Ask HN: How likely it is that the AI bubble will pop next year?

7 points·laurentiurad··6 comments

I think a pivotal moment will be the launch of GPT-5 and deepseek R2. If these models will not be exponentially better than the current frontier models, I doubt that investors will still pour money into the current AI model.

How likely do you think it is that this comes to an end soon? IMHO current valuations are ridiculous indicating we have reached the peak.

View on HN

OpenAI, Anthropic, Meta, Google are are making $100M - billions in revenues.

Wouldn't call that a bubble. Sure some companies may run out of fuel but the Hyperscalers got a huge warchest to deploy for decades.

Revenue is one thing but profit is the most important metric here. I remember when the dot com bubble popped. Back then everything you need to have was a landing page to raise 200 million, but now it's the same to get 2 billion (see this example: https://www.perplexity.ai/discover/finance/mira-murati-s-thi...). It's really the same case but due to the inflation we see higher numbers.

I depends how you define a bubble popping. Stock prices could fall but I think AI progress will crack along. It's quite close to moving from chatbots to doing real work.

Same as it was for dot com bubble. Internet didn't end there but it generated a lot of job losses and economic slowdown.

1) CEOs are still dreaming about replacing labor with AI

2) AI still hasn't been integrated into everything yet (agentic)

3) Nvidia is promising 1m x compute over the next 10 years.

Seems far from over.

sure, CEOs act as sales agents, but they still rely on investors' money. If they stop believing, then it's all over.

Reply on news.ycombinator.com