Vi Labs

1.1 Identity and Business Model

Vi, the brand for Vi Labs, is an enterprise artificial-intelligence platform built for healthcare, life sciences, and wellness organizations and operated from New York. The company describes itself as "productized AI for health enterprises" and positions its software as the "AI execution layer" that sits on top of customers' existing systems of record. Its commercial proposition is to convert a large proprietary dataset and domain-specific models into measurable patient and operational return on investment across the enterprise value chain, from patient engagement through operational efficiency. The platform is organized around five named modules — Data Web (the data layer), Activate, Engage, Operate, and Pulse — and in May 2026 Vi layered a suite of vertically specialized AI agents across the Activate, Engage, and Operate applications to drive next-best actions for patients, care teams, and operational workflows. Revenue is generated through enterprise relationships with health systems, pharmaceutical and life-sciences companies, and wellness organizations rather than a direct-to-consumer model, though Vi does not publicly disclose its pricing structure or contract economics.[CO001, CO021, CO010, CO026, CO023, CO035]

1.2 Leadership, Founders, and Governance

Vi's founder and chief executive officer is Omri Yoffe, who is credited in executive interviews and company databases with building Vi around the integration of artificial intelligence and data science into healthcare and life-sciences outcomes. The company's origin story is unusual: Vi describes itself as "born from aerospace," inspired by a silent systems failure that killed a co-founder's Air Force pilot friend, and frames its technology as a "neural radar for health" that surfaces critical signals buried in noise. Beyond the chief executive, public-facing pages identify a functional leadership bench including Product Lead Yiftach Meitar, Senior Director of Client Performance Chelsea Pincus, VP of Growth for Healthcare Davis Miller, and Client Strategy and Activation Lead Laurén DiVenere, while a partner interview references a chief revenue officer named Spencer. Board composition, equity holdings, and formal governance arrangements are not disclosed, so key-person dependence on Yoffe and the completeness of the executive roster cannot be fully verified from public sources. This leadership opacity is a recurring theme across the company-overview evidence base and is preserved as a gap.[CO009, CO027, CO028, CO037, CO013]

1.3 Funding, Valuation, and Investors

In May 2026 Vi announced it had completed a $145 million transaction valuing the company at $1.64 billion. The company states the deal combined both primary and secondary capital and was aimed at supporting talent retention and acquisition, investing in the platform and new products, and strengthening the balance sheet. The press release, reprinted across PRNewswire and Morningstar and covered independently by Pulse 2.0, lists Vi's shareholders as General Atlantic, Revelstoke, 1902 Capital (managed by The Pritzker Organization), Square Peg, Savano Capital, Island Green, and others; Revelstoke Capital separately lists Vi Labs in its investment portfolio. The most recent prior financing visible in databases is a $131 million later-stage venture round that PitchBook records as completed on 28 June 2024, alongside a 2022 secondary transaction. Combining the disclosed 2024 round and the 2026 transaction implies on the order of $276 million of cumulative capital, although Vi has never published a complete round-by-round funding history, revenue, or margin, leaving the underlying financial trajectory that justifies the $1.64 billion mark unverifiable from public evidence alone.[CO002, CO003, CO008, CO018, CO019, CO020]

1.4 Scale and Traction Metrics

Vi's public scale claims are large and consistent across its press release and founder blog. The company states it serves more than 100 large-scale enterprise customers, including Fortune 500 companies and many leading healthcare, life-sciences, and wellness organizations. It says it supports more than 190 million patients and members, has contributed to the development and commercialization of more than 50 drugs across oncology, pediatric epilepsy, and diabetes, and has generated more than $2 billion in measurable value across its partner ecosystem. The differentiating asset underpinning these claims is the Data Web, which Vi describes as one of the world's largest datasets of clinical, behavioral, and operational signals integrating physiographic, demographic, usage, and licensed partner data into a single AI-ready layer; database descriptions also reference household-level data coverage. These figures are company-reported and not independently audited, and core financial metrics such as revenue, annual recurring revenue, gross margin, and net revenue retention remain undisclosed, so the quality and durability of the traction cannot yet be corroborated by third parties.[CO004, CO005, CO006, CO007, CO030, CO032]

1.5 Milestones, History, and Open Questions

Vi's documented history is uneven because the company has historically maintained a low public profile. Company branding references roots reaching back to roughly 2011, Crunchbase has associated the company with a mid-2010s founding, and PitchBook's first recorded institutional round is a 2018 later-stage venture financing — a genuine conflict in the founding chronology that diligence should resolve. The clearest dated milestones are financial and product events from 2024 onward: a $131 million round in June 2024, the May 2026 launch of the AI agent suite, and the simultaneous $145 million transaction at a $1.64 billion valuation. Vi has also published "State of AI" reports for healthcare, biopharma, and wellness in 2025 and 2026, and its privacy policy governs handling of de-identified data. PitchBook lists roughly 123 employees, a figure that sits in tension with the company's very large scale claims and reinforces the need to verify headcount, revenue, and the founding timeline directly. The Pulse module provides real-time analytics, and marketing partners cite a "4X ROI" figure that remains company-sourced and unverified.[CO012, CO014, CO016, CO017, CO029, CO031]

1.6 Exhibits

2.1 Market Boundary and Substitutes

Vi sits inside the artificial-intelligence-in-healthcare market but is best understood through a narrower boundary: enterprise AI and data platforms that serve healthcare providers, life-sciences and pharmaceutical companies, and wellness organizations. Included spend covers healthcare data platforms and analytics, patient engagement and activation software, population-health and care-navigation tools, and pharma commercialization and clinical-trial analytics. Excluded from Vi's core boundary are clinical-grade diagnostic AI and medical-imaging devices, robot-assisted surgery, electronic health record systems of record, and consumer wellness apps sold direct to individuals, even though some of these are counted in the broadest AI-in-healthcare market totals. The status-quo substitutes Vi displaces are in-house data-science teams, point solutions stitched together by systems integrators, legacy analytics suites, and manual care-management and commercialization workflows. Vi positions itself as a horizontal "AI execution layer" that orchestrates these functions on top of existing systems of record, which both widens its addressable footprint and exposes it to competition from incumbents in each adjacent category. Defining this boundary before sizing is essential because headline AI-in-healthcare totals overstate the spend Vi can realistically capture.[CM001, CM010, CM011, CM012, CM013, CM030]

2.2 Market Sizing Across Multiple Lenses

Independent market researchers disagree materially on the size and growth of AI in healthcare, which is why this report preserves a range rather than a single figure. Grand View Research estimates the global AI-in-healthcare market at $36.67 billion in 2025, growing at a 38.9% CAGR to $505.59 billion by 2033. Precedence Research puts 2025 revenue at $36.96 billion and projects $744.34 billion by 2035 at a 35.0% CAGR. Mordor Intelligence sizes 2025 at $40.14 billion and forecasts $251.36 billion by 2031 at a 36.2% CAGR, while Fortune Business Insights is the most aggressive, projecting growth from $39.34 billion in 2025 to $1.03 trillion by 2034 at a 44.0% CAGR. North America consistently accounts for the largest share — roughly 44-54% — and software is the dominant component at roughly 46% of spend. Applying a top-down haircut to isolate the enterprise healthcare/life-sciences data-platform layer that Vi actually sells into yields a serviceable market on the order of $8-18 billion in 2026, of which Vi's realistically obtainable share is a low-single-digit-billion SOM. These constrained estimates are derived, not source-stated, and are flagged as such.[CM001, CM002, CM003, CM004, CM005, CM006]

2.3 Buyers, Users, Payers, and Adoption Path

Vi sells to three distinct enterprise buyer types, each with a different budget owner and adoption trigger. Health systems and provider organizations buy population-health, patient-engagement, and care-navigation capabilities, typically sponsored by a chief medical, digital, or operating officer and justified by value-based-care economics and operational efficiency. Pharmaceutical and life-sciences companies buy drug-commercialization, patient-finding, and clinical-trial-acceleration analytics, with budget owned by commercial and medical-affairs leaders and triggered by launch ROI and trial timelines. Wellness and payer-adjacent enterprises buy member engagement and lifetime-value optimization. The user is usually an analyst, care manager, or commercial operator, while the economic payer is the enterprise itself rather than an individual patient. The adoption path runs from data integration through pilot, proof of ROI, departmental deployment, and finally enterprise standardization — a multi-quarter cycle gated by data-governance, security, and procurement reviews. Long enterprise sales cycles and the need to connect to existing systems of record are the principal friction points, and they materially affect how quickly Vi can convert its large addressable market into recognized revenue.[CM010, CM012, CM014, CM015, CM016, CM017]

2.4 Growth Drivers and Adoption Constraints

Several secular drivers expand Vi's market. The shift to value-based care, reinforced by CMS Innovation Center payment models, rewards organizations that can predict and manage population health, directly favoring data-and-analytics platforms. Federal interoperability rules advanced by the Office of the National Coordinator, including TEFCA and FHIR-based exchange, increase the availability and portability of clinical data that platforms like Vi's Data Web depend on. Enterprise AI adoption and the appetite for measurable ROI in a cost-pressured sector accelerate budgets toward vertically specialized tools. Against these tailwinds sit real constraints: data-privacy and HIPAA obligations raise the cost of handling protected health information; AI-specific regulatory uncertainty around clinical decision support creates caution; switching costs and integration complexity lengthen sales cycles; and a broader skepticism about whether near-term AI adoption can match the hype tempers buyer urgency. Some analysts also warn of crowding as incumbents and startups converge on the same enterprise budgets. The net effect is a large, fast-growing, but contested market where execution speed and trust, not just technology, determine which platforms capture the available spend.[CM018, CM019, CM020, CM021, CM022, CM023]

2.5 Exhibits

3.1 Competitive Landscape and Substitutes

The competitive set around Vi spans four layers. Direct peers are venture-backed healthcare data and population-health platforms — Komodo Health, Innovaccer, Arcadia, and Lightbeam — that, like Vi, aggregate clinical and claims data and sell analytics and engagement to providers, payers, and pharma. Incumbents are large public companies: Veeva Systems, the dominant life-sciences cloud and CRM vendor; IQVIA, the largest healthcare data and contract-research organization; and Health Catalyst, a provider-focused data-and-analytics company. Adjacent competitors include electronic health record vendors (Epic, Oracle Health) that increasingly add analytics, and the major cloud and horizontal AI providers whose general-purpose models Vi explicitly positions against. The most persistent substitute is the status quo: enterprises building data-science teams in-house or stitching together point solutions and systems integrators. Vi's strategic bet is that a vertically specialized, productized execution layer spanning provider, pharma, and wellness use cases will out-compete both single-vertical specialists and generic horizontal tools. That bet is credible but unproven, because most named rivals disclose revenue, customer counts, or public-market scale that Vi does not, making head-to-head comparison difficult and tilting near-term advantage toward incumbents with established distribution.[CP001, CP002, CP010, CP011, CP012, CP030]

3.2 Capability and Feature Comparison

On capability breadth, Vi's pitch is unusually horizontal: a single platform combining a large data layer (Data Web), patient activation and engagement, operational workflows, real-time analytics, and a new suite of AI agents spanning provider, pharma, and wellness use cases. Most competitors are deeper but narrower. Veeva is strongest in life-sciences commercial and clinical software and CRM but is not a provider-side population-health platform. IQVIA's advantage is the scale and depth of its real-world data and analytics for pharma, plus contract-research capabilities Vi does not offer. Komodo Health centers on its "Healthcare Map" of patient-level data for life sciences. Innovaccer, Arcadia, Health Catalyst, and Lightbeam are provider- and payer-focused population-health and data-platform companies with strong analytics but less emphasis on pharma commercialization. Where Vi can credibly claim an edge is the combination of breadth across all three verticals and an agent layer that drives next-best actions; where it is most exposed is depth and proof, since incumbents have larger datasets, deeper regulatory and compliance track records, and disclosed enterprise customer bases. The capability matrix should therefore be read as breadth-versus-depth rather than a simple feature checklist.[CP013, CP014, CP015, CP016, CP017, CP031]

3.3 Pricing, Distribution, and Switching Costs

None of these vendors, including Vi, publishes transparent list pricing; healthcare data and analytics platforms are sold as multi-year enterprise contracts negotiated per deployment, with value tied to data access, seats, and outcomes. Distribution power is a key differentiator. Veeva and IQVIA enjoy deep, sticky relationships across nearly the entire pharmaceutical industry, and their data and CRM systems carry high switching costs once embedded in regulated commercial and clinical workflows. Provider-focused platforms like Innovaccer, Arcadia, and Health Catalyst lock in health systems through data integration, custom analytics, and care-management workflows that are expensive to rip out. Komodo's moat is its proprietary patient-level Healthcare Map. Vi's switching costs are plausibly real once its Data Web is integrated and its agents are embedded in operational workflows, but with limited public evidence of contract length, retention, or multi-homing behavior, its lock-in is harder to verify than incumbents'. Buyers frequently multi-home — using one vendor for provider analytics and another for pharma commercialization — which both opens a wedge for Vi's cross-vertical breadth and limits any single vendor's ability to fully displace the others.[CP018, CP019, CP020, CP021, CP032]

3.4 Moat Durability and Competitive Risk

Vi's potential moats are its large proprietary Data Web, cross-vertical breadth, and an early agent layer, reinforced by blue-chip investors. Their durability is uncertain. The data moat is meaningful only if Vi's dataset is genuinely differentiated against IQVIA's and Komodo's at-scale real-world data assets, which is unproven from public evidence. The breadth advantage could become a focus disadvantage if specialists out-execute Vi in each vertical. Most importantly, the AI-agent layer that Vi is marketing as a differentiator is precisely the capability every competitor and the horizontal AI providers are racing to add, creating real commoditization risk. Incumbents also have structural advantages Vi lacks: public-company balance sheets, decades of regulatory and compliance track record, and entrenched distribution. The principal adverse scenarios are that incumbents bundle comparable AI agents into existing contracts at low marginal price, that Vi's undisclosed scale proves smaller than its narrative implies, and that long sales cycles let better-resourced rivals win the enterprise standard. These risks make competitive positioning the most consequential swing factor in underwriting Vi's $1.64 billion valuation.[CP022, CP023, CP024, CP025, CP033, CP034]

3.5 Exhibits

4.1 Revenue Streams and Monetization

Vi's revenue model is enterprise software monetization layered on a proprietary data asset. The primary stream is multi-year platform subscriptions to its modules — Data Web, Activate, Engage, Operate, and Pulse — sold to health systems, pharmaceutical companies, and wellness enterprises. A second stream is data licensing and access to its de-identified Data Web, a high-margin asset class comparable to how IQVIA and Komodo Health monetize real-world data. A third, newer stream is the AI-agent suite launched in May 2026, which Vi positions as an execution layer that can be sold as expansion on top of existing deployments. Professional services and implementation likely contribute lower-margin revenue, as is typical for enterprise health-IT vendors. Vi markets outcome-oriented value — more than $2 billion in measurable value delivered and roughly 4x return on investment for customers — which implies some willingness to anchor pricing to outcomes, though there is no public evidence of formal outcomes-based contracts. Critically, Vi discloses none of the revenue mix, recognition policies, or contract terms that would let an investor verify revenue quality. The composition described here is inferred from the company's product structure and sector norms, not from disclosed financial statements, and should be confirmed directly with management before underwriting.[CI001, CI002, CI003, CI004, CI020]

4.2 GTM Motion and Unit Economics

Vi sells through a direct enterprise motion to large healthcare, life-sciences, and wellness buyers, a model characterized by long, complex sales cycles, multi-stakeholder procurement, and security and compliance review. With more than 100 enterprise customers and a $1.64 billion valuation, implied average contract values are likely in the high-six- to seven-figure range, but Vi discloses neither average contract value, customer acquisition cost, payback period, nor net revenue retention. As a result, every unit-economics figure in this chapter is an estimate triangulated from public-company comparables: enterprise health-data SaaS peers such as Veeva operate at roughly 70-75% gross margins, while services-heavier analytics vendors such as Health Catalyst run lower. Vi's blended gross margin is plausibly in the 55-70% range depending on the mix of high-margin data licensing versus lower-margin services, but this is unverified. Sales efficiency is similarly opaque; the only proxies are the company's claimed ROI and value-delivered figures, which are marketing assertions rather than audited metrics. The absence of disclosed CAC, payback, and retention is the most important gap in assessing whether Vi's growth is efficient or capital-intensive, and it materially limits confidence in the business's underlying profitability and scalability.[CI005, CI006, CI007, CI008, CI021]

4.3 Cost Structure and Public Traction

Vi's cost structure is inferred rather than disclosed. As an AI-and-data platform, its largest costs are likely engineering and data-science talent, data acquisition and de-identification, cloud infrastructure to run models across 190 million-plus patient records, and an enterprise sales and customer-success organization. Its roughly 123-person headcount, per PitchBook, is small relative to public incumbents, suggesting either high revenue-per-employee efficiency or a smaller revenue base than the valuation implies — the two cannot be distinguished without disclosure. On public traction, the verifiable facts are limited to non-financial proof points: 100-plus enterprise customers, 190 million-plus de-identified records, 96% US-household coverage, 50-plus drugs supported to market, and the $2 billion-plus measurable-value claim. None of the financial traction metrics an investor would normally rely on — revenue, ARR, growth rate, gross margin, billings, or active-user economics — are public. This gap between strong operational proof points and absent financial disclosure is the defining feature of Vi's financial profile: the narrative is compelling, but the numbers that would confirm it are unavailable, leaving valuation to rest on scale claims and comparable-company multiples rather than on observed financial performance.[CI009, CI010, CI011, CI012, CI022, CI023]

4.4 Capital Adequacy and Financial Verdict

Vi's disclosed capitalization is its strongest financial signal. The company completed a $131 million later-stage round in June 2024 and a $145 million primary-and-secondary transaction in May 2026 at a $1.64 billion valuation, backed by blue-chip investors including General Atlantic and Revelstoke Capital. Because the 2026 transaction included a secondary component providing liquidity to existing shareholders, not all $145 million flowed to the balance sheet, so the precise cash added is unclear. Vi discloses neither cash on hand, burn rate, nor runway, so capital adequacy cannot be quantified; however, raising at a stepped-up valuation from premier investors signals access to capital and reduces near-term financing risk. The financial verdict is mixed. Revenue quality is plausibly high given the data-licensing component, the margin path is likely healthy by sector norms, and capital intensity appears moderate, but none of this is verified. The decisive diligence blocker is the complete absence of audited financials: without revenue, margin, retention, and cash data, the $1.64 billion valuation cannot be reconciled to fundamentals and must be underwritten on scale claims and comparables. Obtaining audited financials under NDA is the single most important condition for any investment.[CI013, CI014, CI015, CI016, CI017, CI024]

4.5 Exhibits

5.1 Product Definition and Module Map

In customer-workflow terms, Vi is the intelligence-and-execution layer that sits between an enterprise's systems of record and the decisions its teams need to make about patients, members, and markets. A health system uses Vi to identify and activate at-risk patients; a pharmaceutical company uses it to find and engage the right providers and patients for a therapy; a wellness enterprise uses it to personalize member engagement. These workflows are delivered through five named modules. Data Web is the data foundation — a de-identified dataset spanning more than 190 million patient records with roughly 96% US-household coverage. Activate drives population identification and patient activation. Engage powers digital and multi-channel engagement. Operate handles operational and care-coordination workflows. Pulse provides real-time analytics and insights. Layered on top is a suite of AI agents, launched in May 2026, that Vi describes as an execution layer automating next-best actions across these modules. The modules are sold as a connected platform rather than standalone tools, which is central to Vi's positioning: the value comes from combining a large proprietary dataset with applications that act on it. Because most product detail comes from Vi's own pages, technical depth and exact module boundaries should be confirmed in a product demonstration during diligence.[CE001, CE002, CE003, CE004, CE020]

5.2 Architecture and Operating Model

Vi's architecture can be read as a stack. At the base is the Data Web data layer: ingestion, de-identification, linkage, and storage of clinical, claims, and consumer signals across hundreds of millions of records. Above it sits a machine-learning and analytics layer that builds predictive and segmentation models on that data — the company's heritage is described as applying signal-processing and "neural" techniques, reflecting roots its leadership frames as coming from aerospace and radar. On top of the model layer is the newer agentic execution layer: AI agents that translate model outputs into recommended or automated actions. The five modules are the application surface that exposes these capabilities to specific user personas, and an integration layer connects the platform to enterprise systems of record. Interoperability with electronic health records and adherence to healthcare data standards such as HL7 FHIR are effectively prerequisites for deployment in provider environments, and US interoperability policy under ONC/HealthIT.gov is steadily raising the baseline for data exchange. Operationally, Vi runs as a multi-tenant cloud platform serving 100-plus enterprise customers, which implies meaningful investment in data engineering, model operations, and cloud infrastructure. The public record does not disclose Vi's specific cloud provider, model stack, or uptime metrics, so the architecture described here is reconstructed from the company's product pages and sector norms rather than from technical documentation.[CE005, CE006, CE007, CE008, CE021, CE022]

5.3 Deployment, Roadmap, and Differentiation

Deploying Vi into an enterprise involves connecting the customer's data and systems to the platform, configuring modules to the customer's workflows, and validating model outputs against the customer's population — a process that, like most enterprise health-IT, is integration-heavy and bounded by security and compliance review. Vi's reliability, support model, and formal SLAs are not publicly documented, a gap for diligence. On roadmap, the clearest signal is the May 2026 launch of the AI-agent suite, which extends the platform from analytics-and-engagement toward autonomous execution and is the company's primary stated direction; further roadmap detail is not public. Vi's differentiation rests on four pillars. First, data: the scale and household coverage of Data Web is a genuine asset if it is differentiated against IQVIA's and Komodo's data. Second, vertical specialization: Vi's agents and models are purpose-built for healthcare workflows rather than general-purpose. Third, breadth across provider, pharma, and wellness on one platform. Fourth, an early-mover position in agentic healthcare AI. The durability of these is uncertain: the data moat is unverified against incumbents, and the agent layer is the capability every competitor is racing to add, creating commoditization risk. The most important product-diligence questions are how defensible the data asset is and how rigorously the models and agents are validated for clinical safety.[CE009, CE010, CE011, CE012, CE023, CE024]

5.4 Trust, Security, Privacy, and Compliance

Because Vi processes protected health information at scale, trust, security, privacy, and compliance are core to the product, not peripheral. Vi's central privacy mechanism is de-identification: it describes its Data Web as composed of de-identified patient records, which under US law (HIPAA) can be used and shared more freely than identifiable PHI provided de-identification is done to standard. Vi publishes a privacy policy governing its data practices. Operating in healthcare also subjects Vi to HIPAA's security and privacy rules, evolving FDA thinking on AI/ML in healthcare, and FTC scrutiny of health data — areas covered in this report's risk chapter. For the product, the key trust questions are the robustness of de-identification (re-identification risk rises as datasets grow and link more signals), the governance and validation of AI models and agents to prevent unsafe or biased recommendations, and the security controls protecting a high-value health dataset from breach. Vi's public materials assert compliance and de-identification but do not publish independent audits (for example SOC 2 or HITRUST), model-validation methodologies, or breach history, so these must be verified directly. The combination of a large PHI-derived dataset and an action-taking agent layer raises the stakes: errors or breaches would carry regulatory, reputational, and clinical consequences, making demonstrable trust controls a gating condition for enterprise adoption and for investment.[CE013, CE014, CE015, CE016, CE025, CE026]

5.5 Exhibits

6.1 Customer Base and Segmentation

Vi segments its customer base primarily by vertical: health systems and providers, pharmaceutical and life-sciences companies, and wellness and consumer-health enterprises. Within each vertical the buyer and user profiles differ. In provider organizations, buyers are typically population-health, care-management, and clinical-operations leaders, with care teams as end users. In life sciences, buyers are commercial, medical-affairs, and patient-services leaders seeking to identify and engage providers and patients. In wellness, buyers are member-engagement and growth leaders. Geographically, Vi's footprint is anchored in the United States, consistent with a Data Web that claims roughly 96% US household coverage; international presence is not documented publicly. By size, the customers implied by a $1.64 billion valuation and enterprise pricing are large organizations rather than small clinics, though the named Minnesota mental-health customer shows Vi also serves mid-sized specialty providers. The use cases span patient identification and activation, provider and patient engagement for therapies, operational care coordination, and analytics. What Vi does not disclose is the distribution of its 100-plus customers across these segments, the revenue contribution of each, or any revenue-band breakdown, so the segmentation described here is qualitative and drawn from Vi's positioning and a single detailed case study rather than from a disclosed customer roster.[CU001, CU002, CU003, CU004, CU020]

6.2 Adoption and Deployment Trajectory

The clearest quantitative traction signal Vi provides is its claim of more than 100 enterprise customers, combined with platform-scale metrics: a Data Web spanning 190 million-plus de-identified records, roughly 96% US-household coverage, support for 50-plus drugs brought to market, and more than $2 billion in measurable value delivered. These figures imply meaningful adoption depth, but Vi does not publish a time series showing how customer count, active usage, or deployments have grown, nor does it disclose utilization, repeat-purchase, or expansion rates. The company's 2024 later-stage round and 2026 $145 million transaction at a stepped-up valuation are indirect evidence that investors saw continued commercial momentum, and the May 2026 AI-agent launch suggests an expansion motion into the existing base. The named Minnesota mental-health customer, which Vi says scaled care roughly fivefold, is the most concrete deployment proof and indicates the platform can drive large operational gains in a live production setting. Still, a single dated case study and aggregate counts are thin evidence of a durable, broad-based adoption trajectory. Without a customer-growth time series or cohort-level usage data, diligence cannot distinguish between a rapidly compounding base and a flatter one padded by early logos, making the adoption trajectory one of the most important things to verify directly with management and customer references.[CU005, CU006, CU007, CU008, CU021]

6.3 Named Customer Proof and Reference Quality

Vi's named-customer evidence is led by a detailed story about a Minnesota mental-health provider that Vi reports scaled its care delivery roughly fivefold using the platform — a production deployment with a specific, sizable outcome claim. Vi also markets aggregate proof points, including the 50-plus drugs supported to market and the $2 billion-plus measurable-value figure, and a partner case study cites roughly 4x customer ROI. The quality of this proof is mixed. The Minnesota story is a genuine, named, production reference with a concrete outcome, which is strong. But it is largely a single deep reference; Vi does not publish a broad logo wall, multiple independent named references, or third-party validated outcome studies, and the strongest numbers (value delivered, ROI) are company- or partner-sourced rather than independently audited. Evidence freshness is reasonable — the funding and agent-launch material is dated to 2026 and the case study is current — but the breadth and independence of references are limited. For an enterprise platform claiming 100-plus customers, the gap between the headline count and the small number of publicly verifiable, named, outcome-backed references is notable. Diligence should request a full customer list, multiple reference calls across verticals, and any third-party or peer-reviewed outcome evidence to corroborate the marketed value and ROI claims.[CU009, CU010, CU011, CU012, CU022, CU023]

6.4 Retention, Expansion, and Concentration Risk

Retention and durability are the weakest-evidenced parts of Vi's customer profile. Vi discloses no net revenue retention, gross retention, churn, renewal rate, contract length, or cohort-survival data — the metrics that would show whether customers stay and expand. For an enterprise platform with multi-year contracts and an embedded data layer, switching costs are plausibly high once deployed, which would support retention, and the 2026 AI-agent launch provides a natural expansion lever into the installed base via land-and-expand. But none of this is quantified publicly. Concentration risk is similarly opaque: Vi does not disclose customer concentration, so it is unknown whether a handful of large accounts drive a disproportionate share of revenue, which would be a material risk. Channel and partner dependence are also undocumented, though the direct-enterprise motion implies limited channel reliance. Procurement friction is a structural headwind — long, security- and compliance-heavy enterprise sales cycles in healthcare slow both new bookings and expansion. The net picture is that Vi's retention and concentration profile is a near-total disclosure gap layered on a plausibly sticky product. Because customer durability and concentration are central to underwriting recurring revenue quality, obtaining cohort retention, NRR, and top-customer concentration data is among the highest-priority diligence asks, and the absence of any adverse churn evidence should not be mistaken for confirmation of low churn.[CU013, CU014, CU015, CU016, CU024, CU025]

6.5 Exhibits

7.1 Regulatory and Legal Risk

Regulatory and legal risk is Vi's most material exposure because it sits at the center of US health-data regulation while extending into AI. The foundational regime is HIPAA, administered by HHS, which governs the use and disclosure of protected health information; Vi's reliance on de-identification is what allows broad data use, so any weakness in its de-identification methodology, or a regulatory tightening of the de-identification standard, would directly threaten its core data asset. Layered on top is the evolving regulatory treatment of AI in healthcare: the FDA's framework for AI/ML-enabled software as a medical device could, depending on how Vi's agents are used, bring some functionality into a regulated medical- device pathway, and the agency's posture is still maturing. The FTC has signaled aggressive enforcement on health-data privacy and the misuse of sensitive data, including against companies that share health information without adequate consent or disclosure. State privacy laws (such as California's CCPA/CPRA) and, for any international data, the EU's GDPR add further obligations, and the federal AI-policy landscape is unsettled. IP risk and litigation are not evidenced in the public record, but the absence of disclosed litigation is not proof of none. The net regulatory picture is high-likelihood, high-impact exposure that Vi asserts it manages through de-identification and a published privacy policy, but which it does not document with audits, attestations, or regulatory correspondence — leaving the most consequential risk category the least verifiable from public sources.[CR001, CR002, CR003, CR004, CR005, CR020]

7.2 Operational, Quality, and Security Risk

Vi's operational and security risks center on the same asset that is its strength: a large, sensitive, PHI-derived dataset feeding AI models that increasingly take or recommend actions. The most severe operational risk is a data breach. Healthcare is the most-breached sector in the US, and HHS maintains a public breach-reporting portal precisely because incidents are common and consequential; a breach of Vi's 190 million-plus-record Data Web would carry regulatory penalties, customer loss, and reputational damage. Closely related is re-identification risk: as de-identified datasets grow and link more signals, the theoretical ability to re-identify individuals rises, which is both a privacy and a compliance hazard. A second category is model quality and safety. Vi's agents produce recommendations that influence patient identification, engagement, and operations; errors, drift, or bias in those models could cause clinical, equity, or commercial harm, and Vi does not publish model-validation or bias-testing methodologies. A third is reliability: an action-taking platform embedded in enterprise workflows must meet stringent uptime and support expectations, yet Vi publishes no SLAs, uptime history, or incident record. Finally, security certifications (SOC 2, HITRUST) that enterprise buyers typically require are not publicly evidenced. Switching costs and Vi's track record of scale suggest competent operations, but the public record provides almost no verifiable operational-control evidence, so operational and security risk must be treated as material and largely unmitigated until proven otherwise.[CR006, CR007, CR008, CR009, CR010, CR021]

7.3 Partner and Dependency Risk

Vi's platform depends on a chain of external relationships, each a potential point of failure. The most fundamental is data supply: Vi's Data Web is assembled from clinical, claims, and consumer data sources, and the company's economics and product both depend on continued, compliant access to that data. If a key data-supply relationship were curtailed — by contract, by regulation, or by a partner building a competing product — Vi's core asset would erode. Second is cloud and infrastructure dependency: like all modern AI platforms, Vi relies on cloud providers and, for its agent layer, likely on third-party model and tooling providers, creating exposure to pricing, availability, and policy changes outside its control. Third is customer concentration: Vi does not disclose whether a small number of large accounts drive a disproportionate share of revenue, a classic dependency risk that cannot be assessed publicly. Fourth is capital-provider dependency: while Vi's blue-chip investors are a strength, a private company without disclosed profitability ultimately depends on continued investor support, and a tighter funding environment raises that risk. Fifth is regulatory dependency: Vi operates at the pleasure of a regulatory regime that could change the rules on de-identification or AI. The dependency map is wide and, because Vi discloses neither its data-supply contracts, cloud arrangements, nor customer concentration, largely unquantifiable — making third-party and concentration diligence a high priority.[CR011, CR012, CR013, CR014, CR023, CR024]

7.4 Financial, Model, and People Risk

Financial and execution risks compound the regulatory and operational picture. The defining financial risk is transparency: Vi discloses no revenue, margin, retention, burn, or runway, so an investor cannot verify the unit economics or capital adequacy underpinning a $1.64 billion valuation, and cannot rule out margin compression, high burn, or weak retention. Although the 2026 transaction included a secondary component and premier investors, the lack of disclosed profitability means financing dependency is real if growth slows. Model risk is a distinct financial-adjacent hazard: if Vi's AI agents are central to its value proposition and they underperform, are commoditized by incumbents bundling comparable agents, or require heavy ongoing investment to maintain, the economic thesis weakens. People and execution risk is meaningful in a company of roughly 123 employees: dependence on founder-CEO Omri Yoffe and a small senior team creates key-person risk, and scaling an enterprise healthcare-AI business requires deep regulatory, clinical, and commercial talent that is scarce and expensive. There is no public evidence of governance weaknesses or executive departures, but neither is there disclosed succession planning or board composition. Taken together, the financial, model, and people risks are individually manageable but collectively raise the bar on diligence: without financial disclosure and organizational depth evidence, the durability of both the business model and the team cannot be confirmed.[CR015, CR016, CR017, CR018, CR025, CR026]

7.5 Mitigations, Monitoring, and Kill Triggers

Vi's publicly evidenced mitigations are partial. On the regulatory and privacy front, it relies on de-identification and a published privacy policy; on operations, on the implicit competence of running a 100-plus-customer platform; on capital, on blue-chip investor backing. What is missing from the public record is the verifiable apparatus enterprise buyers and investors expect: independent security certifications, documented model governance and bias testing, breach and incident history, SLAs, and disclosed financial and concentration data. The most useful framing for an investor is a set of monitoring indicators and thesis-break triggers. Monitoring indicators include any enforcement action or inquiry from HHS/OCR or the FTC, any reported breach on the HHS portal, changes to the HIPAA de-identification standard or FDA AI/ML device policy, evidence of customer churn or concentration, and signs of agent commoditization by incumbents. Thesis-break (kill) triggers include a material data breach or regulatory enforcement action, a regulatory change that undermines de-identification, loss of a key data-supply relationship, evidence that revenue or retention is far below what the valuation implies, or departure of the founder-CEO without a credible successor. The corresponding diligence asks are direct: obtain audited financials, security and compliance attestations, model-validation documentation, data-supply and customer contracts, and litigation and incident history. Until these are satisfied, Vi's risk profile should be underwritten as high-uncertainty with several high-impact, poorly verified exposures.[CR016, CR019, CR027, CR028, CR029, CR030]

7.6 Exhibits

8.1 Investment Thesis and Anti-Thesis

The bull thesis for Vi rests on five reinforcing pillars established earlier in this report. Market: AI in healthcare is large and growing fast — multiple independent forecasts put the market in the tens of billions today, compounding at 30-40% annually — giving Vi a long runway. Product: Vi pairs a large proprietary Data Web (190 million-plus de-identified records, ~96% US-household coverage) with modules and a 2026 AI-agent layer, a coherent data-plus-action platform spanning provider, pharma, and wellness. Customers: 100-plus enterprise customers, a named production reference that scaled care roughly fivefold, and $2 billion-plus in claimed measurable value suggest real demand. Financials: enterprise subscriptions plus high-margin data licensing imply an attractive model, and premier investors validate access to capital. Competition: vertical specialization and breadth differentiate Vi from both single-vertical specialists and horizontal AI tools. The anti-thesis is equally structured. The financial story is unverifiable: no revenue, margin, retention, or cash disclosure. The data moat is unproven against IQVIA and Komodo. The AI-agent differentiator is the capability every incumbent is racing to commoditize. Regulatory and security exposure (HIPAA, FDA AI/ML, FTC, breach risk) is high-impact and poorly evidenced. And the company is small (~123 staff) with key-person dependence. The investment question is whether Vi's genuine data-and-distribution assets justify a $1.64 billion mark despite a near-total disclosure gap — a question that public evidence can frame but not resolve.[CV001, CV002, CV003, CV004, CV005, CV020]

8.2 Recommendation, Confidence, and Valuation Stance

Our recommendation is a qualified BUY with medium confidence, a medium-to-high risk rating, and a valuation stance of fair-to-slightly-rich at $1.64 billion. The rationale is that Vi combines a genuinely differentiated data asset, a credible product expansion into agentic execution, demonstrated enterprise demand, and top-tier investor backing — a combination that, if the undisclosed fundamentals are healthy, supports the current mark and meaningful upside. Confidence is capped at medium, not high, precisely because the fundamentals are undisclosed: an investor cannot verify revenue scale, growth, margins, or retention, so the recommendation is explicitly conditional. The valuation stance reflects that $1.64 billion is reasonable against private healthcare-AI comparables on a scale-and-narrative basis but cannot be anchored to a revenue multiple without disclosure; if Vi's revenue is at the low end of plausible ranges, the multiple would be rich. On entry discipline, a late-stage private round at a stepped-up valuation carries the usual considerations: liquidation-preference and dilution overhang are not publicly detailed and must be examined, and the secondary component of the 2026 deal means some capital funded shareholder liquidity rather than growth. The practical recommendation is to proceed to confirmatory diligence with a term sheet contingent on audited financials, retention and concentration data, and security and compliance attestations — buying the thesis but pricing and protecting against the disclosure gap through structure and conditions rather than taking the headline valuation at face value.[CV006, CV007, CV008, CV009, CV021, CV022]

8.3 Scenarios and Comparable Valuation

We frame three scenarios. In the bull case, Vi's data moat proves durable, the agent layer drives land-and-expand across its base, and it compounds toward category leadership; on a successful trajectory a future valuation of $4-6 billion-plus at exit is plausible, implying strong returns from a $1.64 billion entry. In the base case, Vi remains a strong specialized player with solid but not dominant share, exiting via strategic acquisition or IPO at roughly $2-3 billion — a moderate positive return. In the bear case, competition commoditizes the agent layer, a regulatory or security event impairs trust, or undisclosed financials prove weak, and the valuation stalls or contracts below the current mark, producing flat-to- negative returns. The comparable set anchors these. Public peers Veeva (tens of billions in market cap at high revenue multiples) and IQVIA (roughly $40 billion) set the ceiling for the category, while Health Catalyst (smaller, lower-multiple) shows the downside of a services-heavier model. The most relevant private comparables are Innovaccer and Komodo Health, each valued in the low-single-digit billions, which makes Vi's $1.64 billion mark broadly consistent with where venture-backed healthcare-data platforms have priced — provided Vi's scale is comparable. Because Vi discloses no revenue, the comparison is necessarily on valuation level and qualitative scale rather than on multiples, so the comparable analysis supports the reasonableness of the mark without confirming it. The decisive swing factor across all scenarios is whether the undisclosed financials validate the scale that the valuation and comparables assume.[CV010, CV011, CV012, CV013, CV023, CV024]

8.4 Exit Readiness, Triggers, and Final Diligence Asks

On exit readiness, Vi has the profile of a company that could exit by strategic acquisition or IPO within a multi-year horizon: strategic acquirers (large health-IT, pharma-services, or data companies) would value its dataset and cross-vertical footprint, and a continued strong AI-healthcare market would keep an IPO path open. But exit timing and valuation depend on the same undisclosed fundamentals, and a tighter funding or exit environment in 2026 adds uncertainty. The thesis-break (kill) triggers that should void the recommendation are concrete: discovery that revenue or retention is materially below what the valuation implies; a material data breach or regulatory enforcement action; a regulatory change undermining de-identification; loss of a key data-supply relationship; commoditization of the agent layer by incumbents bundling comparable capabilities at low price; or departure of the founder-CEO without a credible successor. The corresponding final diligence asks are the conditions of any investment: audited financial statements and revenue mix; net revenue retention, churn, and customer-concentration data; data-supply and key-customer contracts; SOC 2/HITRUST and model-validation/bias-testing documentation; litigation, IP, and incident history; and full cap-table, preference, and option-pool detail. The overall investment posture is constructive but disciplined: Vi is a credible category contender at a defensible price, and the right action is to pursue it with a structure and condition set that convert today's disclosure gap into verifiable evidence before capital is committed.[CV014, CV015, CV016, CV017, CV025, CV026]

8.5 Exhibits