[DigitalToday reporter Chi-gyu Hwang (황치규)] Anthropic’s explosive growth shows how quickly demand for large language models can emerge. Some say it is risky to assume the trend will continue.

Anthropic generated $4.8 billion in first-quarter revenue, Reuters Breakingviews said. Second-quarter revenue is expected to nearly double to about $10.9 billion. It is forecast to post adjusted operating profit of $559 million, marking its first profit. The figure includes the cost of training new models but excludes stock-based compensation.

Still, it must spend heavily to train and operate chatbots, so additional losses are expected for several years. xAI, led by Elon Musk, was shown to have spent $2.4 billion on research and development in the first quarter through SpaceX listing documents. That is about three times its revenue of $818 million.

Simply comparing revenue at Anthropic and OpenAI is also not easy.

Anthropic reports gross revenue without excluding the share paid to distribution partners. OpenAI reports net revenue after subtracting revenue shared with Microsoft. That means the $5.7 billion for the first quarter disclosed by Sam Altman and the figures cited by Amodei are different in nature.

AI demand is surging, but large corporate customers are still in an experimental stage. Block carried out large-scale layoffs, citing chatbot-driven productivity gains, but Uber used up its annual AI budget in April. Microsoft is also cutting back Claude Code as token costs rise faster than budgeted.

On a conservative basis, Anthropic’s annualised revenue is about $31 billion. The touted corporate value of $900 billion is about 30 times revenue. AI defence company Palantir trades at 41 times revenue, but its operating profit margin is about 60%. By comparison, Anthropic’s projected margin is only 5 percent in the second quarter. If costs keep rising, even that level could be hard to maintain. Breakingviews said such a valuation still remains far from reality.