Anthropic's sales soar to $16.3 billion, but most firms see no AI payoff

Anthropic, the maker of the Claude AI assistant, reported $16.3 billion in revenue for the first half of this year. That is up from $1.2 billion in the same period of 2025. The numbers come from its filing ahead of a stock market debut. Around the same time, a well-known economist argued that big US companies outside tech have seen no real payoff from AI.

What the filing shows

The figures appear in an S-1, the document a company files when it plans to go public. Anthropic has raised $141 billion so far and still holds more than $112 billion in cash. Its rumored valuation is close to $1.5 trillion. If the IPO goes ahead, it would be the largest ever, bigger than SpaceXAI's by billions of dollars.

The jump is tied to Claude Code, a tool that helps write software. It kicked off an AI coding boom in early 2026. Companies from Google to OpenAI then shifted largely toward business customers.

The costs ahead

Anthropic has committed to spending hundreds of billions of dollars with Google, Amazon, Microsoft and SpaceX for cloud computing. It expects to be short on computing power over the next year. Big cloud providers are putting their own AI work first. That may push Anthropic toward newer, possibly less reliable providers, with data centers largely powered by natural gas to keep costs down.

Outside tech, the payoff is missing

Torsten Slok, chief economist at Apollo, wrote a post that went viral last week. He says the seven biggest tech stocks are propping up the US economy. The other large companies, which he calls the S&P 493, are not seeing rising profits or meaningful productivity gains. His post says AI has delivered "zero financial payoff" to corporate America outside of tech.

He also notes that 80% of the top 1% of AI spenders in the US are VC-backed, meaning funded by venture capital investors. That suggests the newest AI models still serve Silicon Valley more than Main Street.

What to do

Both things can be true. The AI labs are making a lot of money, while many of their customers have yet to see returns. For a business paying for AI tools, that argues for a small trial first. Measure the hours saved or costs cut before spending more. The material does not say when the IPO will happen.