Menlo puts healthcare AI spending at $3.7 billion as growth shifts toward patient care
Clinical copilots and patient-facing tools are growing faster than administrative software, while Microsoft and Epic capture most new spending on AI scribes.
Loading page…
Clinical copilots and patient-facing tools are growing faster than administrative software, while Microsoft and Epic capture most new spending on AI scribes.
Listen to this story
Menlo estimates U.S. healthcare AI spending reached $3.7 billion in 2026, but the competitive picture varies by workflow: EHR vendors are gaining ground in ambient scribing, while patient-facing tools offer specialists more room to compete. Longer provider procurement cycles raise the sales hurdle, and buyers’ preferences are not equally tied to their EHR vendor across tasks. Pharma AI spending and deployment are also scaling, though expected drug-development savings and approval timelines remain forecasts rather than demonstrated outcomes.
Microsoft’s Dragon Copilot and Epic’s Art helped established vendors capture 68% of new ambient-scribing dollars, though startups still grew 28%.
Clinical copilots grew 5.5 times to $550 million, while patient-engagement and front-office AI grew 4.5 times to $450 million.
Menlo estimates specialized pharma AI spending reached $700 million; $600 million of external drug-discovery spending excludes drug assets.
Healthcare’s AI spending is moving beyond paperwork to tools that assist clinicians, serve patients and support drug research. Menlo Ventures’ 2026 State of AI in Healthcare report, published November 16, estimates spending grew 2.5 times to $3.7 billion. Established software vendors captured most new spending on AI-generated clinical notes.
Healthcare Dive reports that the study polled more than 550 U.S. healthcare and life sciences executives. Menlo’s spending figures are market estimates; adoption findings and forecasts reflect executives’ responses. Drug-development expectations of faster results remain distinct from demonstrated medical advances.
Healthcare also buys a different mix of AI than other industries. Menlo says industry-specific applications receive 50% of healthcare AI spending, compared with 24% across enterprises overall. The firm points to regulation, isolated datasets and difficult connections to existing systems as reasons specialized products have a stronger foothold.
Ambient scribes, which turn clinical conversations into notes, show how quickly the market can change. Menlo estimates the category reached $950 million in 2026, up from $600 million in 2025. Every major electronic health record vendor now offers a built-in scribe, according to the report.
Led by Microsoft’s Dragon Copilot and Epic’s Art, incumbent offerings captured 68% of new scribing dollars. Established vendors held 46% of the overall market. That is a gain in momentum, not yet a majority of total spending: startup scribes still grew 28%, even as incumbent offerings grew 120%.
Menlo links the shift to incumbents’ distribution and long-term platform relationships as note-taking becomes less distinctive. Providers’ procurement cycles have lengthened to 10–11 months, from five–seven months. The market is growing, but suppliers face a tougher sales environment—not a simple collapse in demand for startup products.
Providers account for $2.9 billion, or 77%, of healthcare AI spending, by Menlo’s estimate. Administrative tools still command the largest budgets, while categories closer to patient care are growing faster. Two areas stand out:
Buyers’ preferences vary sharply by task. More than two-thirds of providers preferred keeping coding, billing and chart review within their electronic health record system. Only 16% said the same for patient navigation. That contrast leaves a different competitive opening from scribing: providers are not equally committed to their existing vendor across every workflow.
Menlo estimates pharmaceutical spending on specialized AI rose from $30 million to $700 million in 2026. Within that broader figure, external spending on drug-discovery AI totaled $600 million. That narrower measure includes models, software platforms, tools for coordinating AI work and data, but excludes drug assets themselves.
Deployment is further along in pharma than biotech: 51% of pharmaceutical companies were running AI at scale, compared with 35% of biotechnology companies. The spending increase therefore accompanies a reported shift into production, not just a larger collection of experiments.
The ambitions extend well beyond software adoption. Among surveyed pharma research and development leaders, 84% expect AI to remove at least a year from early drug development over the next three to five years. Those are expectations, not established time savings.
Even the respondents’ own timelines put medical validation further out: 93% expect the first approval of AI-discovered drugs within five years, but only 28% expect it within two. Deployment and spending are advancing faster than that anticipated approval milestone.
Loading discussion...
Join the conversation
Explain when you would favor a specialist instead.
Be the first to share a perspective or an experience.
Reader comments
Newest comments first. Replies stay oldest first.