Anthropic recorded a staggering $80B net loss in the first half of 2026

According to that Reuters article $72B of the $80B loss is an "accounting charge reflecting an increase in the estimated value of bonds that could eventually convert into Anthropic shares," i.e. they borrowed money in return for stock options. That's bad for Anthropic investors, but essentially free for the company.

Their revenue "more than doubled from the first quarter to $11.6 billion in the second quarter" which is less than almost tripling so first quarter revenue might've been $11.6B/2.5 ≈ $4.6B for $16.2B total revenue in the first half. To get that revenue, they "spent nearly $9 billion on computing capacity and infrastructure in the first half." Scaling up for the $16.8B third-quarter revenue suggests $9.3B compute spend. Meanwhile to get the $8B remaining first-half loss, they must've had $24.2B expenses, $15.2B of which are not compute, so about $7.6B per quarter. If that stays fixed, total third-quarter expenses should be about $16.9B, fairly close to revenue. I think reporting that their second quarter was profitable focused on the operating income exceeding operating expenses only.

So there are actually several senses in which Anthropic could be profitable or fail to be:

1. Are operations bringing in more money than they cost? (Supposedly already happened.)

2. Are operations bringing in enough money to fund capital expenses? (Maybe close, maybe barely so.)

3. Does investing in Anthropic yield good returns?

Note that the last point crucially depends on their stock price. If it drops a little, those "bonds that could eventually convert into Anthropic shares" start showing up on the positive side of the balance sheet instead. Insofar as the stock market is efficient, if they IPO, it will be priced such that it's no longer obvious whether it's better or worse than any other investment. But early investors and founders will make bank.

Ahhhh I had someone parroting to me "anthropic is profitable now" in real life the other day. I guess they got it from hn.

Just a few years ago there was perhaps just as great a crisis in enterprise compute pricing. The end of "unlimited" storage accounts for enterprise customers. No one ran from gsuite and microsoft et al. They still kept paying for all the subscriptions to these sometimes redundant suites of services as they always did, only they now pay for additional storage solutions (in whatever shape those ended up manifesting as) as well.

> And we haven’t even got to the part where AI consumer pricing corrects to realistic models. Once that hits, companies will drop AI like a hot potato.

That ship has sailed. If you're a mid-sized company, with the resources of a mid-sized company, you'll always be able to download Deepseek's new Flash model and run it on your own hardware.

With all of the competition out there, prices arguably can't rise much further than they already have.

Well it's from a prediction market so could just as easily flip to like 5% a few minutes after a suitable story in the news breaks.

Indeed. We eval models regularly on their ability to record the amenities an apartment building has (ex. Fireplace, In-Unit Laundry, Hardwood Floors etc...). Classic structured data extraction task.

This allows https://fless.io to match what a renter is looking for without looking up a building's site (aggregators like zillow frequently have incorrect or non-unit specific amenities).

The cost at API rates is astounding.

An order of magnitude or more what coding plans go for. And everything is going up - coding plans and API rates both. Chinese or American frontier. Everything up.

API rates, currently, are so expensive for this particular workload that only self hosted or human classification are viable despite improvements in model capabilities.

Capabilities for this task at the current prices (even with claimed $ / token reductions) are nowhere close to viable as API in a business sense - just outsource the classification or host models yourself.

The implied valuations of Oai and Athr dictate either massive price increases or step changes to model performance.

Given model performance gains have been decidedly linear this leaves only price increases to fit the goal.

Not clear to us given that a basic task is barely price competitive as is, while requiring open weights and self hosting, that the valuations can be justified.

Has Anthropic been taking on debt? I thought they've been primarily selling equity.

Why would you buy the dip of an unprofitable enterprise?