The Strategic Reconfiguration of Healthcare AI Business Models: ARR, Tokenised Consumption and the Value Based Transition

Nelson Advisors
1 hour ago
11 min read

The global healthcare technology ecosystem is currently undergoing a structural transformation that industry analysts have categorised as the transition from Health Tech 1.0 to Health Tech 2.0. This evolution is not merely technological but fundamentally commercial, representing a shift from speculative, hype-driven growth to a rigorous, margin-centric paradigm defined by the "Health AI X factor".

As of 2025 and moving into 2026, the industry is witnessing a decoupling of healthcare growth from traditional labour constraints, driven by a new generation of AI-native companies that reach significant revenue milestones with unprecedented velocity. While the previous generation of healthcare software providers often required over a decade to achieve $100 Million in Annual Recurring Revenue (ARR), contemporary AI-native firms are hitting $200 Million ARR in under five years.

This acceleration is underpinned by a radical shift in labour productivity, where AI-native healthcare entities generate between $500,000 and $1 Million in ARR per full-time equivalent (FTE), compared to the $100,000 to $200,000 typically seen in traditional services.

The Macro-Economic Foundations of Health Tech 2.0

The reopening of the IPO window in 2024 and 2025 marked a turning point for the sector, adding approximately $36.6 Billion in fresh market capitalisation. These newly public entities differ from their predecessors in their emphasis on sustainable growth and free cash flow (FCF) margins. Elite performers in the 2025 market cycle demonstrated a "Rule of 40" performance, calculated by the sum of annualised revenue growth and LTM FCF margin.

This financial rigour has led to a divergence in market performance; the Bessemer Health Tech Index rose 18% in 2025, matching the S&P 500 and outperforming the broader emerging cloud indices, while the "Health Tech 1.0" index, comprising companies that went public before 2022 remained essentially flat.

The valuation gap between traditional healthcare and healthcare SaaS reflects these differing economic engines. Traditional healthcare valuations are predominantly shaped by operational efficiency and regulatory factors, often measured via EV/EBITDA multiples, whereas healthcare SaaS prioritises recurring revenue, scalability, and net revenue retention (NRR).

Aspect	Traditional Healthcare Focus	Healthcare SaaS / AI Focus
Revenue Model	Fee-for-service / Reimbursements	Subscription / Token-based Recurring
Primary Valuation Multiple	8x - 15x EV/EBITDA	6x - 20x EV/ARR
Margin Profile	Operating Margin (15% - 25%)	Gross Margin (70% - 85%)
Growth Metric	Patient Volume (5% - 15% YoY)	Net Revenue Retention (95% - 140%)
Risk Factors	Regulatory compliance / Payor changes	Churn rate / Tech obsolescence / Hallucinations

In the private markets, deal-making remained steady in 2025 with 527 venture deals totalling an estimated $14 Billion. A critical trend within this deployment is the concentration of capital into AI-native firms, which captured 55% of all health tech funding in 2025, up from 29% in 2022.

Furthermore, for every dollar invested in the broader AI landscape, $0.22 was deployed specifically to healthcare AI startups, signalling investor confidence that AI will capture a disproportionate share of value in the coming years.

The Three Pillars of Modern Monetisation: ARR, Fixed Fees and Tokens

The commercial architecture of healthcare AI is currently organised around three distinct but often overlapping pillars: Annual Recurring Revenue (ARR) through subscriptions, fixed fees for implementation and professional services and token-based or usage-based metered consumption.

The Subscription Engine: ARR and the SaaS Hierarchy

Annual Recurring Revenue remains the gold standard for valuation in the healthcare technology sector. It provides the financial predictability necessary for accurate long-term planning, hiring and product development. However, as AI integration becomes more intensive, the standard seat-based model is being refined. Finance leaders now categorise revenue into a hierarchy to maintain valuation integrity. Subscription revenue sits at the top of this hierarchy, while variable and service revenue are treated with greater scrutiny by investors.

A well-structured P&L for a 2026 healthcare AI firm must distinguish between these streams to avoid "margin leakage." Subscription revenue typically carries gross margins above 70%, while professional services often operate at much lower margins. If services revenue becomes too high a percentage of the total, the business risks being valued as a consulting agency rather than a high-multiple SaaS company.

The Fixed Fee Paradigm: Implementation and EHR Integration.

Fixed fees play a critical role in the enterprise healthcare sales cycle, particularly regarding implementation and data migration. The cost of implementing an Electronic Health Record (EHR) system or a large-scale AI platform is not merely financial but includes significant "indirect" costs such as staff training and workflow realignment. For large clinics and hospitals, ready-made EHR solutions can cost millions in upfront fees, while smaller practices might spend approximately $400,000 on average.

The Token Factory and Usage Based Economics

The most significant shift in the 2025-2026 timeframe is the emergence of what Jensen Huang described at GTC 2026 as the "Token Factory" era. This model meters AI consumption through tokens, which represent units of text or data processed by large language models (LLMs). Token-based pricing is ideal for infrastructure and developer tools where usage varies significantly.

However, pure token pricing can create budget unpredictability for non-technical buyers, leading to "bill shock" if interactions exceed projections.

To mitigate this, hybrid models have become the most popular monetisation strategy in 2026, with 85% of SaaS leaders adopting some form of usage-based or metered pricing. These hybrid models typically combine a base subscription fee (for predictability) with usage-based charges or credit systems (for scalability).

The Transformation of Unit Economics: Revenue per FTE

The shift from manual, service-heavy workflows to AI-augmented processes has fundamentally altered the unit economics of healthcare. Traditional healthcare services are labour-intensive, resulting in low revenue per employee. AI-native companies leverage "agentic AI" to handle repetitive administrative and clinical tasks, allowing human clinicians to focus on high-value decision-making.

Historical data on ARR per FTE illustrates this jump:

Traditional Healthcare Services: $100,000 - $200,000
Pre-AI Healthcare SaaS: $200,000 - $400,000
AI-Native Healthcare: $500,000 - $1,000,000+

This increased productivity translates directly into higher valuations. Companies demonstrating high ARR per FTE are perceived as more scalable and less susceptible to the labor shortages and burnout currently plaguing the medical profession. For example, AI ambient scribes have been shown to reduce "pajama time" (after-hours documentation) by up to 60% at the University of Vermont Health Network, effectively increasing the "billable capacity" of every clinician on the platform.

Hospital Financial Strategy: Navigating the CapEx to OpEx Migration

The adoption of AI and cloud-based technologies is forcing hospital CFOs to navigate a complex migration from Capital Expenditure (CapEx) to Operational Expenditure (OpEx). Traditionally, hospitals favored CapEx because it allowed them to own physical assets and depreciate them over many years. However, the rapid evolution of AI technology means that hardware-based AI infrastructure often becomes obsolete before it can be fully depreciated.

Factor	CapEx (Capital Expenditure)	OpEx (Operational Expenditure)
Upfront Cost	High initial investment	Lower upfront costs
Cash Flow Impact	Large immediate outflow	Predictable recurring payments
Tax Treatment	Depreciated over asset life	Immediate tax deduction
Balance Sheet	Appears as asset	Hits operating budget immediately
Agility	Less flexible; manual upgrades	More flexible; easier to scale

Many hospitals are now freezing new IT CapEx in favour of OpEx models that align costs with usage and outcomes. Shifting an EHR or AI suite to a cloud model allows a hospital to shrink its CapEx line by up to half, although this transition requires careful management of the profit and loss statement, as OpEx costs hit the operating budget immediately. To defend these shifts to the C-suite, IT leaders are focusing on "Total Cost of Ownership" (TCO) analyses that quantify the savings from reduced downtime, enhanced security, and avoided hardware refreshes.

Regulatory Unlocks and the Reimbursement Landscape

In 2026, the primary "unlock" for healthcare AI is the emergence of direct reimbursement pathways. For years, AI was considered a cost centre; however, the launch of clinical AI payment codes by CMS (Centers for Medicare & Medicaid Services) has turned AI into a billable service.

CPT Taxonomy: Assistive, Augmentative, and Autonomous AI

The American Medical Association (AMA) introduced a taxonomy in 2022 to classify AI medical services, which has been fully implemented in the 2025-2026 CPT code sets. This classification determines the level of clinical responsibility the AI assumes:

Assistive: AI detects data that might be missed by a physician but does not perform primary analysis.
Augmentative: AI analyses and quantifies data to produce clinically meaningful insights, supporting the physician's work.
Autonomous: AI makes clinical decisions without immediate human intervention.

As of January 2026, there are 26 active CPT codes for clinical AI solutions. While many are Category III (temporary codes for emerging technologies), three have been upgraded to Category I (permanent reimbursement). These include FFR-CT for cardiovascular risk prediction and automated detection of diabetic retinopathy.

The Challenge of Carrier Pricing

Despite the growth in CPT codes, reimbursement remains fragmented due to "carrier pricing". Because CMS often declines to set national Relative Value Units (RVUs) for new AI software, it relies on Medicare Administrative Contractors (MACs) to determine pricing regionally. This results in significant price variation; an AI service might be reimbursed at a level that supports profitability in one region while being financially unsustainable in another.Furthermore, CMS still uses Physician Practice Information (PPI) survey data from 2007-2008 to allocate indirect costs, which predates modern software-as-a-service models, creating disputes over whether "per-click" licensing fees should be treated as direct or indirect costs.

Specialised Domain Economics: Radiology, Diagnostics, and Remote Monitoring

The commercial models for healthcare AI vary significantly by clinical domain, with radiology and remote monitoring serving as the most mature markets.

Radiology and Diagnostic AI: The Value of "Incidental" Discovery

Radiology AI demonstrates value by increasing diagnostic accuracy and facilitating the detection of incidental findings that might otherwise lead to downstream medical emergencies. These tools are often billed on a per-use basis via CPT codes or New Technology Add-on Payments (NTAPs) in inpatient settings. The economic value of AI in radiology is frequently tied to the reduction in "Length of Stay" (LOS). With a national average cost of $\$2,500$ per hospital bed-day, even a minimal reduction in LOS across a patient population can yield massive annualised savings.

To refine this value proposition, researchers have proposed an "AI Score" model for reimbursement. This framework links payment to the complexity of the data processed and the scope of automation provided.

AI Score Dimension	High Complexity / High Value	Low Complexity / Low Value
Data Diversity	Multi-center, multi-population data	Homogeneous, single-center data
Multimodality	EHR + Genomic + Imaging	Single modality (e.g., text only)
Clinical Impact	Direct diagnosis of life-threatening events	Basic administrative summarization
Human Oversight	Highly autonomous (level 4-5)	Purely assistive (level 1-2)

Remote Patient Monitoring (RPM) and Therapeutic Monitoring (RTM)

Remote monitoring has emerged as a high-margin opportunity for providers, with CMS allowing compensation of over $120 per patient per month for Medicare patients. A practice managing $100 RPM patients can generate over $140,000 in annual revenue without significant additional infrastructure.

For vendors, this has created a sustainable "revenue-share" or "zero-upfront" model. Companies like Accuhealth and Optimize Health provide devices (BP monitors, glucose meters, scales) at no upfront cost to the provider, taking a percentage of the reimbursed CPT codes in exchange for handling logistics, billing, and 24/7 monitoring services.

New rules proposed for 2026 aim to simplify this further by relaxing the "16-day data rule," allowing clinics to bill even if patients send data for fewer days, and introducing new codes for shorter-duration monitoring.

Value-Based Care and Gain-Sharing: The Future of Stakeholder Alignment

The ultimate goal of many healthcare systems is to move entirely away from volume-based fee-for-service models toward Value-Based Care (VBC). In these models, clinicians are paid for keeping patients healthy and achieving better long-term outcomes.

Gain-Sharing Models

Gain-sharing refers specifically to direct payments by hospitals to physicians based on reducing hospital costs while meeting quality standards. Unlike shared savings, which impacts revenue from insurers, gain-sharing focuses on lowering internal costs on inpatient services. This model is lower risk for providers because incentive payments are paid out of actual cost reductions already achieved. For instance, the New Jersey Medicare Gainsharing Demonstration project effectively decreased inpatient costs by 8.5% over three years.

In the context of AI, gain-sharing is increasingly common in "agentic" contracts, where vendors are paid a percentage of the savings they generate. A mid-tier US carrier recently used AI to migrate a closed block of insurance and launch a new product in just 10 weeks, reducing the cost of migration by 20-40%. By pooling operations across multiple clients, these strategic partnerships can reduce total operating expenses by 45-65% over the life of a block.

Global Commercial Models: The UK NHS and Centralized Commissioning

The United Kingdom's NHS provides a different commercial lens, moving toward centralised digitisation and national funding mandates.

The MedTech Funding Mandate (MTFM)

The MTFM policy was established to accelerate the uptake of proven, cost-saving technologies by removing local financial barriers. Technologies must be recommended by NICE, be cost-saving within three years, and be "affordable" (national budget impact under £20 million). Once a technology is selected, NHS commissioners and providers are mandated to agree on local funding arrangements. This ensures equitable access across England, preventing the "postcode lottery" often seen in fragmented systems.

The AI Diagnostic Fund and NHS Online

The NHS AI Diagnostic Fund has already achieved significant milestones, with $100\%$ of stroke units in England using AI for scan analysis and 50% of hospital trusts deploying AI for lung cancer diagnosis. Looking toward 2027, the NHS aims to establish "NHS Online," an "online hospital" connecting patients to expert clinicians and utilizing AI-powered tools for health advice and triage. This digital transformation is supported by a £10 Billion commitment through 2028-29, representing a 50% increase from previous levels.

Strategic Risks and Competitive Moats

As the market matures, the nature of competition in healthcare AI is shifting. Startups face three primary hurdles to scalability: eroding competitive moats, regulatory complexity, and clinician trust.

The Data Ownership Moat

The "Health AI X factor" is not just about the model but about data ownership. As general-purpose LLMs from horizontal players (OpenAI, Google, AWS) commoditise basic AI tasks, the value shifts to companies that own longitudinal patient data and can underwrite risk. Startups that lack enterprise pilots or access to proprietary data silos are finding it increasingly difficult to compete with incumbents like Epic or Cerner, who are building AI directly into their core platforms.

The EU AI Act and Global Regulation

Regulatory frameworks are becoming more stringent, with the EU AI Act classifying most healthcare AI as "high-risk".This necessitates rigorous conformity assessments and post-market surveillance, which can add significant operational overhead to smaller ventures. In the US, the MHRA and other bodies are rewriting the regulatory rulebook to allow for quicker access to AI assistants while maintaining patient safety.

Clinician Disengagement and Burnout

The success of any commercial model in healthcare depends on clinician adoption. Developing AI tools without clinician input has led to a trust deficit, with 80% of AI tools lacking prospective validation. To mitigate this, successful vendors are establishing clinician advisory boards and focusing on "explainable AI" (XAI) to ensure that the AI's reasoning is transparent and actionable for the medical professional.

Conclusions: The Future of Commercial Alignment in Health AI

The commercial future of healthcare AI and technology is defined by a convergence of metered consumption (tokens), recurring stability (ARR), and outcome-aligned incentives (Gain-sharing/VBC). The transition to Health Tech 2.0 has moved the industry past the era of speculative investment toward a period where financial success is tied to measurable improvements in clinical productivity and patient outcomes.

The decoupling of healthcare growth from labor through AI-native "FTE productivity" is the most significant economic shift of the decade. For providers, the shift from CapEx to OpEx enables the rapid adoption of these tools, provided they can navigate the fragmented reimbursement landscape of CPT codes and carrier pricing.

For vendors, the focus must remain on building "specialist" agents for high-value verticals, securing proprietary data moats, and aligning pricing models with the actual value delivered to the system, whether that is measured in bed-days saved, incidental findings discovered, or "pajama time" reclaimed.

In 2026, the question is no longer whether AI will transform healthcare, but which commercial architectures will most effectively capture the value of that transformation.

Nelson Advisors > European MedTech and HealthTech Investment Banking

Nelson Advisors specialise in Mergers and Acquisitions, Partnerships and Investments for Digital Health, HealthTech, Health IT, Consumer HealthTech, Healthcare Cybersecurity, Healthcare AI companies. www.nelsonadvisors.co.uk

Nelson Advisors regularly publish Thought Leadership articles covering market insights, trends, analysis & predictions @ https://www.healthcare.digital

Nelson Advisors publish Europe’s leading HealthTech and MedTech M&A Newsletter every week, subscribe today! https://lnkd.in/e5hTp_xb

Nelson Advisors pride ourselves on our DNA as ‘Founders advising Founders.’ We partner with entrepreneurs, boards and investors to maximise shareholder value and investment returns. www.nelsonadvisors.co.uk

#NelsonAdvisors #HealthTech #DigitalHealth #HealthIT #Cybersecurity #HealthcareAI #ConsumerHealthTech #Mergers #Acquisitions #Partnerships #Growth #Strategy #NHS #UK #Europe #USA #VentureCapital #PrivateEquity #Founders #SeriesA #SeriesB #Founders #SellSide #TechAssets #Fundraising #BuildBuyPartner #GoToMarket #PharmaTech #BioTech #Genomics #MedTech