summary

1. Key Themes

AI drug discovery is over-hyped relative to clinical results

Despite massive capital inflows, AI has not meaningfully improved the odds or speed of getting drugs through the hardest, most expensive stage of development. The peer-reviewed paper "calls the clinical impact of AI 'disappointingly limited.'" The core issue is data quality, not algorithmic capability: "We don't know how to clean this stuff up and categorize it for real ML/AI, and honestly, we don't even know if it can be," per biotech journalist Derek Lowe. Despite "$40 billion into AI biopharma this decade," per PitchBook, "there are precious few medicines that have moved beyond Phase II."

The real bottleneck is regulatory and operational, not computational

Industry insiders argue AI is being applied to the easy parts of drug development, not the structurally expensive parts. Bijan Salehizadeh of NaviMed Capital states: "The biggest costs are in the D of R&D, where Phase 2 and Phase 3 costs are massive and inflated by CROs and baked-in inefficiencies and regulation ... I'm afraid the AI applications in development will be surface level, focused on low-hanging fruit like biostats and data lock ... But not the big stuff of how to enroll a phase 2 or phase 3 program at hundreds to thousands of global sites."

Legacy consumer brands are shedding international units to refocus on core turnaround

Starbucks' potential Japan stake sale mirrors its earlier China divestiture, signaling a broader strategic pattern of using foreign unit sales to fund and focus a domestic reinvention. "This highlights how the ubiquitous coffee chain seeks to eliminate distractions while trying to revitalize its core U.S. business." Notably, this is happening even as the unit performs well — "the company crediting Japan as a key driver" of international comp sales growth of 5.7%.

Crypto's top legislative priority is stalling amid political entanglement

The Clarity Act's failure to advance shows how founder/administration conflicts of interest can jeopardize industry-wide policy wins regardless of technical bipartisan support. "Much of the opposition relates to the Trump family's crypto investments, and a lack of ethics rules governing them." Even earlier optimism was tempered: "Clarity is on the one-yard line ... But to borrow a quote from Teddy Roosevelt: Nothing in the world is worth doing unless it means effort, pain, and difficulty."

2. Contrarian Perspectives

AI's biggest near-term value in biopharma may be in targeting existing drugs, not creating new ones. Rather than accelerating novel drug discovery, the most credible late-stage success story is using AI for patient/target selection with already-approved medicines: the Moderna/Merck Phase III trial "is using AI to choose tumor targets in cancer patients — and then treating them with an existing medicine... this is more about decision-making for specific individuals than it is wholesale drug development."
Optimism about AI "disrupting" biology may require infrastructure that doesn't yet exist — and may take a generation. Arch's Robert Nelsen frames the upside as conditional on future breakthroughs, not current tools: "We need more data for models of biology. We need regulatory reform to move faster. And when we have the cellular data and more population data, and eventually quantum computing, it will fundamentally disrupt the industry" — implicitly conceding it hasn't yet.
Rhetoric about AI safety regulation slowing life-saving innovation is undercut by lack of actual results. The author pushes back on the AI industry's framing directly: "AI advocates have pushed back hard on the past week's safety debate, arguing that slowing the frontier would delay life-saving medicines. Rhetoric, however, is no substitute for results."

3. Companies Identified

Starbucks (Nasdaq: SBUX) — Global coffee chain. Mentioned as weighing a majority stake sale of its Japan business (~$3B valuation) to refocus on its U.S. turnaround. "Starbucks stock is up 15% in 2026, but remains down nearly 15% over the past five years."
Moderna / Merck — Pharmaceutical companies. Cited as running one of the few credible late-stage AI-enabled trials: a "new Phase III trial ... using AI to choose tumor targets in cancer patients."
Arch Venture Partners — Biotech VC. Featured for its view on the future of AI in biology via founder Robert Nelsen's commentary on data and regulatory needs.
NaviMed Capital — Healthcare-focused PE/VC firm. Featured for partner Bijan Salehizadeh's skeptical, operationally grounded critique of AI drug development hype.
OpenAI — Reported in "early talks to raise new funding at a $1.2 trillion valuation," signaling continued mega-scale AI investor demand.
Factory — SF-based AI coding startup that "raised $200m at a $5b valuation" from a broad syndicate including Sequoia, Khosla, and Blackstone — notable for high valuation in AI coding tools space.
Wonder — Food retailer/tech platform that raised additional Series D funding from DoorDash and sold its Grubhub Dining Campus to DoorDash for $300m, indicating consolidation in food-tech.
Coinbase — Referenced via chief legal officer Ryan VanGrack's earlier optimistic comments on the Clarity Act, contrasted with the bill's actual stalling.

4. People Identified

Robert Nelsen — Co-founder, Arch Venture Partners. Cited for his view on what's needed to unlock AI's real potential in biology: "We need more data for models of biology... eventually quantum computing, it will fundamentally disrupt the industry."
Bijan Salehizadeh — Managing Partner, NaviMed Capital. Cited for a sharply skeptical, tactically detailed critique of where AI is (and isn't) being applied in drug development trials.
Derek Lowe — Biotech journalist (Science.org). Cited for his blunt assessment of AI drug-discovery data quality: "We don't know how to clean this stuff up and categorize it for real ML/AI... we don't even know if it can be."
Ryan VanGrack — Chief Legal Officer, Coinbase (ex-Citadel Securities, Obama administration). Cited for his earlier confident prediction on the Clarity Act's passage, now undercut by events.
Brian Niccol — CEO, Starbucks. Cited regarding his turnaround strategy: "efforts to lure customers back through store upgrades and heavier marketing have helped stabilize demand but have also increased costs and pressured margins."

5. Operating Insights

Founders in AI/biotech should be wary of overselling timeline compression. The article notes "many AI proponents have claimed these timelines will be meaningfully shortened," yet Phase II/III bottlenecks remain unchanged — a caution for pitch decks that lean on speed claims without data-quality evidence.
Unsexy infrastructure work (data cleaning, standardization) may be the actual moat/opportunity in AI-for-science, since "this mostly boils down to cellular data sets, an unsexy thing to devote resources to when you could instead be swinging for blockbuster treatments." Startups tackling data curation, not just model-building, may be underinvested relative to their importance.
Divesting strong-performing but non-core geographic units can be a legitimate capital-discipline strategy during a turnaround, even when those units are contributing positively — Starbucks' Japan unit is a "key driver" of growth yet still a sale candidate to focus resources domestically.

6. Overlooked Insights

Regulatory reform is being framed by top biotech VCs as a prerequisite for AI's promise in drug development, not just a hoped-for accelerant — Nelsen's comment that "we need regulatory reform to move faster" suggests VCs increasingly see policy risk/opportunity as equally important as technical progress.
PE is quietly becoming a major force in Japanese corporate carve-outs, per the aside that a Starbucks Japan sale "could spark a major bidding war in Japan, where private equity has become very active" — a market-structure shift with implications beyond this one deal.