Anthropic CEO Says 2026 Will Have a Radical Acceleration That Will Surprise Everyone

Anthropic CEO Dario Amodei stated at the Morgan Stanley Technology, Media and Telecom (MS TMT) Conference that 2026 could bring a dramatic acceleration in the progress of artificial intelligence that may catch many observers off guard.

“We do not see hitting the wall. I think this year is going to have a radical acceleration that surprises everyone,” Amodei said during the conference, as per excerpts from his talk shared by Apoorv Agrawal, partner at Altimeter Capital on X, and a Substack post on TMT Breakout. 

Explaining how AI development may accelerate, he pointed to the nature of exponential growth. “Exponentials catch people off guard—there’s the old parable of the second half of the chessboard, where you have one grain of rice in the first square, two on the second, four on the fourth,” he said.

Amodei argued that current AI progress may already be well into that exponential curve. “We’re standing on square 40 out of 64, and from square 40 to square 64, it’s going to go faster than you think—even having seen how fast it’s gone so far,” he said.

“I don’t think people are ready for it. I think we are on the precipice of something incredible.”

Amodei, despite the debate and growing scepticism around whether scaling laws are hitting limits, has maintained that they continue to hold. In a recent podcast with Dwarkesh Patel, he argued that both pre-training and reinforcement learning still show clear scaling trends, indicating that increasing compute and data continue to translate into stronger AI capabilities.

At the MS TMT conference, Amodei also pointed to coding as the fastest area of enterprise adoption for AI systems. “The standout winner—I guess this one’s pretty obvious—is code. We actually had a thesis that this would be the thing to focus on and that it would go fast, but I think it’s exceeded even our very high expectations,” he said.

He added that the technology is already reshaping productivity within Anthropic itself. “We’ve reached the point where it doubles and triples the rate at which we’re able to produce new things end-to-end,” Amodei said, referring to how AI tools are being used internally across development workflows.

He stated that the rapid progress in AI has also intensified competition for top talent in the sector. “Last summer, a competitor was trying to hire our researchers by paying numbers ranging from $100 million to $500 million—prices that greatly exceed those of pro athletes,” he said.

He argued that retaining talent ultimately comes down to culture and mission, not compensation alone. “We said to everyone: you’re here for the mission. We’re not going to give you 10 times as much as you’re earning right now just because someone threw a dartboard and your name came up—that’s very divisive and destructive for the culture,” he said. 

“At the end of the day, we lost two employees to Meta. The best comparison: OpenAI lost several, and they’re maybe 1.5x as big as us—so even normalised by size, there’s a 10x or 20x difference. We are just clearly doing something different,” Amodei added. 

The comments come as Anthropic is experiencing rapid business growth. According to reports, the company’s annualised revenue run rate has climbed to nearly $20 billion, more than doubling since late 2025. 

The report said Anthropic has recently reached about $19 billion in annualised revenue, up from roughly $9 billion at the end of 2025 and around $14 billion only a few weeks earlier. 

The growth has been fueled largely by increasing adoption of the company’s AI models and developer tools, including its coding-focused product Claude Code, as demand for AI systems continues to expand.

The company recently raised $30 billion at a valuation of $380 billion.