Verily Adds NVIDIA as Investor in Precision Health AI

Verily Health extended its March 2026 fundraising on September 9 with new investment from NVIDIA and additional capital from CU Healthcare Innovation Fund II. The company did not disclose the size, valuation, or security type of the new investments.

The distinction matters because Verily's $300M figure belongs to the round it announced in March, led by Series X Capital with participation from Alphabet, UCHealth, the University of Colorado Anschutz, and other investors. This extension adds owners who already sit inside Verily's operating thesis: NVIDIA on the computing side and a CU-affiliated healthcare fund on the clinical and institutional side.

For Verily, the wager is bigger than training another model. The company is trying to connect fragmented health data, controlled research environments, AI development, clinical workflows, and patient-facing software without asking healthcare organizations to lower the standards that make the work worth doing.

What Verily Announced

Verily said NVIDIA became a new investor as the company extended its March fundraising, while existing investor CU Healthcare Innovation Fund II made an additional investment. Current Chairman and CEO Stephen Gillett said the backing would strengthen Verily's effort to build a trusted data and intelligence foundation for healthcare.

The announcement did not provide an incremental dollar amount. A secondary market-data report associated the September closing with $300M, but Verily's primary release assigns that number to the March round, so the cleaner accounting is an undisclosed strategic extension rather than a second $300M raise.

That March round changed Verily's ownership structure as well as its capitalization. Series X Capital led the $300M investment, Alphabet became a significant minority investor instead of the controlling owner, and the business converted from an LLC into Verily Health Inc.

Why NVIDIA Fits the Cap Table

NVIDIA did not arrive as a stranger with a term sheet. Verily and NVIDIA announced a technical collaboration in October 2025 that placed NVIDIA tools and accelerated computing inside Verily's research environment. Researchers using Verily Workbench can access NeMo, Parabricks, RAPIDS, and NVIDIA GPU infrastructure for biomedical analysis and model development.

The two companies also worked on Forecast 1.0, a foundation model that combines genomic and electronic health record data to predict long-term health risks. Verily says NVIDIA's stack improved the model's predictive performance and computational efficiency, but the company has not published enough independent evidence in the investment announcement to treat those gains as broadly validated clinical outcomes.

That qualification sharpens the investment logic. NVIDIA benefits when healthcare organizations move larger, more complex workloads onto accelerated computing, while Verily needs compute, software, and engineering support that can operate inside a regulated data environment. Equity aligns those incentives beyond a vendor contract.

The Platform Verily Is Financing

Verily Pre is the center of the strategy. Its Refinery product ingests and standardizes structured and unstructured healthcare data, Exchange supports discovery and access to research-ready datasets, and Workbench provides a governed environment for analysis, collaboration, and AI development.

The unglamorous work lives between those components. Health records, genomic files, device data, laboratory results, clinical notes, permissions, and data-use policies rarely arrive speaking the same language. Verily's Syntax data model uses a clinically informed, FHIR-native structure to make those inputs more interoperable and accountable before models start making inferences from them.

The patient-facing side is Verily Me, which brings medical-record context, clinician recommendations, and the Violet AI companion into one application. Pre is therefore being asked to support both institutional research and individual health navigation, two settings that share data dependencies but carry different expectations around consent, interpretation, and responsibility.

The Clinical Side of the Investment

CU Healthcare Innovation Fund is affiliated with the University of Colorado Anschutz Medical Campus and invests in healthcare IT, software-enabled services, medical devices, diagnostics, and therapeutics. Its additional investment gives Verily more than capital; it deepens a relationship connected to researchers, clinicians, and health-system workflows.

Verily previously announced work with UCHealth, CU Anschutz, and RefinedScience using Pre for oncology data extraction and model development. The company reported more than 95% accuracy and 30x faster extraction in initial acute myeloid leukemia work. Those figures remain company-reported, yet they identify the practical standard the partnership must keep meeting: better data preparation has to survive contact with clinical judgment and operational governance.

This is where strategic investors can become useful beyond their logos. NVIDIA can influence the speed and economics of model development. The CU ecosystem can pressure-test whether those models, data pipelines, and controls fit the institutions expected to use them.

From Moonshot to Healthcare Infrastructure

Verily began as a Google X moonshot in 2015 under founder and former CEO Andrew Conrad. The current leadership team is led by Stephen Gillett, with Scott Burke serving as CTO over software engineering, AI and data science, security, and IT.

The company has spent a decade moving through devices, research platforms, chronic-care programs, data products, and several portfolio resets. Its current focus is narrower and more infrastructural: make health data usable for AI, give researchers a governed place to work with it, and connect the resulting intelligence to care and patient software.

The September investment is evidence that two partners want exposure to that direction, not proof that Verily has solved the deployment problem. Healthcare organizations still have to reconcile technical performance with privacy, regulation, workflow fit, clinician trust, and measurable outcomes.

Verily's next chapter will be written in those handoffs. A faster model matters when the data is fit for purpose, the research can be reproduced, the care team understands the output, and the patient experiences something better than another layer of software asking for trust.