Warsh: Fed Models 3 AI Scenarios, 2 With Rising Unemployment

Federal Reserve Chair Kevin Warsh said on 16 November that the Fed Board’s staff is modelling how super intelligence could play out in the US economy, and described three outcomes: broadly shared gains, labor market disruption, and a slowdown after equity valuations fall. “The Board’s staff has begun to evaluate various contingencies, applying stylized, model-based frameworks to generate scenarios”, he said in opening remarks published by the Fed for its Productivity and Jobs Task Force Conference in San Francisco.

In two of the three, by our count, unemployment rises: in one the jobless rate goes above levels consistent with full employment, and in the other Warsh described “a more notable rise in unemployment”. The third scenario starts where equity valuations fall after heavy investment; our nine-stock basket of US-listed AI data-centre suppliers fell 39.4% from its 22 June peak to the 13 November close, by our arithmetic in our piece on the $4.2 trillion sell-off. Warsh called the three “illustrative scenarios, not forecasts”.

Business Pill · Financial Conditions Explained

An evergreen Business Engineer explainer on financial conditions. It teaches the general idea only and does not discuss the Fed speeches.

The key insight: As we read the remarks, the Fed has put a valuation-led slowdown on its list of super-intelligence outcomes alongside the good case, and two of the three scenarios Warsh described include rising unemployment, by our count.

The Three Scenarios, in Warsh’s Words

The first is the good case. “At one end, SI leads to an increase in productivity and economic gains that are shared broadly across labor and the owners of capital”, Warsh said. “In that event, real GDP growth rises persistently above its long-run average, and the U.S. stays close to full employment”.

The second keeps growth but loses jobs. “An SI-driven expansion could bring substantial labor market disruption as some jobs are displaced, rather than complemented, by technology”. GDP growth would still rise above its long-term average, he said, but “widespread automation, combined with labor market frictions, could push the jobless rate above levels consistent with full employment”.

The third is a disappointment. “Or SI could disappoint. After significant capital investment and stock market gains, expected returns could turn down and equity valuations could fall, leading to an economic slowdown and a more notable rise in unemployment”. “In time, the economy would recover, but without the productivity gains”.

He added that “many other outcomes—probably including one that will prove most accurate—could be drawn”. The text gives no probabilities and no figures for growth or unemployment in any of the three.

The three Board staff scenarios for super intelligence as Fed Chair Kevin Warsh described them on 16 November
The three Board staff scenarios for super intelligence as Fed Chair Kevin Warsh described them on 16 November 2026; Warsh called them “illustrative scenarios, not forecasts”. Quoted words are his; the scenario labels, summaries and the two-of-three count are ours.

What Warsh Said About Risk and Who Decides

Warsh said “the risks in this case, if overlooked, can be highly disruptive and economically costly” and that “SI empowers all, including bad actors, so we are thinking especially of critical infrastructures and their potential vulnerabilities”.

On the Fed’s role, he said: “At the Fed, our role in all this is neither advocacy nor opposition”. He said wide adoption of the technology “will have significant impact on the economy, with important implications for monetary policy down the road”. He gave no guidance on interest rates.

On people, he said machines “do not take the place of human judgment, morality, or responsibility”, and that “Humans are the designers and the builders, and humans must be the deciders”.

He listed the questions the Fed is asking, including “Will this new technology deliver a rise in total factor productivity?” and “When will the surge start, and when will it taper?” He also said research efforts “have increased more than 20 times since the 1930s” by some estimates, while research productivity has fallen, citing Nick Bloom, Chad Jones and coauthors.

Barr, Same Day: How a Repricing Would Spread

In a separate speech in Milan on the same day, Governor Michael Barr said he had “discussed three different scenarios” in an earlier speech, and that in one of them the question is “whether investors will see returns on the AI buildout on a timeline and to an extent consistent with their expectations”.

Barr said losses from a large-scale repricing could lower output growth “first, from the direct effect of a drop in investment” and “second, from the knock-on effects of this loss of wealth in lowering consumer and potentially business spending”.

He put on-balance-sheet debt of the seven largest AI firms on course to “exceed $800 billion before year’s end”, and said the Fed’s Financial Stability Report estimates the potential maximum nominal value of their off-balance-sheet obligations “now exceeds $3 trillion”. Hyperscalers and GPU makers, he said, also have “over $100 billion in cumulative equity investments” in AI labs and cloud start-ups.

Where the Third Scenario Meets the Market

The Fed’s 13 November Financial Stability Report said equity valuations had remained high, “partly driven by market enthusiasm for super intelligence (SI)”, and that the largest US banks have close to $1 trillion of exposure to large technology firms, data centre developers and related vehicles, as our piece on the report shows. Vice Chair for Supervision Michelle Bowman warned on 10 November of a possible large correction if AI returns fall short, covered in our Bowman piece.

Some prices have already moved. Our nine-stock basket of data-centre suppliers fell 39.4% from its June peak, by our arithmetic. Nvidia, which reports on 17 November, traded at 13.3 times next-four-quarter consensus earnings on Monday morning against FactSet’s 18.3 times for the S&P 500, by our arithmetic in our forward P/E piece. Warsh’s conference and its panels were previewed in our 13 November piece.

The Structural Read

As we read the speech, Warsh framed the Fed’s job as contingency planning rather than a forecast: he called the scenarios illustrative, gave no probabilities and gave no rate guidance.

As we read the two speeches together, Barr set out how a repricing would reach output, through lower investment and lost wealth, which is the mechanism behind the third scenario Warsh described.

Kevin Warsh, Chair, Federal Reserve, 16 November 2026

“Or SI could disappoint. After significant capital investment and stock market gains, expected returns could turn down and equity valuations could fall, leading to an economic slowdown and a more notable rise in unemployment”

Three Implications

FOR WORKERS Two of the three scenarios include rising unemployment, by our count; Warsh gave no figures for any of them.

FOR INVESTORS The third scenario starts with equity valuations falling; Barr said a large-scale repricing could lower output growth through investment and lost wealth.

FOR MONETARY POLICY Warsh said the technology will have important implications for monetary policy down the road and gave no rate guidance.

The Business Engineer Lens

This story maps onto the Business Engineer framework The AI Capex Map.

The framework argues that the capex race “is a sovereign-scale financing problem dressed in tech-company clothing” and asks “what happens when revenue can’t catch up”.

As we read Warsh’s third scenario, it is the Fed’s own version of that question; the framework does not discuss this speech, and the Fed gave no figures for the scenario.

What Is Not Established

Warsh did not say which scenario the Fed thinks most likely, and he said he was “not here to offer predictions”. The Fed has not published the staff models, their assumptions or any numbers behind the three scenarios; the remarks describe them in words only.

The viral post by @DeItaone, a news relay account, summarised the remarks as Warsh stressing that “humans must retain decision-making authority”; that is the relay’s wording. Warsh’s own words are “humans must be the deciders”. The two-of-three count and the market figures are ours, not the Fed’s.

This Week In The AI Supercycle

AI Lost Its Central Banker, the Lab Oligopoly, and the Future of Open Source

This week’s issue: the Nvidia leadership change, read as the AI build-out losing its central banker: who now finances the GPU landlords, backstops the data centres and sponsors open models, and why two labs are becoming the bank.

Read the full issue →

Business Engineer Framework

The AI Capex Map

The Business Engineer’s map of how the AI buildout is financed, where it cracks and who can absorb it.

Read the Map of AI →

The Bottom Line

Fed Chair Kevin Warsh said Board staff are modelling three illustrative super-intelligence scenarios: shared gains near full employment, growth with job displacement, and a valuation-led slowdown. Two of the three include rising unemployment, by our count. Warsh gave no probabilities, figures or rate guidance.

95,000+ executives read Business Engineer for the AI strategy frameworks cited by ChatGPT, Claude, and Perplexity.

A note on sourcing. Warsh’s and Barr’s words are quoted from the full texts of their 16 November speeches published on federalreserve.gov. The @DeItaone post is a relay summary and is attributed as such. The two-of-three count, the 39.4% basket fall and the Nvidia multiple are our arithmetic or count, from the pieces linked. We did not contact the Federal Reserve for this piece.

Sources: Federal Reserve: Chair Kevin Warsh, Opening Remarks, Productivity and Jobs Task Force Conference (16 Nov 2026) · Federal Reserve: Governor Michael S. Barr, Some Potential Financial Stability Implications of Artificial Intelligence (16 Nov 2026) · X: @DeItaone post (16 Nov 2026, 17:00 UTC)

Scroll to Top
FourWeekMBA