Lightspeed Healthtech Summer Newsletter

September 01, 2026

Welcome to the 3rd Edition of Lightspeed's Healthtech Newsletter! This is our semi-annual look behind the curtain of what we're building and where we're headed in Healthtech.

The first half of 2026 was about conviction: putting more capital behind companies we backed early. That included clinical AI (Doctronic and Neko Health), administrative AI (Assort Health), AI rollups (Gyde), frontier health (Science), and frontier AI (Anthropic).

We also started fresh at the Seed stage with founders we've been waiting to back: Almanac's Cyril Zakka, who wrote one of NEJM AI's most-cited papers on retrieval-augmented models for clinical medicine, and Signal Labs' Raj Ronanki, former CEO of Lyric and former president of Carelon Digital Platforms at Elevance. We have more to announce in the weeks ahead.

Toward the AI Doctor

Our team delivered a keynote at the FDA on clinical AI at the frontier, and put our AI-doctor thesis on paper. On Lightwork, our weekly podcast about AI, we mapped AI's path through healthcare by the patient's journey rather than by technology category. The patient is the atomic unit of impact in a $5 trillion system. Administrative work is where AI landed first: Abridge, Assort, Tennr. More than $1 trillion in spending still sits in administrative work, and 80% of scheduling still happens by phone, even as most U.S. physicians now use AI in their workflows.

The clinical frontier is opening now. We see a continuum from physician-in-the-loop tools, such as decision support, to agentic systems with guardrails, such as triage, and eventually to more autonomous uses, beginning with tightly bounded tasks such as prescription renewals and moving toward the “AI doctor.” Abridge's Clinician Intelligence Platform, unveiled in June, is the clearest example: what began with ambient documentation now includes in-visit clinical decision support and other functions across the encounter, all in a single platform that can deploy across an enterprise.

To explore where this shift is going, Abridge, Anthropic, and Lightspeed hosted a hackathon on the future of agentic AI in healthcare, bringing together clinicians and engineers from around the world. From more than 100 projects, three winning teams focused on home care, behavioral health, and emergency triage.

Bringing the Ecosystem Together

This summer we hosted our Annual General Meeting in Napa Valley. Our global team was joined by some leading AI founders and LPs, including health systems that together care for millions of patients, to exchange ideas and expertise.

The healthtech sessions covered two recent milestones: Assort Health's $120M Series C, with its voice-AI platform now built on more than 190 million patient interactions, and early results from Doctronic's Utah prescription-renewal pilot. During the pilot's first phase, every AI-generated renewal requires physician approval before it reaches a pharmacy. In 97% of cases in which the AI recommended renewal, either the first or second reviewing physician found the recommendation appropriate based on the information the AI had gathered. We were also candid about many of the bottlenecks that remain: clinical-trial data loops in drug discovery, the jump from wedge to platform in administrative AI, and regulation and reimbursement in clinical AI.

We also spent more time with academic medicine. Ravi Mhatre joined UCSF's board. Galym Imanbayev joined Stanford Medicine's faculty as an adjunct professor. Brenton Fargnoli also co-authored a piece in NEJM Catalyst on why clinical training should include real exposure to venture and innovation.

For the World Cup, we brought founders together for matches in San Francisco, New York, Los Angeles, Atlanta, and Dallas. We talked about how important a team approach is and what we believe strong teams do well: adapt quickly, stay coordinated under pressure, and draw intensity to win from a love for the game.

Where We’re Headed Next

As we move into the fall, we're focused on five areas:

1

AI-Native Specialty Care:
Amidst a widening clinician shortage, we continue to see AI as a critical supply lever. Building upon our AI-native primary and proactive care investments with Doctronic and Neko, we’re looking next to AI-native care in bounded specialties, particularly those taking a vertically integrated vs. vendor approach.

2

Payer Adoption:
Employer health costs are projected to rise near double digits for the fourth straight year. Yet payer adoption of AI remains patchy. Adoption at scale runs through the national payers, so we're focused on the startups working most closely with them.

3

Pharma Commercial:
Specialty drugs now represent the majority of drug spending. That shifts the emphasis in pharma commercial from promotion toward operations such as benefits checks, prior authorization, and medication adherence. Each specialty prescription is a managed case, and hubs built on manual, services-heavy models may not be able to keep pace.

4

Robotics:
Recent advances in robotics have been striking; Lightspeed-backed Skild's in-context learning video is one example. In healthcare, robotics usually brings surgical systems to mind. We think the physical layer will follow the same arc software did: administrative before clinical, starting in places like manufacturing.

5

Data & Models:
As frontier labs look for healthcare-specific data for pretraining, post-training, and reinforcement learning, a market is emerging to acquire, curate, and supply it. The biggest winners may be the intelligence layers built on each healthcare organization's own data, workflows, and trust.

If you're working in these areas, whether as a founder, operator, customer, or investor, we'd like to talk. 

This fall, you'll find us speaking at HLTH, the World Medical Innovation Forum, Mayo Clinic Investment in Healing, the Digital Health New York Summit, the Aegis Summit, and the Second Opinion Summit. We're also launching Venture Doctors, a podcast hosted by Lightspeed's physician-investors.

Best,

Lightspeed Healthcare
Brenton Fargnoli, Galym Imanbayev, Jonathan MacQuitty, Naveed Matinfar, Ravi Mhatre, Shaurya Aggarwal, Shelley Chu, and Will Kohler