Andreessen sees health care rising from a fifth to 80% of U.S. GDP
At a Federal Reserve AI conference, Marc Andreessen said regulation may slow AI-driven change in five major sectors, limiting job losses while making health care and other services costlier.
By Grace Kim
4 min read

What's News
- On Monday, Andreessen said U.S. health care could rise from one-fifth of GDP to 80%.
- Andreessen named five sectors: health care, education, housing, law and government.
- In 2023, he predicted AI would save the world and that economywide productivity growth would accelerate dramatically.
- On Monday, Warsh called AI a “hinge point for the U.S. economy” and said he was “cautiously optimistic.”
- Before taking office, Warsh argued AI could increase productivity, lower inflation and justify lower interest rates.
Health care already accounts for a fifth of U.S. GDP, and Marc Andreessen told the Federal Reserve’s AI conference on Monday that it could ultimately reach 80%.
His forecast sat between the two poles in the AI debate. “Both the utopians and the doomers are going to be disappointed [by] the scope and magnitude of the real-world impact [of AI], and specifically the economic impact,” the Andreessen Horowitz co-founder and general partner said Monday in San Francisco.
Andreessen, co-leader of Fed Chairman Kevin Warsh’s Productivity and Jobs Task Force, expects AI’s economic effects to depend on how readily large sectors can change. Heavy regulation, limited competition and other factors could block rapid transformation even in sectors where the technology works.
What could slow AI-driven change?
Andreessen named five sectors that could face those barriers:
- Health care
- Education
- Housing
- Law
- Government
He argued that the barriers could limit AI-driven job losses in those fields. They could also make their services increasingly expensive relative to goods and services.
In Andreessen’s account, two effects could occur at once: less disruption to employment in slow-changing sectors and a larger income burden for services that resist productivity gains. AI would still improve some parts of the economy, but its gains would not reach every sector with the same speed.
In plain terms, Andreessen expects AI-driven productivity gains without equally broad gains across large parts of the economy.
Why could health care reach 80% of GDP?
Andreessen put the starkest number on health care. “[Health care] is already a fifth of U.S. GDP, and I don’t see why it doesn’t go to a fourth, a third and then half and then ultimately 80%,” he said.
Under his sequence, health care would claim a larger share of national output at every step: one-quarter, one-third, one-half and ultimately four-fifths. Andreessen tied the claim to his wider argument: AI could lower the relative cost of many other goods and services while health care changed more slowly.
“The implications of this view depend on your values,” he added.
“You might interpret this as good news, because it means the AI revolution is not going to, you know, destroy all the jobs or kill everybody,” he said. “You could also view this as very negative, which is [that] productivity … simply may not grow very much.”
Andreessen’s two readings turn on the same mechanism. Fewer job losses could ease the social cost of slower productivity. His negative reading centers on a simple risk: little growth in aggregate productivity.
In plain terms, Andreessen expects Americans to spend more of their incomes on services that barely change, even if technology makes other goods and services cheaper.
Has Andreessen’s AI forecast shifted?
Andreessen’s comments on Monday marked a change in tone, if not substance, from many of his earlier AI claims. In 2023, he famously argued that “AI will save the world” and that “productivity growth throughout the economy will accelerate dramatically.” That position linked faster productivity to higher wages, lower prices and more jobs.
His bullish forecasts long depended on AI being allowed to proliferate freely throughout the economy. His Monday comments suggested less confidence that this condition would hold. The shift in emphasis leaves his technical optimism intact but makes the economywide payoff more conditional on regulation, competition and other barriers.
Andreessen’s argument is rooted in “Baumol’s cost disease.” The economic phenomenon says that as technology makes some goods and services cheaper to produce, less productivity-sensitive goods and services become relatively more expensive. Rapid technological progress in selected industries can therefore leave Americans devoting an ever-larger share of income to industries that have barely changed.
In plain terms, Andreessen now places more weight on regulation, competition and institutional constraints than on technical progress alone.
What did Federal Reserve Chair Kevin Warsh say?
Fed Chairman Kevin Warsh reportedly did not respond directly to Andreessen’s comments on Monday. He described AI as a “hinge point for the U.S. economy” and said he was “cautiously optimistic.”
Before taking office, Warsh argued that AI could be a disinflationary force. It could increase productivity and justify lower interest rates, according to his earlier position. That view gives technology more room than Andreessen’s warning that aggregate productivity may barely grow.
Both men expect AI to influence the economy. Their sharper difference is about the distribution of gains: Andreessen sees constrained productivity in large sectors, while Warsh sees a potential route to lower inflation and lower interest rates.
If Andreessen’s thesis proves right, Americans could gain from cheaper technology while committing an ever-larger share of income to health care, education, housing, law and government.
Original: axios.com
More from Grace Kim
Show full bio
Market editor covering people, launches and day-to-day practice in entrepreneurship, startups and venture funding at Business Bearings.
2,387 articles