$110 Million Pearl Health Raise Powers AI Platform as CMS Rewrites ACO Rules

Two Triggers: Capital and Rule Changes

Pearl Health closed $110 million in July, including $50 million in equity led by Andreessen Horowitz with Viking Global Investors, AlleyCorp, and Ulysses Capital, plus a $60 million credit facility, to expand an AI platform that helps independent primary-care physicians manage Medicare risk. In the same month, the Centers for Medicare & Medicaid Services published proposed rules reshaping the Medicare Shared Savings Program, adjusting the ACO REACH model for its 2026 performance year, and opening a pathway for ACOs to reduce or eliminate beneficiary cost-sharing starting April 2027. The funding round priced the platform just as the federal rulebook is being rewritten to make value-based risk both more accessible and more unavoidable for physicians who have so far stayed on the sidelines.

Pearl was founded in 2020 on the premise that independent physicians — not health systems, not private-equity roll-ups — are the right risk-bearers for Medicare's shift to value-based care, provided they get the data infrastructure and actuarial support that large organizations build for themselves. As of the July announcement, the company supports more than 10,000 providers across over 40 states, caring for upwards of 250,000 Medicare beneficiaries and managing approximately $3.6 billion, Pearl Health's July press release reported, in annualized medical spend, up from $2.4 billion, the release's data shows, the prior year and $1.6 billion, the same release found, the year before that. The company reached profitability in 2025, a milestone its investors note is rare in this space at this growth rate, and projects $500 million, the company's July figures put projected savings at, in gross healthcare system savings while tripling its patient base through the end of 2026. The new capital is earmarked for advancing the AI platform, specifically the "Performance Intelligence" layer that surfaces real-time insights on total cost of care, quality, and utilization, and the "Care Orchestration" agents designed to automate annual wellness visit scheduling, post-discharge follow-ups, and care management outreach, as well as for expanding into Medicare Advantage and deepening enterprise health system and payer partnerships.

More than 70 million people rely on Medicare, with program costs exceeding $1 trillion, according to Pearl Health's July press release, and climbing. Reimbursement is increasingly tied to outcomes rather than utilization, creating what Pearl's CEO Michael Kopko describes as bonuses for keeping patients out of hospitals. But the infrastructure to act on those incentives, including predictive risk modeling, FHIR-enabled data pipelines, actuarial pricing of clinical interventions, and automated care orchestration, is beyond the reach of most independent practices. That is the gap Pearl and its competitors are racing to fill. The July funding round and the July rule proposals are not separate stories; they are the capital and regulatory sides of the same market formation. The independent PCP who does not plug into an AI-enabled platform in the next 18–24 months faces a binary outcome: join an ACO backed by a technology partner, or be consolidated by a larger risk-bearing entity that already has one.

The Doctor Who Decides Admission

The independent primary-care physician is the only actor in Medicare who actually decides whether a patient goes to the hospital — and the only one who eats the cost when that decision is wrong. Yet the payment rails still treat them like volume widgets. As of 2021, only 4.4 percent of traditional Medicare dollars and 2.5 percent of commercial reimbursement flowed through capitated, population-based alternative payment models; 17 percent of Medicare Advantage did. The rest, nearly 40 percent of all insurer payments as recently as 2018, remained pure fee-for-service with no link to quality or value. A PCP who wants to keep a frail diabetic out of the ER gets paid the same whether the patient stabilizes or codes. The incentive to invest in prevention simply does not exist on the dominant rails.

CMS has spent a decade trying to change that. The Medicare Shared Savings Program, launched under the ACA, grew to roughly 500 ACOs serving over 11 million beneficiaries by 2020 — about one in five Medicare enrollees. But the majority of MSSP contracts are still one-sided: the ACO shares savings if costs come in below benchmark, but owes nothing if they exceed it. Downside risk, where the provider also covers losses, climbed from under 10 percent of MSSP participants in 2017 to 37 percent in 2020 and 41 percent in 2021, even as total ACO enrollment and attributed lives declined. CMS wants two-sided risk. The market is not volunteering for it.

Why not? Start with the math. A typical independent practice lacks the actuarial capacity to price its own risk, the data infrastructure to track total cost of care across hospitals, specialists, and post-acute settings, and the clinical workflows to intervene before a $50,000 admission becomes a $500,000 ICU stay. Brent Davis, who ran a 330-provider network across roughly 100 sites in central Florida, described the subsidy required to operate in a fee-for-service market as "eight figures — a very large subsidy." His hospitals and physicians "don't get rich off of Medicare." When that system tried to take on risk, the margin for error was razor-thin. "Risk cuts both ways," Davis said. "We were mindful to keep that from increasing."

The administrative burden compounds the problem. MSSP and the newer LEAD model require whole-TIN participation, meaning every clinician under a single tax ID must be in the ACO. But billing TINs rarely map to clinical relationships. A cardiologist who sees a patient once a year gets bundled into the same risk pool as the PCP who manages that patient's diabetes, hypertension, and heart failure every month. The PCP bears the clinical accountability; the specialist bears none. MIPS reporting burdens inflate further. The LDI white paper called the VBP landscape's complexity "a significant barrier to participating in APMs and to evaluating each model."

Physician-led ACOs have historically outperformed hospital-led ones on savings, largely by reducing hospitalizations, and those savings grow the longer the ACO stays in the program. Hospital-led ACOs face a structural conflict: every avoided admission is lost revenue. The independent PCP has no such conflict. They are the natural risk-bearing entity. But they cannot build the infrastructure alone. They need actuarial modeling to set benchmarks, FHIR-based data pipes to ingest claims and clinical feeds in near-real time, predictive models to flag rising-risk patients before they hit the ED, and workflow automation to turn those flags into outreach, scheduling, and documentation without hiring an army of care managers.

Pearl's platform was built for exactly that gap: individual physician exposure to financial risk rather than insulation from it; technology infrastructure to support fundamentally new workflows; a marketplace to create risk liquidity anchored on physician decision-making. Its AI platform identifies patients who could benefit from attention before costs rise, suggests next steps for care teams, and automates those administrative actions.

The independent PCP is the demand side of this market. They hold the clinical leverage. They bear the financial exposure. And they are the ones CMS is pushing through rule changes, benchmark reforms, and the coming LEAD transition into two-sided risk whether they have the infrastructure or not. The platforms that can deliver that infrastructure at scale will lock in the physicians. The ones that can't will watch them consolidate into health systems or private-equity roll-ups that can.

CMS Rewrites the Rulebook

CMS has moved three levers at once: a finalized physician fee schedule, an updated ACO REACH methodology, and a July proposed rule that sketches the next generation of accountable care. Together they widen the on-ramp for independent primary-care physicians and raise the cost of staying on the sidelines.

The CY2026 Physician Fee Schedule final rule, published October 31, 2025, carries the most immediate changes. CMS cut the maximum time an ACO can remain in a one-sided BASIC track from seven performance years to five, effective for agreement periods starting January 1, 2027. The agency said the change is "intended to encourage participation in two-sided risk models" — a direct nudge toward the downside exposure that platforms like Pearl are built to manage. At the same time, CMS relaxed the 5,000-beneficiary minimum for benchmark years, adding safeguards against normal expenditure variation while giving smaller practices a plausible entry point. The rule also stripped the health equity adjustment from the quality score beginning in performance year 2026, calling it duplicative after prior changes to the Complex Organization Adjustment and eCQM reporting incentives. Quality reporting got lighter in other ways: the Medicare CQM beneficiary definition was revised so the eligible population overlaps more cleanly with assigned beneficiaries, reducing patient-matching burden; the CAHPS survey moves to a web-mail-phone protocol in 2027; and Quality ID 487 (Screening for Social Drivers of Health) was dropped from the APP Plus measure set. CMS also extended extreme-and-uncontrollable-circumstances protections to cyberattacks and ransomware for PY2025 forward, and required mid-year reporting of ownership changes, both practical concessions to the operating reality of independent groups.

ACO REACH, the model designed for newer entrants and higher-risk populations, entered its 2026 performance year with 74 ACOs covering 1.7 million Traditional Medicare beneficiaries and 125,909 participating providers. CMS updated the model's financial methodology in 2025 to "ensure future cost savings," projecting lower net spending for 2026 without disrupting care. But REACH is time-limited: the Long-term Enhanced ACO Design (LEAD) model launches in 2027 at REACH's conclusion, explicitly targeting "smaller, independent or rural-based practices" and those with specialized patient populations. That transition creates a policy bridge — physicians who join REACH-aligned platforms now position themselves for LEAD's broader risk tracks without a cold start.

The July 2026 proposed rule, released July 14 and elaborated in a July 16 CMS blog, signals where the next fee-schedule cycle is headed. CMS floated giving approved Shared Savings Program ACOs the option to do so for certain Part B services starting April 2027 — a direct financial incentive for patients to stay in-network. It proposed higher payment for "quarterback" clinicians who coordinate longitudinal care, acknowledging the workflow burden that drives independent doctors toward employment. The agency also opened comment on prospective primary care payment, outcomes-based models, and a benchmark adjustment for ACOs that bring in new providers and beneficiaries — a growth lever that rewards recruitment. Two structural fixes drew particular attention: a "ratchet effect" proposal to soften rebasing penalties that punish previously successful ACOs, and guardrails for the Accountable Care Prospective Trend to stabilize benchmarking against system-wide cost growth. CMS also committed to transitioning quality reporting to digital quality measures across programs and sought input on electronic prior authorization requirements.

The participation data confirms the tailwind. As of January 2025, 477 Shared Savings Program ACOs served 11.2 million beneficiaries through 650,000-plus providers. By performance year 2026, the count rose to 511 ACOs, with 72 new and 62 renewing, covering 12.6 million beneficiaries, a 12.3 percent jump and the largest population ever. Over 82 percent of those ACOs now sit in Level E of the BASIC track or the ENHANCED track, both qualifying as Advanced APMs under the Quality Payment Program, the highest share since the program began in 2012. In PY2024, the most recent reconciled year, SSP ACOs generated $4.1 billion in shared savings and $2.5 billion in net Medicare savings. Total Medicare beneficiaries in accountable care arrangements reached 14.3 million as of January 2026, up 4.4 percent year over year.

The rule arc is coherent: lower the barrier to entry, compress the timeline to two-sided risk, subsidize the coordination work that independent practices struggle to fund alone, and build a glide path from REACH to LEAD that keeps the door open for smaller groups. For a platform betting on AI-enabled risk management, the policy window is not theoretical — it is codified in final rules, measured in beneficiary counts, and backed by a proposed rule that doubles down on the same direction.

Who Else Is Chasing These Doctors?

Pearl's $110 million raise lands in a market that is already crowded and consolidating fast. The independent primary-care physicians Pearl targets are being pursued by three distinct buyer types: established technology enablement platforms, health-system-owned networks, and private-equity roll-ups. Each has a different playbook, and each is moving to lock in practices before the CMS rule changes that take effect in 2026 and 2027 make the switch harder to reverse.

The enablement-platform tier is the most direct competitive set. Innovaccer, founded in 2014, is the category heavyweight.

In the past year Innovaccer acquired CaduceusHealth (May 2026) to add full-stack revenue-cycle management for ambulatory care and Story Health (September 2025) to deepen specialty-care coordination. It also announced a multi-year AWS collaboration to scale agentic AI and launched "Gravity," a healthcare intelligence layer. The company's own tear sheet lists ten direct competitors, including Navina, Smile Digital Health, Datavant, Lifen, Lumeris, Health Catalyst, Arcadia, Apixio, Verily, and IQVIA, signaling how fragmented the vendor map remains. Navina, for its part, is sponsoring AMGA webinars on "fragmented to connected" clinical workflows and has built an AI copilot that surfaces diagnostic gaps at the point of care. Health Catalyst and Arcadia have long histories selling population-health analytics to health systems and ACOs; Lumeris operates its own value-based care enablement arm. Pearl's bet is that a lighter, physician-first platform, with profitability achieved in 2025 per the company, can out-execute the enterprise sales cycles of these incumbents.

Health-system-owned networks represent a second front. Sutter Health outsourced its entire revenue-cycle operation to R1 RCM in a deal that transitioned roughly 1,150 RCM employees, about 10 percent of Sutter's physician base, to the vendor. That model locks affiliated practices into the system's chosen stack. Meanwhile, Humana and Providence launched a FHIR-based data exchange in October 2025 targeting Medicare Advantage members, automating patient-data flow for value-based contracts. The partnership is explicitly designed to "enable providers to deliver more effective care and helping our members spend less time on paperwork," per Humana's announcement. For an independent PCP, joining a health-system ACO often means adopting the system's EHR, analytics, and care-management workflows — a package that can feel like employment by another name. Pearl's pitch is that its platform works across any EHR and any ACO model (MSSP, ACO REACH, LEAD, Medicare Advantage), letting the practice stay independent while still accessing the data infrastructure CMS increasingly demands.

The third front is private-equity roll-ups. Oak Street Health, backed by General Atlantic and others, built 169 centers in 21 states before CVS acquired it for $10 billion in 2023; CVS has since signaled further expansion. The KFF analysis of CMS data shows 99 percent of Medicare Advantage enrollees now sit in plans with prior-authorization requirements — a lever PE-backed groups use to justify centralized utilization management. Research from the Leonard Davis Institute and PMC confirms that private equity and hospital employment are steadily reducing the pool of truly independent practices. A 2019 Health Affairs study found a 4-percentage-point increase in large practices (50+ physicians) in counties with high ACO penetration, and a 2026 Sage Journals paper noted that prioritizing independent smaller practices for ACO engagement yields smaller clinician-level ACO participation gains — suggesting the market naturally funnels toward scale. Pearl's that projection through 2026 and its push into Medicare Advantage and new risk offerings are explicitly aimed at offering an alternative: the data and actuarial muscle of a roll-up without the ownership change.

As of January 2025, the Shared Savings Program had 477 ACOs; by that performance year, the total reached 511, including 72 new and 62 renewing ACOs. The proposed July 2026 reforms, including LEAD launching after ACO REACH ends in 2026, expanded the addressable market further. Every platform, system, and roll-up is now racing to be the default "yes" when an independent PCP asks, "Who helps me take risk without selling my practice?" Pearl's raise buys runway to make that answer its own.

The Talent Scramble: Actuaries Meet Machine Learning

Pearl Health's $110 million raise landed in July 2026 with a specific mandate: triple the patient base from 2024 levels through year-end 2026 while pushing gross system savings toward $500 million. That scale does not run on spreadsheets. It requires a three-legged talent stool: risk actuaries who can price Medicare benchmarks and model shared-savings mechanics, machine-learning engineers who can turn claims and clinical feeds into prospective risk scores, and FHIR-fluent data engineers who can stitch together fragmented payer and provider systems in near real time.

The Society of Actuaries has been explicit about the shift. In its 2026 guidance on provider use of AI, the SOA wrote that actuaries "can play a critical role in validating these models, ensuring fairness, and aligning them with business objectives in value-based care arrangements," adding that as AI systems grow more autonomous, actuaries "would significantly benefit from expanding their influence beyond traditional forecasting and cost modeling." Arbital Health, a competing enablement platform, has already branded the fusion: it markets "Actuarial AI" that "immediately increase[s] productivity and help[s] manage risk in real time" by unifying claims, benchmark, utilization, and contract data with predictive models. The job market is pricing that fusion at a premium. Pearl's own board listings show a Lead Actuary, Valuation role posted at $95,000–$180,000 (New York, Boston, or remote) — a band that sits well above traditional insurance actuarial entry points and signals the scarcity of professionals who understand both CMS risk-adjustment methodology and gradient-boosted tree ensembles.

On the ML side, the volume is staggering. As of late 2026, LinkedIn listed 11,000-plus "Machine Learning Engineer and Healthcare" roles in the United States; Indeed showed 868 dedicated "Machine Learning Engineer in Healthcare" postings and another 418 broader "Machine Learning Healthcare" openings. But the VBC enablement niche demands a narrower profile: engineers who have built features on top of 837/835 transaction sets, who know the difference between HCC and CDPS risk models, and who can deploy inference pipelines that satisfy HIPAA and CMS audit trails without a six-month compliance review. That intersection is thin. Pearl's February 2026 update noted the company is "advancing development of Care Orchestration AI agents to further automate administrative workflows such as those workflows" — work that sits squarely at the junction of FHIR-based interoperability, clinical NLP, and reinforcement-learning-driven workflow optimization.

FHIR data engineers are the connective tissue. The 21st Century Cures Act final rule and CMS's advancing interoperability mandates have made FHIR R4 the de facto exchange standard for claims, clinical, and enrollment data. But "FHIR compliance" in a slide deck is not the same as normalizing a firehose of Blue Button 2.0 exports, payer-specific extensions, and legacy CCDA feeds into a feature store that an actuary can trust for PY2026 benchmark certification. Platforms that solve that plumbing problem and hire the engineers who have done it lock in provider networks faster. The talent pool is currently split: health-system IT shops with Epic/Cerner depth but no risk-modeling mandate, and insurtech veterans with actuarial rigor but thin clinical-data chops. The VBC enablement platforms are bidding for the overlap.

For frontier-tech talent, including robotics, autonomy, and simulation, the signal is clear. Modern ML toolchains and the same production pressures (latency SLAs, drift monitoring, regulatory evidence packages) now appear in a market backed by $1 trillion in Medicare spend and a CMS rulebook that is actively expanding the addressable population. The hiring wave is not theoretical. Pearl added two roles in a single week in July 2026, including the Lead Actuary slot. Competitors from Innovaccer to Navina to health-plan-owned arms are posting similar blends. The consolidation of independent primary care, from 76 percent physician ownership in 1983 down to 51 percent by 2014, has created a buyer's market for platforms; the platforms are now fighting for the builders who can make the risk math work in production.

What Pearl Must Still Prove

Pearl Health's press releases and founder interviews make a clean causal chain: AI-driven risk prediction → earlier intervention → lower total cost of care → shared savings for independent PCPs. The $110 million raise bets that chain holds at scale. But the public record leaves three links unproven.

First, Pearl's own shared-savings performance. The company says it manages roughly that amount across more than 250,000 Medicare beneficiaries and projects that amount by the end of 2026. Those are forward-looking figures. CMS publishes Performance Year Financial and Quality Results for every MSSP ACO, but Pearl has not released a consolidated, independently verified ledger showing how much of its managed spend actually converted to shared savings versus how much was benchmark movement or risk-score inflation. Without that, the $500 million projection is a model output, not an outcome.

Second, risk selection. The academic literature on value-based payment is explicit: "adverse selection, cherry picking, cream skimming, and patient dumping… has been found in a variety of contexts related to quality reporting or pay for performance" (JAMA Network), and "risk selection is a multi-dimensional phenomenon that occurs at the patient, hospital, and system level, and is a source of inefficiency and inequality in healthcare" (PubMed). Pearl's platform assigns beneficiaries to participating PCPs through the MSSP attribution algorithm; it does not control which patients walk through the door. If the platform's predictive models mainly identify high-cost patients who are already coded aggressively, or if participating practices quietly steer healthier seniors into the ACO, the savings signal confounds selection with intervention. CMS's quality gate (the 40th-percentile MIPS score of 73.85 for PY 2026) screens for process measures, not for case-mix neutrality. No public audit has tested whether Pearl's attributed population is risk-equivalent to the benchmark.

Third, the AI attribution gap. CEO Mike Kopko told Columbia Business School in February 2026 that "almost everything in senior care is predictable and therefore preventable" and that AI can "automate the actions required to prevent those potential issues occurring." The July 2026 raise announcement describes "Performance Intelligence" chat-enabled insights and "Care Orchestration AI agents" for wellness-visit scheduling and post-discharge follow-up. Those are product claims. The research contains no peer-reviewed study, no CMS evaluation report, and no third-party actuarial opinion linking Pearl's specific model outputs to marginal reductions in avoidable utilization. CMS's own Shared Savings Program data shows ACOs overall hit record shared-savings rates in PY 2024 while maintaining quality scores — but the program-wide trend conflates maturation, benchmark resets, and the 2026 rule changes that ease entry. Isolating Pearl's platform effect would require a counterfactual the company has not published.

When an independent PCP signs a participation agreement with Pearl, what audited evidence tells her that the platform, not the benchmark, not the attribution, not the cohort, will deliver the shared savings she's betting her practice on? The CMS rulebook has opened the door. The capital has arrived. The talent war is on. But the receipt that proves the platform, not the policy tailwind, is the engine of savings has not yet been printed.

Working in frontier tech? Zero G Talent tracks the openings: see every open Pearl Health role, browse frontier tech jobs, the companies hiring, and the people building the field.