Anthropic Advisor Says AI Productivity Gains Are Vastly Exaggerated, Valuations Are ‘Crazy’

Eric Ries, author of the legendary book The Lean Startup and, more recently, Incorruptible, and a governance advisor to Anthropic since 2021, delivered a sharp critique of the AI investment boom. He told CNBC’s Squawk Box on June 8, 2026, that he believed corporate AI productivity claims are largely unproven.

He believes many AI-attributed layoffs are repackaged cost-cutting, and that private AI valuations have become detached from reality. Coming from someone who helped shape Anthropic’s governance, the skepticism carries weight beyond typical AI critiquing.

The “Relabeled Layoffs” Argument

Ries’s most provocative claim targets the corporate AI narrative itself. “I think so many of these corporate layoffs are just the typical corporate purging. But if you put the AI label on it, then your stock goes up instead of down,” he said. If companies dress routine workforce reductions in AI language, part of the “AI efficiency” story propping up share prices may be more spin than substance.

He noted that Dario Amodei’s sister (Dario is Anthropic’s CEO) recently stated that job losses attributed to AI have not yet materialized. That observation cuts against louder voices in the debate. NVIDIA‘s (NASDAQ:NVDA | NVDA Price Prediction) Jensen Huang has called fears of AI-driven job loss “nonsense” and pointed to rising engineering hiring, while Andrew Yang has said his own company, Noble Mobile, is completely replacing junior employees with AI. Ries takes a different stance: he’s skeptical that either the productivity gains or the displacement are showing up cleanly in the data yet.

The Productivity Reality Check

On the empirical question, Ries was blunt. “It’s very, very, very early to see the impacts of AI. We see the data that many companies that have adopted AI have not seen any productivity improvements, let alone 40% across the board improvements,” he said. The entire AI capex cycle rests on an assumption of massive enterprise productivity gains. If those gains are slow or uneven, the return on tens of billions in infrastructure spending becomes a real question, echoing Sam Altman’s recent acknowledgment that there is significant waste in current AI spending.

Recently, Goldman Sachs (NYSE:GS) noted that AI-oriented companies were more active in the corporate bond market in 2025, raising questions about whether AI-fueled growth can continue to offset weaker parts of the underlying real economy.

The Valuation Skepticism

Ries’s firsthand account of Anthropic is the most striking part of his case. “I worked so hard to get people into Anthropic at a $5 billion valuation when nobody wanted to do it. And then I had those same people begging me for the chance to get in at $500, $800, $1 trillion at any price. I’m reflexively skeptical anytime people are willing to pay any price,” he said.

That “pay any price” behavior is a classic late-cycle signal. It aligns with a broader debate about a bubble in public markets, where some strategists brace for a reality check while others argue that current conditions differ from those in 1999.

Still Bullish on Anthropic

Although Ries thinks we might be in the early innings to understand AI’s impact, he’s still bullish on Anthropic. “Of the current AI leaders, I’m very bullish on Anthropic, and I think they seem like they act with the most integrity. They’ve been the most consistent in their statements. They seem the most focused and have the lowest cost basis,” he said, pointing to the company’s long-running stance on safety and containment risk.

Ries advises Anthropic on governance, so he might have a bias when praising the company. Anthropic has said roughly 80% of its code is written internally by its own AI agents. That suggests real productivity gains for AI-native companies, even as typical enterprise adopters are still searching for results.

For investors, Ries’ message is to be careful about assuming that every company discussing AI is actually seeing meaningful productivity gains. Investors should focus on measurable results rather than AI narratives, paying close attention to whether companies are growing revenue, improving margins, or increasing output because of AI.