Anthropic nears 17 billion in quarterly revenue and loses 23.3 billion: the IPO will test the cost of the frontier

By Bloomberg · November 2, 2026.

Anthropic has reported, according to Bloomberg, revenue of nearly $17 billion in the third quarter, which ended September 30. That is a twelvefold increase over the same period a year earlier. The figure comes from the updated version of its filing for an initial public offering (IPO), which Bloomberg says is just weeks away.

The other side of the ledger is just as striking. According to the same report, the company posted a net loss of $23.3 billion in the quarter. Its operating expenses rose to $17.5 billion, up 292% from a year earlier. In other words, it spent more than it took in on operations alone, and the net loss was even larger than revenue.

The company also disclosed, again according to Bloomberg, that it expects to spend more than $895 billion over the coming years. The article does not break down where that money will go, although its own IPO filing frames the issue in terms of rising computing costs. Bloomberg adds that the filing reflects growing ambition at a time when Wall Street is showing more concern and revenue growth is slowing.

The figure deserves careful reading. Multiplying revenue twelvefold in a year is an extraordinary pace, but the slowdown Bloomberg points to indicates that this pace is no longer where it started. With such a large base, sustaining rates like these becomes mathematically harder and harder. In addition, the available material details neither sequential growth nor the composition of revenue, so we cannot say how much comes from enterprises, subscriptions or the API (the interface that lets other programs use the model).

**Our take.** This IPO is, above all, an experiment in the economics of the frontier. Training and serving top-tier models demands capital on a scale no previous tech industry had needed so early. Anthropic's implicit thesis is that every dollar of compute turns into capability, capability into demand and demand into revenue. This quarter's numbers support half of that thesis, demand, and leave the other half, profitability, open.

The short-term risk is real and should not be sugarcoated. If public markets turn skeptical of committed spending in the hundreds of billions, pressure to monetize could push the company to prioritize immediate revenue over caution, and that matters for a company whose positioning rests on safety. There is also a concentration effect: few companies can afford to play at this scale, which deepens dependence on a handful of providers. As industry context, this is the same dilemma facing its direct rivals.

In the long term, however, there are grounds for down-to-earth optimism. The fact that customers are willing to pay nearly $17 billion in three months shows that AI already generates measurable value, and that base is what funds the path toward systems capable of accelerating medical and materials research. The abundance we advocate requires exactly this kind of colossal investment, and it also requires that it eventually become sustainable. The coming weeks will tell whether investors buy that promise or demand proof of margins before backing it.