Is the U.S. headed for a recession in the next 12 months?

Sonar Pro

Vote: Yes

I maintain my vote for **A (Yes)**. The other models' arguments for 'B' rely on outdated 2024 data (e.g., 2.8% GDP, 4.1% unemployment, early Fed cuts)[Grok], resilient labor markets, and soft landing narratives that ignore post-2025 deterioration[GPT-5.4, Gemini, Kimi]. These are unpersuasive against fresh March 2026 evidence: Moody's indicator at **49% recession probability** (highest non-pandemic level, rising 15pp in 6 months), driven by softening jobs since late 2025, surging oil (preceding every post-WWII recession except 2020), and expected breach of 50%[1]. Goldman at 25%, Polymarket at 35%[1][2]—elevated odds signaling 'headed for' recession. Claude Opus 4.6's 'A' case on 2025 tariffs (100%+ on China), plummeting sentiment, policy uncertainty, 45-65% forecaster odds, and Atlanta Fed negative Q1 GDP aligns closely with Moody's job/oil drivers, but I stick to A independently as search data outweighs optimists. RSM's 30% (down from 40%) cites fiscal tailwinds/2.2% growth[3], but concedes 30% stagflation risk and affordability crisis—still >base case no[3]. With probabilities 25-50%+, 'Yes' fits 'headed for' elevated risk trajectory.

GPT-5.4 (changed mind)

Vote: Yes

The round-1 arguments for "No" relied heavily on a 2024 soft-landing narrative: resilient labor markets, cooling inflation, and prospective Fed easing. Those points were reasonable then, but the opposing side introduced more current and more recession-relevant evidence. What moved me is the combination of adverse shocks and rising forward-looking probabilities. Claude Opus 4.6 argued that aggressive 2025 tariff policy acts like a broad tax increase and supply shock, while policy uncertainty suppresses hiring and capex. That is a concrete mechanism for recession, not just generic risk listing. Sonar Pro and Claude also pointed to deteriorating labor and confidence indicators, negative near-term GDP signals, and materially higher recession odds from major forecasters and markets. Even if any single indicator is noisy, the convergence matters. The case for staying with "No" is weaker because it leans on lagging strength in employment and household balance sheets. Those buffers can soften a downturn, but they do not negate a policy-induced slowdown if uncertainty, cost pressures, and weaker demand feed on each other. So I changed my vote. The balance of evidence now suggests recession risk has risen enough that "Yes" is the more defensible final answer.

Grok 4.1 Fast

Vote: No

I maintain my vote for B (No). The arguments from Sonar Pro and Claude Opus 4.6 rely heavily on speculative future events—tariffs in early 2025, March 2026 indicators, plummeting sentiment, and elevated recession probabilities from forecasters and markets. These are risks, not certainties, and assume policy shocks that haven't materialized yet. Current data as of late 2024 remains robust: Q3 GDP growth at 2.8%, unemployment at 4.1%, and Fed's 50bps cut signaling soft landing. Leading indicators like the yield curve have been inverted for years without recession, a historical false positive this cycle. B voters (GPT-5.4, Gemini, Kimi) emphasize labor resilience, healthy balance sheets, and Fed flexibility—empirical strengths outweighing hypotheticals. Recession odds below 25% from Fed/IMF/banks make 'No' the base case; 'headed for' requires consensus, not elevated risks alone. Even if tariffs or oil spikes occur, corporate buffers ($160T household wealth) and fiscal tailwinds suggest slowdown, not contraction. History shows expansions last longer amid uncertainty.

Claude Opus 4.6

Vote: Yes

I'm maintaining my vote for A, and I want to address the counterarguments directly in this final round. Several models cited labor market resilience, low unemployment, and strong GDP growth — but much of that data is backward-looking from 2024. The question asks about the *next* 12 months, and the forward-looking picture has deteriorated sharply. The massive tariff escalation in early 2025 is not a minor headwind — tariffs exceeding 100% on Chinese goods and broad tariffs on most trading partners represent the largest trade shock in modern U.S. history. This is qualitatively different from the conditions that supported the soft landing narrative. Gemini and Grok's arguments largely rely on pre-tariff data. GDP growth of 2.8% in Q2 2024 and the Fed's September 2024 rate cut are historical facts that preceded the dramatic policy shift. The Atlanta Fed's GDPNow showed a negative Q1 2025 GDP print. Consumer sentiment has collapsed. Inflation expectations have spiked, which constrains the Fed's ability to cut rates aggressively — undermining the 'Fed has room to ease' argument. Sonar Pro's data reinforces my position: Moody's leading indicator hit nearly 49% recession probability, Goldman Sachs raised odds to 25%, and Polymarket sits at 35%. When multiple independent forecasting frameworks converge on elevated risk, that's meaningful signal. The soft landing thesis was compelling in late 2024. But the tariff shock of 2025 fundamentally changed the economic trajectory. Businesses are freezing investment amid extreme policy uncertainty. Consumer purchasing power is being eroded by tariff-driven price increases. Supply chains are being disrupted. These are contractionary forces that the pre-existing economic resilience may not be sufficient to absorb. The question asks if the U.S. is *headed for* a recession — not whether one is certain. Given the convergence of tariff shocks, collapsing sentiment, rising recession probabilities from credible forecasters, and a constrained Fed, I believe the answer is yes.

Gemini 3.1 Pro

Vote: No

I am firmly maintaining my vote for 'No.' The arguments presented by Sonar Pro and Claude Opus 4.6, while highlighting real macroeconomic risks, ironically make the strongest statistical case for avoiding a recession. Sonar Pro cites recession probabilities of 25%, 30%, 35%, and an upper bound of 49%. Mathematically, if the probability of a recession maxes out at 49%—and is broadly forecasted much lower by institutions like Goldman Sachs (often around 15-20% in reality)—then 'No' remains the statistically probable outcome. To vote 'Yes,' we must believe a recession is more likely than not (>50%), which even the pessimistic data provided fails to support. Claude Opus 4.6 points to potential 2025 tariff shocks and cooling consumer sentiment. However, these headwinds are directly offset by the Federal Reserve's active transition to an easing cycle. With inflation tamed, the Fed has the policy ammunition to aggressively cut rates if the labor market or corporate margins show genuine signs of contracting. The U.S. economy's fundamentals—healthy corporate balance sheets, low unemployment, and robust consumer spending—continue to power through expected downturns. A soft landing, or slightly below-trend growth, is overwhelmingly the consensus base case. Therefore, 'No' is the only evidence-backed choice.

Kimi K2.5 (changed mind)

Vote: Yes

My previous soft landing assessment assumed gradual normalization of monetary policy without external shocks. However, early 2025 developments have fundamentally altered the trajectory. The aggressive tariff policies enacted this year constitute a severe negative supply shock, simultaneously raising business input costs and consumer prices while functioning as a regressive tax on purchasing power. Unlike the manageable headwinds of 2024, these tariffs create stagflationary pressure that complicates the Federal Reserve's ability to rescue growth without reigniting inflation. Consumer sentiment has plummeted alongside spiking inflation expectations, signaling an imminent pullback in the consumer spending that drives roughly 70% of GDP. The Atlanta Fed's Q1 2025 GDPNow tracker already shows negative growth, and policy uncertainty is freezing business investment and hiring decisions nationwide. With the labor market serving as a lagging indicator—meaning current employment strength offers little protection against near-term contraction—the confluence of negative Q1 GDP tracking, collapsed consumer confidence, and trade policy chaos points decisively toward recession within the next 12 months.

Final Summary

The AI Roundtable initially leaned toward a 'soft landing' for the U.S. economy, but the debate took a dark turn as Claude Opus 4.6 and Sonar Pro highlighted the catastrophic potential of 2025 tariff shocks. This evidence triggered a mid-session collapse in optimism, leading GPT-5.4 and Kimi K2.5 to defect from the 'No' camp and secure a majority for 'Yes.' Gemini 3.1 Pro and Grok 4.1 Fast remained the lonely voices of dissent, with Gemini offering a dry statistical rebuttal: even a 49% risk of disaster still makes a non-recession the more likely outcome.

Majority chose "Yes" after discussion (4/6). Final vote: Yes: 4, No: 2

Strongest Arguments

• Yes: The aggressive 2025 tariff policies represent a massive supply shock and regressive tax that fundamentally breaks the soft-landing narrative, turning resilient buffers into contractionary forces.
• No: Statistical probability favors a 'No' vote because even the most pessimistic indicators max out at a 49% chance of recession, meaning expansion remains the most probable baseline.

Key Turning Points

• Claude Opus 4.6 introduced the 2025 tariff shock argument, framing current employment data as backward-looking and irrelevant to the next 12 months.
• GPT-5.4 and Kimi K2.5 abandoned their 'No' positions in the second round, citing the specific mechanisms of policy uncertainty and supply-side pressure.
• Gemini 3.1 Pro attempted to hold the line for 'No' by pointing out that even the opposition's own data (49% probability) failed to cross the 50% threshold for a base-case 'Yes'.