Microsoft, Alphabet, Meta and Amazon are expected to spend more than $700 billion on capital expenditures this year — largely on data centers and computing hardware — as they build out the infrastructure needed to support artificial intelligence. For outlays on that scale to pay off, AI will have to deliver meaningful productivity gains across the economy — and few industries offer more room to prove it than healthcare.

U.S. healthcare spending reached $5.3 trillion in 2024, equivalent to 18% of GDP. By some estimates, close to $1 trillion of that total goes toward administration alone. Meanwhile, demand for care continues to outpace the available clinical workforce. If AI can lower the cost of delivering care while freeing clinicians to serve more patients, the economic opportunity could be enormous.

Major AI players are already moving deeper into the sector. Microsoft and Mayo Clinic are developing a frontier AI model designed specifically for healthcare, while Google Cloud and CVS Health have entered a long-term partnership to power CVS's new Health100 platform with Gemini, Google's family of AI models. Both moves follow the same logic: bringing AI into the specific administrative and clinical workflows where costs accumulate.

For healthcare providers, however, the AI opportunity comes with a clock. As greater efficiency becomes reflected in reimbursement, what creates an advantage today can become the baseline providers expected to meet tomorrow.

The window to capture AI's gains is narrowing

For providers, the economics of AI can be straightforward: reducing the administrative work required to deliver care lowers the cost of delivering it. The challenge is that the window to benefit from those efficiencies may be limited before reimbursement catches up.

Medicare already incorporates productivity into payment updates across hospitals, skilled nursing facilities, home health and other settings. For 2026, CMS — the agency that administers Medicare — extended that logic to physician payment, applying a 2.5% efficiency adjustment to the work component of certain non-time-based services. The adjustment accounts for efficiencies CMS expects to accrue in how those services are delivered over time.

That gives providers a reason to capture AI-driven savings early. Those that use AI to reduce documentation, intake or coordination costs can lower their cost structures now, before greater efficiency is more fully reflected in how they are paid. Those that wait risk facing the same payment pressure without having captured the savings. Over time, more efficient providers can compound that advantage by reinvesting in staff, capacity and care.

AI's real payoff is what happens after the work gets faster

Healthcare has something many industries lack: a clear place to put the productivity gains AI creates. Because demand already exceeds the available clinical workforce, time freed from documentation, scheduling or intake does not have to mean fewer people doing the same work. It can mean more patients getting care sooner.

Capturing that value requires starting with the work, not the technology. The best opportunities are often hiding in plain sight: documentation that keeps a clinician at a screen, an intake process that delays access, or a referral that stalls between care settings. Success should be equally concrete. AI should give clinicians meaningful time back, allow them to spend more meaningful time with patients, or eliminate unnecessary work rather than simply shifting it somewhere else.

Consider a patient moving from a hospital to post-acute care. A delayed referral or missing information can trigger phone calls and manual follow-up, delay care, and increase the risk of readmission. If AI helps that handoff happen correctly the first time, the value extends beyond a faster referral: it can eliminate downstream work and cost while getting the patient into care sooner. A tool that saves five minutes but adds another login, data silo or handoff has simply shifted the burden.

Providers also need to decide upfront what they will do with the capacity they create. Time returned to clinicians can mean seeing more patients or easing pressure on an already stretched workforce. When productivity gains lower the cost of delivering care, they can strengthen margins and create more room to invest in staff, expand access and improve care.

The $700 billion bet needs real-world returns

The next phase of the AI boom will be measured less by how much computing capacity gets built than by what businesses can do with it. Healthcare offers an unusually consequential test: can AI take enough cost and friction out of a massive industry to change its economics while expanding the amount of care it can deliver?

Big Tech has already committed hundreds of billions of dollars to building the infrastructure for AI. The harder question is whether that technology can produce productivity at a scale that makes the investment worthwhile. Healthcare may be one of the clearest places to find out. The markers to watch are concrete: whether AI tools move from pilots into the documentation, intake and referral work where costs sit, and how provider cost structures change as efficiency becomes embedded in payment rules.

The opinions expressed in Fortune.com commentary pieces are solely the views of their authors and do not necessarily reflect the opinions and beliefs of Fortune.

This story was originally featured on Fortune.com.