Alphabet's earnings show where AI healthcare investments go

Alphabet’s next earnings discussion lands with an unusually large number hanging over it: the company has outlined roughly $180 billion to $190 billion of 2026 capital expenditure for AI infrastructure, about double the $91.4 billion spent in 2025, while Google Cloud’s backlog has been reported at $462 billion.[1] For healthcare buyers, researchers, and informatics teams, that scale is useful only after it is disaggregated. Alphabet does not report “healthcare AI” as a clean segment. The better question is where the investment becomes observable: in cloud contracts, drug-design capital, longitudinal research platforms, venture exposure, or ecosystem grants.

That is also the more practical earnings question for Alphabet’s AI healthcare investments. A health system cannot procure a capex line. A life-sciences partner cannot run a trial on a keynote. What matters is whether the AI buildout is tied to a usable platform, a defensible dataset, a regulatory path, a revenue line, or a clinical-stage asset.

Cloud is the financial center of gravity, but not a healthcare segment

Google Cloud is the cleanest place to look for near-term financial evidence because it is an operating business rather than a research promise. The Cloud business has been reported at $20 billion in quarterly revenue, up 63% year over year, with operating margin expanding from 9.4% to 33%.[2] Reuters also reported that 75% of Cloud customers were using AI products.[3] Those numbers do not prove healthcare-specific monetization, but they do show that Alphabet’s AI infrastructure spending is tied to a business line with revenue, backlog, and margin expansion.

For healthcare, Cloud is not just generic compute rented by hospitals. It is where provider data modernization, medical imaging workflows, life-sciences analytics, research environments, model deployment, and enterprise AI governance can be sold as infrastructure. A health-system executive looking at Alphabet’s AI exposure should therefore treat Cloud as the most financially mature channel, while resisting the temptation to infer a healthcare revenue number Alphabet has not provided.

The distinction matters because the $462 billion Cloud backlog is a total Cloud figure, not a healthcare backlog.[1] A portion may be relevant to healthcare and life sciences, but the company’s public reporting does not let outsiders allocate that backlog by clinical vertical. That makes Cloud highly material to the healthcare AI story and simultaneously frustrating as evidence. It is measurable, but not at the level most healthcare market analysts would want.

This is where Alphabet’s capex story becomes procurement-relevant. If a hospital, payer, or biopharma customer already runs data workloads on Google Cloud, Alphabet’s AI infrastructure can reduce friction for model development, analytics, and managed AI services. If the customer’s data estate is elsewhere, the same spending may still improve Google’s competitive posture without immediately changing that customer’s clinical workflow. Capex creates capacity; contracts and implementations decide whether that capacity becomes healthcare infrastructure.

Isomorphic is the clearest drug-discovery bet

Isomorphic Labs is the channel where Alphabet’s AI research lineage becomes most directly connected to drug-design capital. The company raised a $600 million Series A in 2025 and a $2.1 billion Series B in May 2026, led by Thrive Capital with participation from Alphabet, GV, MGX, and Temasek, bringing total disclosed funding to $2.7 billion.[4] Reuters reported that Isomorphic has partnerships with Eli Lilly and Novartis and is targeting its first AI-designed drugs to enter clinical trials by the end of 2026.[5]

That is a different kind of evidence from Cloud revenue. It is not yet proof of clinical effectiveness or commercial return. It is evidence that Alphabet’s AI biology work has been organized into a financed company, with pharma partnerships and a stated clinical-development timetable. For readers who have watched AI drug-discovery companies overpromise on speed, the trial target is useful precisely because it creates a harder test. Once programs enter humans, model confidence has to meet safety monitoring, dosing decisions, endpoints, and regulatory scrutiny.

The AlphaFold background matters here only because it explains why this is not a random venture-style adjacency. Google said AlphaFold received the Nobel Prize in Chemistry, and its 2026 health update also described AlphaGenome as achieving a 14.7% improvement over existing methods for identifying genetic relationships.[6] Those are research and platform signals, not drug approvals. Isomorphic is the vehicle through which that scientific credibility is being tested against the far less forgiving economics of therapeutic development.

The main disclosures stop short of the questions investors and pharma partners eventually need answered. Isomorphic’s Series B valuation was not disclosed.[4] The company’s program-by-program economics are not visible. Its planned clinical trials had not yet produced human efficacy evidence as of July 21, 2026. The channel is therefore strong on capital formation and strategic relevance, but still early on clinical proof.

Verily shows the data-platform side of the portfolio

Verily occupies a less tidy but important part of Alphabet’s healthcare AI map. In March 2026, Verily announced a $300 million raise involving Series X Capital, Alphabet, UCHealth, and the University of Colorado Anschutz Medical Campus.[7] Its Pre platform powers the NIH All of Us Research Program, which Verily said supports more than 21,500 researchers.[7] That places Verily closer to longitudinal research infrastructure and real-world data operations than to a single model demo.

The newer product mix also shows where Verily is trying to make AI operational. Fierce Healthcare reported on Verily Me, a consumer app with a Violet AI agent, as well as Lightpath for chronic disease management, Samsung Galaxy Watch integration, and Salesforce integration.[8] Those products point toward a healthcare AI business built around data capture, patient engagement, chronic-condition workflows, and enterprise distribution.

Verily is not as financially transparent as Google Cloud, and it is not as cleanly clinical-stage as Isomorphic. Its importance is that it keeps Alphabet in the plumbing of health research and care management: consented longitudinal datasets, participant engagement, wearable inputs, chronic disease programs, and integration layers. For health systems, that may be more consequential than another foundation-model announcement, but the public evidence still does not support a precise ROI calculation.

The venture arms are a signal, not an operating result

GV and CapitalG widen Alphabet’s exposure beyond companies it controls. Crunchbase reported that more than 50% of GV’s investment team is focused on healthcare and life sciences, and that GV has led $1.6 billion in health rounds since 2020.[9] The same reporting identified eight healthcare AI unicorns in the relevant portfolio.[9] That is meaningful as a capital-allocation signal: Alphabet-affiliated investors are not treating healthcare AI as an occasional theme.

It is also easy to overread. Venture exposure does not tell a health-system CIO which platform will integrate with the EHR, which model has been externally validated, or which company will survive reimbursement pressure. The venture channel is useful for understanding where Alphabet-linked capital is searching for optionality. It should not be counted as Cloud revenue, clinical evidence, or proof that Alphabet can consolidate healthcare AI demand.

Google.org is ecosystem shaping, not monetization

Google.org’s healthcare AI activity belongs in the map, but in a smaller box. Its $10 million clinician AI education initiative is a real intervention in adoption capacity and workforce readiness.[10] It can help clinicians, administrators, and educators understand AI tools before procurement decisions harden into institutional policy.

That does not make it equivalent to Google Cloud’s operating scale or Isomorphic’s drug-development financing. Education grants shape the environment in which healthcare AI is evaluated. They do not, by themselves, demonstrate product revenue, clinical effectiveness, or return on Alphabet’s infrastructure spending.

What earnings can and cannot prove

Earnings material can clarify the size of Alphabet’s AI buildout and the performance of the businesses most likely to absorb it. It cannot, on its own, answer how much of that buildout is healthcare-specific. The July 22, 2026 Q2 report had not been released as of July 21, so any Q2 figures circulating before then were analyst projections rather than reported results. They may frame expectations, but they should not be used as evidence of actual healthcare AI performance.

The same caution applies to Alphabet’s non-operating gains. The reported $37.7 billion Q1 other-income figure included unrealized mark-to-market gains on minority stakes such as Anthropic and SpaceX.[1] That is non-cash and non-operating. It may affect reported income, but it is not healthcare AI revenue and should not be interpreted as monetization of clinical tools or drug-discovery platforms.

Berkshire Hathaway’s $10 billion investment in Alphabet, reported in June 2026 as part of an $84.75 billion equity raise, is relevant context for market confidence in Alphabet’s broader AI and infrastructure strategy.[11] It is not evidence that healthcare AI is producing returns. The same investor can like Alphabet’s overall position while the healthcare-specific return profile remains unresolved.

The narrower conclusion healthcare buyers can use

Alphabet’s healthcare AI investments are already structurally significant. They are visible through a scaled Cloud business, a well-capitalized AI drug-design company, a health data and chronic-care platform company, venture exposure, and clinician education grants. That is more than speculative adjacency.

The evidence is uneven by design. The strongest financial signal sits in Google Cloud, but healthcare-specific revenue is not broken out. The strongest clinical-stage optionality sits in Isomorphic and Verily, but human efficacy, regulatory progress, and durable commercial ROI remain incomplete. GV, CapitalG, and Google.org help explain the surrounding ecosystem, not the core operating result.

For now, the disciplined reading of Alphabet’s earnings is not that the company has won healthcare AI. It is that Alphabet has built several serious channels through which general-purpose AI infrastructure can become healthcare capability. The unresolved question is how much of the $180 billion to $190 billion AI buildout will ultimately be attributable to healthcare outcomes rather than to the broader economics of AI compute.

References

1. Alphabet 2026 AI capex and Cloud backlog reporting, Reuters and Yahoo Finance, February 2026.
2. Google Cloud revenue and margin reporting, Fortune, October 2025.
3. Google Cloud AI customer adoption reporting, Reuters, April 2026.
4. Isomorphic Labs raises $2.1B Series B for AI drug discovery, Fierce Biotech, May 2026.
5. Isomorphic Labs pharma partnerships and clinical trial target reporting, Reuters, May 2026.
6. The Check Up with Google Health 2026, Google, 2026.
7. Verily announces $300M funding and Pre platform updates, Verily, March 2026.
8. Verily Me, Violet AI agent, Lightpath, Samsung and Salesforce integrations reporting, Fierce Healthcare, March 2026.
9. GV healthcare and life sciences investment reporting, Crunchbase.
10. Google.org clinician AI education initiative, Google.org.
11. Berkshire Hathaway investment in Alphabet reporting, Reuters, June 2026.