The Supply Crisis: Why AI is Healthcare's Only Way Out

Farzad Soleimani MD, MBA, FAAEM Partner, Healthcare

For decades, the healthcare debate has centered on cost. But cost is a symptom. The root cause is supply: we have built a system structurally incapable of meeting the demand placed upon it. AI is the only technology that can close that gap at scale. Fix the supply problem and everything else follows. Our strategy is shaped by the following seven beliefs.

— — —

Healthcare's Cost Crisis Is Fundamentally a Supply Crisis

Healthcare is the largest industry in the United States. We spend $5.3 trillion, nearly 19% of GDP, on care today. That number climbs to $10 trillion by 2036. Yet, 100 million Americans have no access to a primary care provider. The average wait for a new patient appointment is 26 days, up 44% since 2004, and 44% of currently practicing physicians will reach retirement age within a decade. When care is inaccessible, patients delay. Conditions manageable in a clinic become emergencies, far harder and more expensive to treat. This is where the cost crisis originates. It is not a pricing problem or an insurance problem. It is a supply problem expressing itself downstream as cost.

Sources: AAMC 2023 Physician Workforce Report; HRSA Health Workforce projections; Merritt Hawkins Wait Time Survey 2022; AMA Physician Practice Benchmark Survey.

This reframe has direct investment implications. Solutions that target cost alone attack the symptom while leaving the underlying problem intact. The genuine opportunity is in technologies that expand the effective supply of care: allowing one clinician to do the work of many, and enabling the system to serve patients who today have no one to see them.

The companies we have backed are already demonstrating this at scale. Fay Nutrition uses AI to multiply the reach of dietitians across a covered-benefit patient population, putting clinical-grade nutritional care within reach of patients who would otherwise never access it. Robot Health puts a continuous clinical presence in the home without adding a single caregiver, monitoring patients around the clock and surfacing actionable insights to clinicians in real time. Both companies expand the effective supply of care without waiting for the provider shortage to be solved.

— — —

Clinical AI Is Expanding the Effective Supply of Every Clinician

For years, the 60% of total healthcare spending that goes to clinician fees was essentially untouchable. Technology could optimize the margins around care delivery but could not penetrate the clinical encounter itself. That has changed. Clinical AI now functions like a senior resident: interviewing the patient, synthesizing data, developing the differential, and packaging everything before a certified clinician makes the final call. This is where we see the biggest opportunity in healthcare.

Source: CMS National Health Expenditure Accounts 2023.

Contrary to intuition, AI makes care more personal, not less. When AI absorbs the documentation, data gathering, and prior authorization, what remains is the conversation, the relationship, and clinical judgment applied with full attention. The visit becomes more human precisely because the machine has handled everything else.

Future Clinic is the clearest expression of this in our portfolio. It handles the clinical encounter 99% of the way before a physician verifies. That is not a reduction in care quality. It is a reallocation of the most valuable resource in medicine toward the work only a physician can do.

THE EVIDENCE BASE: AI VS. PHYSICIAN PERFORMANCE

LLMs are outperforming clinicians across structured clinical benchmarks, and consumers are rapidly adopting AI for health needs. The data below shows both dimensions of the shift.

In 2024, 16% of Americans used AI chatbots for clinical questions. By 2025 that figure doubled to 32%, with 64% of those users engaging weekly or more.

— — —

Healthcare's Biggest Failures Are Coordination Failures

Many of healthcare's biggest failures are not clinical but coordination failures. Consider a cancer patient. The delay between first symptom and confirmed diagnosis is rarely caused by a lack of specialists or technology. It results from countless handoffs between primary care, imaging, specialists, pathology, and treatment teams. Healthcare has become a coordination problem, and coordination at scale is something machines are uniquely suited to solve.