I just wanna know how the OpenAI/Anthropic shell game works long-term. So both companies made equity deals with infrastructure providers; OpenAI on Azure, Anthropic on AWS, GCloud, and Colossus. They get a loan of compute credits and then pay for the compute with the credits. So the PaaS are effectively giving them free compute, then book it as revenue; and the AI provider lets them do inference and books that as revenue. So, it's like both types of company have a buffet, and let each other eat there for free. But somebody has to actually buy the pasta salad, with real dollars. Afaict, those real dollars are.... the cash reserves of the PaaS.
How long are they going to eat into that cash? Microsoft and AWS don't really have their own models, whereas Google and SpaceX do. And while Google has tons of cash, SpaceX is perpetually looking for cash. So the only player here that can actually afford to keep doing this, or leave the game entirely, is Google.
I know a traditional SaaS company I worked for that IPO’d years ago and still has no signs that they can be profitable (and many others like it) and nobody seems particularly concerned.
The core bottlenecks are power and computing capacity, and they actually trace back to the exact same issue. It all comes down to the physical energy it takes to flip or move a single bit inside the ram or disk storage. This concept is subject to fundamental physical barriers.
There are a few ways to tackle this, like improving power efficiency, reducing model size, or pushing hardware further. However, achieving orders-of-magnitude improvement in any of these areas will cost a massive amount of time and money. I wonder if governments, corporations, and investors have the patience to wait for these tech breakthroughs.
Yet this site suggests that tokens are very unprofitable
This will always be negative for any new business as you are effectively depreciating the assets straight away. Like if you build a hotel and deduct the cost of building it from room income - it would take years before you get the money back but may be quite profitable with GAAP accounting.
GAAP accounting (Generally Accepted Accounting Principles) is what's used for official reporting and tax returns but excludes any increases in IP value or goodwill unless there's a buyout. If you included those the likes of OpenAI or Anthropic would have done pretty well. I'm not sure there's a word for that but basically value of the business less the money that's gone in. It doesn't get reported because 'value of the business' is guesswork and can be prone to BS but is pretty important to real world outcomes. AI is probably doing well on that one. Maybe why
>Is AI Profitable Yet? NO. Everyone's Broke.
doesn't fit with the top companies on the list having many billions in the bank.
Maybe most of them or all of them lose on their bets, but there's potential for a future where revenue grows beyond the immense capex and research investments.
Oracle though... Immensely risky capex to service a startup industry with what will soon be a commodity...
[1] https://www.wired.com/story/spacex-ipo-anthropic-compute-fin...
Sure they're torching money on building consumer LLMs, but they seem to be doing very well optimizing things like ad ranking
https://engineering.fb.com/2025/11/10/ml-applications/metas-...
https://engineering.fb.com/2026/03/31/ml-applications/meta-a...
1. Outspend and outlast your competition until you have market dominance. Win over and lock in your customers with sweetheart deals.
2. Enshittify and squeeze your customers to pay back your debt.
If you're using AI, you're not paying the true cost right now because we're in phase 1. Be ready for phase 2.
Oh it doesn't fit the narrative. Never mind then.
Also many of these companies like Amazon, Google, and Meta drive a lot of incremental value due to both AI powered content suggestion and AI powered ad suggestion. Personalized ads has driven a ton of revenue.
https://www.cnbc.com/2026/05/20/anthropic-revenue-explosive-...
Only the later have something to lose if AI bubble gone by tomorrow. Everyone else will just stay with grown capacity and reuse infrastructure for whatever.
Not listing other hardware companies is just dishinest. AI is not a crypto mining where resources are just burned.