Lux Capital's Deena Shakir on AI in healthcare: scribes are crowded, clinical trials and drug development are the real opportunity

Summary

Deena Shakir, partner at Lux Capital, draws a clear line between where AI in healthcare has already peaked and where the real investment opportunity now sits. AI scribes, once the lowest-hanging fruit in the sector, are now a saturated market with over 100 competing products and more forming daily. The smarter capital, in her view, is moving toward clinical trials and drug development.

The Real Opportunity: Clinical Trials and Drug Development

Pharmaceutical spending on specialty drugs represents tens of billions of dollars in addressable market, and getting those drugs to patients faster remains a massive unsolved problem. Lux is backing Trial Library, which uses AI and technology to match hard-to-reach patients with clinical trials, and Eli Lab (co-invested with Elad Gil), which applies generative AI to regulatory and medical writing to accelerate drug approvals. Drug discovery itself Shakir characterizes as the "V1" of AI in biopharma, already well-populated with early entrants.

Scribes Are Crowded, Not Dead

The scribe category did unlock genuine value. LLMs resolved the adoption friction that had stalled earlier documentation tools, and physician note-taking was a clear pain point. But with 100-plus vendors now competing, differentiation is thin. Shakir notes that in most clinical encounter contexts, audio-based AI is sufficient and hardware add-ons like cameras or wearables introduce unnecessary PHI and privacy complexity without commensurate benefit.

Big Tech and Big Payers as Distribution, Not Innovators

She draws on her own time helping build Google Health to argue that large tech companies will function as distribution infrastructure and digital front doors for healthcare, not as the source of transformative innovation. Google remains where many healthcare journeys begin before patients even self-identify as patients. OpenAI is increasingly entering that same layer, particularly as Fidji Simo publicly positions ChatGPT around therapy and health use cases.

Meanwhile, major national payers are moving faster on clinical AI adoption than many observers expect. Shakir describes a recent board meeting where a senior executive from one of the largest U.S. payers outlined AI bot deployments for patient engagement that were more ambitious than most VC pitches she had seen. AI in payer operations has shifted from a discretionary experiment to a stated mandate, visible in earnings calls across the sector.

Robotic Surgery Is Already Here

On surgical robotics, Shakir pushes back on forward-looking framing: the transition is already underway. Lux was lead investor in Orus, acquired by Johnson & Johnson. Robotic-assisted procedures are now standard across laparoscopic surgery, ultrasound, and increasingly IVF, where Lux holds a computer-vision-enabled software position. The remaining constraints are patient preference and liability, not technology readiness.

Brain-Computer Interfaces: Capital-Intensive, Still Unproven at Scale

Lux was an early backer of CTRL-Labs, subsequently acquired, and its founder Reardon is now a venture partner at the firm. Shakir does not expect BCI to follow the same fragmentation path as AI software. The category remains capital-intensive and scientifically demanding. Lux is seeing early-stage activity through its Lux Labs initiative, but the path from research acceleration to commercialization is still unproven.

AI Therapy: Real Adoption, Real Risk

Studies published in JAMA show AI scoring comparably or higher than humans on empathy metrics in some clinical comparisons. ChatGPT is already being used at scale for mental health support amid a documented epidemic of loneliness. Shakir does not expect AI to displace therapists in the near term, but treats current usage as a fact to be governed rather than a trend to be debated. The key policy question is guardrails, particularly given documented cases linking AI companion platforms to self-harm.