Ken Perez: AI, the healthcare industry’s private-sector-led revolution

With the advent of AI, many industry observers say the U.S. healthcare industry is standing at the brink of a sea change. As industrywide adoption of AI accelerates, now is a good time for the industry’s stakeholders to pause for a moment and take stock of what this new technology means for the healthcare system.

Remember the last time an emerging technology promised to revolutionize healthcare? It was 17 years ago.

In February 2009, in his address to a joint session of Congress, President Barack Obama stated, “Our recovery plan will invest in electronic health records and new technology that will reduce errors, bring down costs, ensure privacy and save lives.”^a The recovery plan was the Health Information Technology for Economic and Clinical Health (HITECH) Act, which was enacted as part of the American Recovery and Reinvestment Act of 2009 (ARRA).

Federal government support for EHRs

Under HITECH, the federal government made available financial incentives to physicians and hospitals to adopt electronic health records (EHRs) and demonstrate their meaningful use. Initially, it was projected that this program would cost $27 billion over 10 years, but subsequent estimates pegged the cost much higher, with a 2019 article by Kaiser Health News and Fortune magazine reporting it as having reached $36 billion.^b

Why did the federal government resort to this unprecedented way to motivate healthcare providers to adopt EHRs? With healthcare expenditures rising, it was thought that use of EHRs could improve the efficiency of the healthcare system, producing a so-called “digital dividend.”

Also, the United States lagged other industrial nations in adoption of EHRs. In 2008, while the United Kingdom, the Netherlands, Australia and New Zealand had nearly universal use of EHRs among general practitioners, only 10% to 30% of ambulatory care physicians in the United States and Canada used EHRs consistently.^c Last, given the healthcare industry’s well-deserved reputation as a technology laggard — slow to adopt new practices, taking an average of 17 years for research evidence to change practice — it was felt that a strong “nudge” by the federal government, a combination of financial “carrots” (adoption subsidies) at first and “sticks” (penalties for non-meaningful use) later, was needed.^d

The EHR incentive program’s results

HITECH achieved its basic goal of spurring EHR adoption. By 2017, 86% of office-based physicians had adopted EHRs and 96% of non-federal acute care hospitals had done so, according to the 2019 Kaiser Health News/Fortune article. However, there have been many downsides of EHRs, including the following.

Patient safety risks. Software glitches, user errors and other system flaws have resulted in medical errors in general and medication errors in particular.

Fraud. The software can be used to upcode, or overcharge, and some healthcare providers have allegedly overstated their use of EHRs, defrauding Medicare and Medicaid.

Lack of interoperability. David Blumenthal, Obama’s national coordinator for health information technology and one of the chief architects of the EHR initiative, acknowledged that interoperability goals took a backseat to just adopting one of the EHRs offered by more than 700 EHR vendors. As quoted in the Kaiser Health News/Fortune story, Blumenthal stated, “We had an expression: They had to operate before they could interoperate.”

Unfortunately, this, combined with perverse business incentives, resulted in EHR systems routinely being unable to talk to one another.

Physician burnout. The KFF Health News/Fortune article cites the findings of a 2017 study in the Annals of Family Medicine that physicians spend an average of 5.9 hours on EHRs compared with 5.1 hours spent with patients (out of an 11.4-hour workday).

Who won as a result of the EHR incentive program? Clearly, the largest EHR vendors benefited the most, prompting some to criticize the program as a huge wealth transfer from the American taxpayer to primarily those companies.

The promise of AI

Today, as AI similarly promises to revolutionize healthcare, the Trump Administration is cheering on this revolution. President Donald Trump opined in America’s AI Action Plan:

In his former role as the director of the Digital Health Center of Excellence, Center for Devices and Radiological Health, in the U.S. Food and Drug Administration (FDA), Troy Tazbaz said:

The Trump administration’s stance on AI for healthcare

However, in striking contrast with HITECH’s subsidization of EHR adoption, the Trump administration is not disbursing funds to healthcare providers to encourage them to adopt AI. This policy is partly due to two fiscal realities: CMS’s budget — which paid for the EHR incentives — is under pressure; and the Medicare Hospital Insurance Trust Fund is just eight years from insolvency.^g

But more fundamentally, it reflects a philosophy of supporting private sector-led innovation, primarily by reducing burdensome regulations. To that end, as of December 2025, the FDA had approved 1,300 AI-enabled medical devices.^h

Private sector investment in AI

The private sector’s AI leadership is evident, as it is pouring unprecedented amounts of money into AI. In 2026, Alphabet, Amazon, Meta and Microsoft — the so-called hyperscalers — are projected to spend $670 billion on AI infrastructure, more than 10 times the Apollo space program (1960-1973) as a percentage of the nation’s gross domestic product.ⁱ Healthcare, accounting for about a fifth of the U.S. economy, has emerged as the primary target of these investments. And digital health startups involving any AI component commanded valuation premiums and captured 62% of all venture funding in 2025.^j

Healthcare’s adoption of AI

So, are hospitals using AI even in the absence of federal subsidies? The answer is, “Yes.” In 2024, 71% of hospitals reported using predictive AI integrated into their EHR.^k And in terms of generative AI, according to a survey of 2,174 non-federal U.S. hospitals conducted in 2024, 31.5% reported using it and 24.7% planned to do so in one year. Thus, if the latter group of hospitals followed through in their plans, it would mean that by the end of 2025, half of non-federal hospitals in the U.S. were using generative AI.^l

What lies ahead for AI in healthcare

These trends are certainly encouraging. But the jury is still out as to whether AI is delivering sufficient value to enable AI-focused digital health startups and other firms to evolve from surviving on venture funding to garnering substantive revenues and achieving profitability.

Reflecting on the heady nature of the current AI bubble, an AI leader at one of the hyperscalers recently told me that market consolidation is inevitable, which obviously would pose some risks for healthcare providers, such as unsupported software and outdated models and algorithms if AI vendors were to cease operations.

However, the value question and the inevitability of market consolidation are no different from what took place with EHRs. What is different is that the AI revolution is being led by the private sector, which incorporates the refining fire of market competition and dwarfs the size of the federal government. That means it has more staying power and greater prospects for success than the EHR incentive program ever had. 

Footnotes

a. CMS, “Electronic health records at a glance,” July 13, 2010.
b. Worzala, C., “Policy update: Federal incentives for the adoption of electronic health records,” Journal of Oncology Practice, September 2009; and Schulte, F., and Fry, E., “Death by 1,000 clicks: Where electronic health records went wrong,” Kaiser Health News and Fortune,
March 18, 2019.
c. Jha, A., et al., “The use of health information technology in seven nations,” International Journal of Medical Informatics, December 2008.
d. Morris, Z., et al., “The answer is 17 years, what is the question: understanding time lags in translational research,” Journal of the Royal Society of Medicine, December 2011.
e. The White House, “Winning the race: America’s AI action plan,” July 2025.
f. Tazbaz, T., “Blog: The promise artificial intelligence holds for improving health care,” FDA, July 25, 2024; Tazbaz is currently senior vice president of corporate strategy & operations for Oracle in San Francisco. 
g. Committee for a Responsible Federal Budget, “Social Security and Medicare trustees release 2025 reports,” June 18, 2025.
h. Murphy, H., “Dozens of new AI-powered devices make FDA’s list of approvals,” Radiology Business, Dec. 14, 2025.
i. Bobrowsky, M., et al., “Big Tech’s AI push is costing a lot more than the Moon Landing,” Wall Street Journal, Feb. 7, 2026.
j. Topaz, M., et al., “The coming clinical correction: Why health care needs its AI bubble to burst,” Health Affairs, Feb. 4, 2026.
k. Chang, W., et al., “Hospital trends in the use, evaluation, and governance of predictive AI, 2023-2024,” HealthIT.gov, Assistant Secretary for Technology Policy, September 2025. Predictive AI refers to using statistical analysis and machine learning to classify or produce a risk score for individuals, such as readmission risk prediction, early disease detection, appointment no-show and treatment recommendations.
l. Everson, J., Nong, P., and Richwine, C., “Uptake of generative AI integrated with electronic health records in U.S. hospitals,” JAMA Network Open, Dec. 12, 2025.