Town Hall Ventures’ 2026 Predictions
By: Andy Slavitt, Josh Loria, & Andie Steinberg
Last year, we published What THV Sees in 2025—predicting a set of transformational changes for the healthcare system. 2025 largely lived up to that thesis, marked by significant transformation across AI adoption, CMMI activity, and a renewed focus on outcomes and prevention. What’s clear is that many of the forces we flagged have only intensified heading into 2026.
We’re publishing our 2026 outlook with a view of the structural forces shaping the healthcare system over the next year—where policy and technology continue to act as the twin engines of change. Our views are informed by sustained, candid conversations we had across CMS leadership, payors, providers, founders, and technologists. In 2026, we look forward to collaborating across the healthcare ecosystem as these dynamics take shape.
1. ChatGPT becomes the patient front door—and a referral engine for healthcare
We believe consumer-facing LLMs—such as ChatGPT and Claude—will emerge as a major distribution channel for digital health companies, and healthcare more broadly, creating a new growth pathway beyond traditional referrals from health systems, payors, providers, and direct-to-consumer marketing.
The behavioral shift is already underway. For many consumers, conversational AI has become the default starting point for health questions. Some needs will be handled directly by the LLMs, but many won’t. The LLMs will need trusted clinical partners that can accept warm handoffs based on clinical need, geography, and insurance coverage.
That dynamic creates a strategic opening. We expect platforms like OpenAI and Anthropic to integrate referral pathways directly into their interfaces, relying on vetted partners for specific use cases. Over time, this becomes a new distribution channel—AI-mediated referrals—where routing happens natively within the conversation and is supported by API-based intake, eligibility checks, and deep-linked scheduling. For example, a user might ask: “I’m 34, live in Florida, have Medicaid, and I’m feeling depressed”, and receive a direct link within the response to a trusted covered behavioral health provider’s intake flow—reducing friction for patients while increasing qualified referrals for providers.
In this world, advantage accrues to companies that are “AI-referral ready”. That readiness looks less like marketing and more like infrastructure: accurate coverage matching, frictionless intake, clinical appropriateness, trust, and robust APIs for connectivity.
2. Drug costs stay center stage
In 2026, prescription drug costs will remain one of the most persistent and consequential focal points in healthcare policy, with pressure coming simultaneously from regulation, market dynamics, and the innovation pipeline.
Policy momentum is reinforcing this trend. Additional rounds of IRA drug price negotiation are rolling out, underscoring that price compression is not a one-time intervention but a sustained direction of travel. The most recent batch—covering 10 additional drugs with reported average price reductions of ~35%—signals bipartisan durability. In parallel, international developments point to early signs of “global cost burden sharing” for pharmaceutical innovation. Reports that the UK is moving to increase net prices for new medicines (~25%+ increase) highlight growing recognition that the U.S. has long borne a disproportionate share of global pharmaceutical R&D costs—reinforcing MFN-style policy thinking around global cost burden sharing. Expect more announcements like this.
Cost pressures are intensifying as the FDA pipeline tilts toward high-cost modalities like cell and gene therapies (C>s), and as GLP-1s demonstrate substantial clinical benefit across expanding indications. As discussed later in Prediction 7, GLP-1s are especially influential because they move outcomes policymakers care about—such as A1c, BMI, and cardiometabolic risk—which are directly rewarded in models like ACCESS.
We expect continued growth in specialty drug carve-outs, including for GLP-1s, alongside increasing consumer and employer interest in alternative access pathways, including cash-pay, as stakeholders search for relief from rising pharmacy spend. Direct-to-consumer and cash-pay pharmacy models are expanding access and reshaping patient expectations, particularly as AI-enabled prescribing and clinical decision support move closer to the point of care (see Prediction 9).
Pharmacy is no longer a downstream function; it is increasingly central to how care is accessed, navigated, and paid for.
3. Medicare changes risk-adjustment again… beginning in 2028
We predict CMS will announce, by the end of 2026, a structural shift away from documentation-driven coding intensity (RAF) and toward a new CMS-standardized inferred risk aided by AI. Under an inferred-risk approach, expected cost would be derived statistically at the population level using claims data, pharmacy utilization, and ideally deeper clinical signals—rather than relying solely on annually documented diagnoses. This transition won't occur in 2026, but we expect the direction to be clearly signaled, with 2027 likely the last year of relative stability under the current model before a new model is launched in 2028.
Billions of dollars of revenue will shift in ways that are deliberately harder to game—and that’s the point. When inference is applied consistently across the population, individual coding advantages will matter far less. Artificial variance driven by who can “out-code” whom flattens, and in the best case, this approach is harder to game, better aligned with policy goals around choice, competition, and protecting taxpayer dollars, and reduces the burden on clinicians.
4. Medicare Advantage shifts to become the most stable risk pool in a destabilized system
After upheaval in the Medicare market over the last few years, Medicare Advantage (MA) is stabilizing just as other government risk pools begin to wobble. Medicaid and the ACA Exchanges face growing volatility as OBBBA-era redeterminations accelerate churn and shift acuity, straining state budgets and MCO economics. At the same time, affordability pressures—especially around ACA subsidy expiration—add uncertainty in Exchange enrollment and risk composition.
By contrast, MA is entering a period of relative clarity. 2026 marks the first year that the disruptive effects of the V28 risk model phase-in have fully worked through plan economics, and, as we had predicted, CMS finalized a 5% average increase in MA payments—more than $25 billion above 2025 levels, materially higher than recent years. Looking ahead, we expect upward rate pressure to continue into 2027, with CMS likely approving an even higher payment increase as benchmarks are further aligned with sustained utilization and unit cost trends.
The levers needed for performance are changing. As rates improve and the V28 uncertainty fades, accountability rises: plans will be expected to demonstrate real health improvement through care management, clinical pharmacy integration, access expansion, home-based touchpoints, and consistent operational execution—not just documentation excellence.
5. Medicaid enters an actuarial reckoning
Medicaid is entering a period of acute uncertainty over the next three years, driven in large part by OBBBA-era dynamics. Work requirements and their downstream effects will reshape risk pools in ways that existing rate-setting frameworks will have a difficult time keeping up with. As acuity shifts and membership turns over more rapidly, actuarial soundness and rate adequacy move to the center of the debate.
State rate-setting methodologies are increasingly struggling to keep pace. Variation across actuarial firms and states compounds the problem: some rely on short lookbacks that amplify volatility, while others use multi-year windows that lag rapidly changing risk. With CMS largely limited to approving or rejecting state submissions, actuarial approaches at the state level must price for the changes in the underlying risk pool.
Plans that cannot secure adequate rates can pull back, reduce benefits, or tighten networks if competitive dynamics allow. Providers will feel the pressure through lower reimbursement and higher bad debt. States will face an increasingly difficult tradeoff between near-term budget control and long-term program viability.
In 2026, the advantage accrues to organizations that can work closely with states to navigate this volatility and provide credible actuarial analytics.
6. Sub-capitation scales from the new ACO flagship model (LEAD)
From Pioneer to NextGen to REACH, CMS’s flagship ACO model now evolves again: ACO REACH sunsets in 2026 and will be replaced by LEAD—a structural redesign with many benefits, not a retreat from two-sided risk.
The most consequential change, and your new acronym of the year, may be CARA (CMS Administered Risk Arrangements). Under REACH, downstream risk transfer largely failed—fewer than ~4% of payments flowed to specialists under capitated or episode-based arrangements, leaving most care delivery effectively fee-for-service despite nominal two-sided risk. CMS aims to address this by providing CMS-managed data sharing, standardized episode definitions, configurable episode design, and CMS-administered payment mechanics.
LEAD also materially de-risks participation. The model introduces 10-year benchmarks with no rebasing, improved high-needs benchmarking, and prospective infrastructure payments for rural providers, alongside lower alignment thresholds for new entrants. These changes directly address the volatility and cash-flow concerns that historically kept rural, independent, and safety-net providers out of ACO models.
The implication is a step-function change in how population risk is executed. Sub-capitation starts to become a reality rather than an experiment. The winners will be those that can align high quality specialists, manage episode economics, and deploy technology that supports attribution, performance tracking, and reconciliation at scale.
7. Medicare’s first AI-native CMMI model (ACCESS) sets a new template for chronic care management and boosts GLP-1 use
CMMI’s ACCESS model (Advancing Chronic Care with Effective, Scalable Solutions) marks Medicare’s acknowledgment that chronic care is best managed through continuous, technology-enabled support—not episodic visits to a doctor’s office. Over the past several years, healthcare AI innovation has accelerated, with companies deploying agents that coach patients, support adherence, navigate care, and meaningfully augment clinicians. The persistent constraint has been payment: no broadly applicable CPT code has mapped cleanly to these services.
ACCESS begins to resolve that mismatch. Like RPM and CCM a decade ago, it creates a scalable reimbursement pathway for technology-enabled care—but with a critical distinction: payment is aligned to outcomes, not activity or time logged.
ACCESS puts pressure on legacy RPM and CCM structures, which we expect to gradually phase out as policymakers favor models that reward measurable improvement over utilization. If the economics prove durable, ACCESS could open a clearer path toward outcomes-aligned PMPMs that support care models built around coaching, navigation, adherence, and longitudinal engagement. For providers, ACCESS creates an economic incentive to partner with digital capabilities such as behavioral health integration, digital therapeutics, remote monitoring, and between-visit engagement.
ACCESS KPIs such as A1c or BMI are directly influenced by GLP-1 use, and behavioral health measures rely on patient self-reporting. At the same time, ACCESS introduces a clear forcing mechanism through the Substitute Spend Adjustment, which disallows duplicative services—such as RPM or CCM for the same condition—billed elsewhere during the care period. This signals CMS’s intent for ACCESS to replace, not layer on top of, legacy volume-based chronic care programs as Medicare shifts toward reimbursing outcomes rather than process.
8. The Administration’s prevention agenda shifts Stars from process measures to outcomes
In our 2025 outlook, we argued that healthcare would move from measuring activity to measuring improvement. CMS’s recent CY2027 proposed rule confirmed our prediction from last year. The proposal would remove 12 Stars measures focused primarily on administrative processes, with the explicit goal of rewarding outcomes, not process measures.
These changes are incremental, but their significance lies in what they signal. We believe they foreshadow a significantly larger shift toward embedding outcomes and prevention directly into Stars in a way that materially changes plan behavior. The MAHA framing reinforces this direction by elevating prevention and measurable health improvement as policy priorities. While CMMI’s MAHA Elevate initiative is directionally aligned, its limited scale—$3 million in grants to a small number of organizations—makes it unlikely to drive population-level change on its own. Stars, by contrast, is one of the few levers powerful enough to operationalize prevention across Medicare.
9. Administrative AI consolidates—clinical care is the next frontier
AI is the most consequential healthcare technology development—even including the EHR. 2025 marked a year of rapid adoption, with much of that uptake resulting in a payor vs. provider “arms race”. Buyers piloted point solutions quickly, often outside formal RFPs or a unified platform strategy. Open budgets, accessible APIs, and abundant venture capital made this an unusually favorable environment for healthcare AI startups. Higher-quality AI platforms will emerge—tools that reduce friction rather than add to it—as consolidation gives rise to mega-AI companies such as Thoreau.
At the same time, AI’s center of gravity will shift into clinical care. Beyond administrative automation, 2026 will see AI directly augment clinicians—supporting assessment, decision-making, prescribing, and longitudinal care amid workforce shortages. The market will move from primarily diagnostic use cases toward therapeutic and workflow-embedded systems of action, including mental and behavioral health agents, adherence support, and care delivery.
In an increasingly crowded marketplace, distribution will matter as much as product in determining market winners.
10. ACA subsidy expiration and employer cost pressure put coverage affordability back on the ballot for midterms
We expect healthcare coverage affordability to become a material issue in the 2026 midterm elections. Market forces across the ACA Exchanges and employer-sponsored insurance now converge on tens of millions of voters, pushing healthcare cost and access back onto the kitchen-table agenda. Rising premiums, higher deductibles, and visible coverage instability elevate healthcare from a policy debate to a lived economic concern heading into the midterm cycle.
At the center of this shift is the expiration of enhanced ACA subsidies. Early indicators suggest growing price sensitivity and churn risk. These dynamics point to a pronounced mix shift toward Bronze plans, a sicker risk pool, rising underinsurance, and eventual increases in uninsurance.
As people churn off Medicaid or Marketplace coverage, they reappear in emergency departments and hospital balance sheets. Uncompensated care and bad debt will rise, particularly in states hardest hit by redeterminations and subsidy roll-off. Health systems will seek relief through higher commercial rate negotiations. Layer in continued growth in specialty drug spending (ex: GLP1s and C>s per Prediction 2), and employer healthcare cost trendlines will continue to climb. With roughly 150 million Americans covered through employer-sponsored insurance, employers feel the heat in a way that bites into wages—giving this dynamic real political weight heading into the midterms.
As we predicted last year, ICHRA adoption remains modest (~500k lives to date) and we don’t expect that to change in 2026.
At Town Hall Ventures, we view this as a period of structural reset. These are moments when enduring healthcare companies are built. Medicare is demanding outcomes. Employers are demanding relief. AI is full of promise. Consumers are demanding affordability that actually works. We’re excited to partner with the founders and operators prepared to meet this moment and help build a healthcare system that delivers better outcomes and durable value.
We can be reached at steinberg@townhallventures.com.
