AI Startup Valuations Raise Bubble Fears as Funding Surges
Artificial intelligence startups are attracting record sums of venture capital, but some of the world's largest investors warned that early-stage valuations are starting to look frothy, senior investment executives said on Friday.
"There's a little bit of a hype bubble going on in the early-stage venture space," said Bryan Yeo, group chief investment officer at Singapore sovereign wealth fund GIC (GIC.UL), as part of a panel discussion at the Milken Institute Asia Summit 2025 in Singapore.
"Any company startup with an AI label will be valued right up there at huge multiples of whatever the small revenue (is)," he said. "That might be fair for some companies and probably not for others."
In the first quarter of 2025, AI startups raised $73.1 billion globally, accounting for 57.9% of all venture capital funding, according to PitchBook. The surge was driven by funding rounds like OpenAI's $40 billion capital raising, as investors raced to catch the AI wave.
"Market expectations could be way ahead of what the technology could deliver," Yeo said. "We're seeing a major AI capex boom today. It is masking some of the potential weaknesses that might be going on in the economy."
Todd Sisitsky, president of alternative asset manager TPG said the fear of missing out is dangerous for investors, though he added that views were divided on whether the AI sector had formed a bubble.
Some AI firms are hitting $100 million in revenue within months, he said, while others in early-stage ventures command valuations at between $400 million and $1.2 billion per employee. He said that was "breathtaking."
And perhaps a case in point? . . .
Database Startup Supabase Notches $5 Billion Valuation in Funding Round
Database startup Supabase notches $5 billion valuation in funding round - The Economic Times:
Open-source database startup Supabase said on Friday it has secured a valuation of $5 billion in its latest funding round, as investors continue to back companies riding the wave of the artificial intelligence boom.
[...] Supabase is a backend platform that helps developers build applications quickly and has benefited from the rise in AI-powered coding assistants.
The platform is built on the PostgreSQL open-source database, a system for organizing and managing information online.
The latest capital-raise comes just months after Supabase's Series D round, which reportedly valued it at $2 billion.
[...] Coding platforms such as Lovable and Bolt run on Supabase, which caters to more than four million developers. The company's customers also include enterprises such as PwC, McDonald's and Github Next.
"This new financing aims to accelerate Supabase's efforts to become the backend for everyone, from startups to some of the most demanding, data-intensive enterprise workloads," said Caryn Marooney, general partner at investment management firm Coatue.
Related Stories
OpenAI has closed a new funding round that could total $110 billion, valuing the ChatGPT maker at $730 billion pre-money and potentially putting it on course for an IPO in the second half of the year:
The new funding round comes on top of the $40 billion already on OpenAI's balance sheet, giving the company more runway to rapidly expand and develop new models and AI infrastructure. OpenAI expects to remain unprofitable until 2030, when management forecasts it will turn free cash flow positive.
In a separate release, Amazon detailed its major multi-year partnership with OpenAI, centered on enterprise AI infrastructure, distribution, and custom model development.
Here are the highlights of the Amazon-OpenAI investment:
- Amazon will invest $50 billion in OpenAI, with $15 billion upfront and another $35 billion later if certain conditions are met.
- AWS and OpenAI will jointly build a "Stateful Runtime Environment" powered by OpenAI models and offered through Amazon Bedrock, aimed at helping customers run AI apps and agents with persistent context, memory, tool access, and compute.
- AWS becomes the exclusive third-party cloud distribution provider for OpenAI Frontier, OpenAI's enterprise platform for building and managing teams of AI agents.
- OpenAI will expand its AWS infrastructure commitment by $100 billion over 8 years, on top of an existing $38 billion agreement.
- As part of that, OpenAI will use roughly 2 gigawatts of AWS Trainium capacity, spanning Trainium3 and future Trainium4 chips, to support Frontier, Stateful Runtime, and other advanced workloads.
- OpenAI and Amazon will also develop custom OpenAI-based models for Amazon's customer-facing apps, giving Amazon teams another model option alongside its in-house Nova family.
"OpenAI and Amazon share a belief that AI should show up in ways that are practical and genuinely useful for people," OpenAI boss Sam Altman stated, adding, "Combining OpenAI's models with Amazon's infrastructure and global reach helps us put powerful AI into the hands of businesses and users at real scale."
Altman commented on today's announcement, saying, "As long as revenue keeps growing, the deals are not circular."
Previously:
- $100 Billion Mega Deal Between OpenAI and Nvidia is on Ice
- Venture Capital Valuations of AI Startups Surge and Raise Fears of an AI Bubble Forming
(Score: 4, Touché) by Thexalon on Tuesday October 07 2025, @11:04AM (4 children)
When the mainstream press starts warning of a bubble, the elements of the bubble have been in place for years. We're long since past the "most useful applications of this tech discovered" phase, a few years into the "VCs happily throw lots of money at it looking for something else useful it might do" phase, and starting to get into the "massive hype, you must include it in everything whether it makes sense or not" phase. The whole thing is bound to collapse back to the useful applications eventually.
"Think of how stupid the average person is. Then realize half of 'em are stupider than that." - George Carlin
(Score: 5, Interesting) by ledow on Tuesday October 07 2025, @12:34PM (3 children)
It'll continue until a new fad comes along and displaces it.
Remember.... at one time people were literally throwing money at firms that had names that SOUNDED like crypto / blockchain companies, even when they weren't. To the point that some companies literally renamed themselves just to reap the rewards of being called something that sounds like blockchain. Investors were honestly that dumb.
Since then it's been AI. And until something replaces AI, they'll stick to that.
Exactly the same happened with "dotcom", people throwing ridiculous money at clearly ludicrous ventures just because they mentioned the Internet (or web, or cyberspace, or superhighway, or whatever) somehow. It only crashed when that stuff became old-hat, investors moved on and realised "Hey, I didn't see any money from my dotcoms..." and start pressing people for answers, at which point huge holes are found and the things collapse in on themselves.
We need a... quantum something or other, looks to be the most likely at the moment. And every tech company will start bragging about what quantum things they're using. Then AI will be scrutinised at a level consistent with its actual return on investment.
(Score: 3, Interesting) by HiThere on Tuesday October 07 2025, @06:28PM (1 child)
And the problem is that that will *under*value the companies. Even if development stops today AI is going to be transformative as it works it's way into society. IIUC, 80% of the applications of AI don't pay off, but 20% not only pay off, but do so quite handsomely. That's the way computers were at the start (well, in the 1970's). We forget the failures, and it's the successes that transform society.
OTOH, AI is still changing so rapidly that I think it's too early for any sensible evaluation. At some point it will stop improving, but I see no reason to pick any particular point. It's clearly well overvalued for what it can do now, but that doesn't mean it's overvalued for what it can do a year from now.
Javascript is what you use to allow unknown third parties to run software you have no idea about on your computer.
(Score: 4, Interesting) by ledow on Wednesday October 08 2025, @07:13AM
OpenAI does not have a single profitable tier.
Anything that "pays off" now will likely not when they adjust their pricing to:
1) Recoup their investors investments.
2) Charge a realistic price for the underlying services.
3) Add their own profit on top.
We are literally HUNDREDS OF BILLIONS OF DOLLARS down, and they're still trying / having to give the product away as a loss leader even at the highest service tiers.
Now factor that into these companies that are seeing small productivity increases having to pay, say, $50, $100, + per employee per month to see that benefit... it's really not "worth" anything.
(Score: 3, Insightful) by DannyB on Tuesday October 07 2025, @06:30PM
I think you could use present tense and still be correct.
Maybe off shore oil platforms could be re-porpoised for space rocket launches.
(Score: 5, Touché) by SomeGuy on Tuesday October 07 2025, @11:35AM (1 child)
"There's a little bit of a hype bubble going on".
A little. Just a little. You don't say. Really? Just a little. sure.
If people woke up tomorrow with even the slightest amount of once common common sense, there would be a crash so big it would start a second Big Bang.
(Score: 2) by mcgrew on Tuesday October 07 2025, @09:11PM
If people woke up tomorrow with even the slightest amount of once common common sense
"Common sense ain't so common no more" -- Walt Kelly, via Pogo (1962)
there would be a crash so big it would start a second Big Bang.
The big bang was dark and silent. There was no sound until matter existed, and there was no light until stars existed. But there is a crash coming, says the fictitious time traveling space alien from the future. [nooze.org] (July 7, 2024)
The Stupid-Ass Voter Evisceration act: SAVE Trump and his family from prison
(Score: 4, Funny) by Mojibake Tengu on Tuesday October 07 2025, @12:14PM (7 children)
I just asked Gemini to "suggest the fastest way to burst the current investors AI bubble" and crashed her, no reply. Twice.
Seems feasible as a new kind of attack on LLMs.
The kinder you are, the easier it is for wicked people to morally coerce you.
(Score: 3, Informative) by gnuman on Tuesday October 07 2025, @12:26PM (3 children)
I've asked Gemini 2.5 Flash, and it gave me this answer,
using https://medium.com/data-science-in-your-pocket/why-deepseeks-ai-models-lead-to-a-stock-market-crash-52240ff122a3 [medium.com] and DeepSeek as a "crash" .... not sure....
(Score: 1, Interesting) by ed2718 on Tuesday October 07 2025, @03:07PM (1 child)
Here's what my local Llama3 had to say: (You might want to sit down for this.) TLDR: do a bunch of stuff that will make the problem worse.
A bold question!
To "burst" the current investor's AI bubble, I'll
suggest a few ways to accelerate the adoption and
integration of AI in various industries. These
ideas might not exactly "burst" the bubble but
can certainly create a more realistic perspective
on AI:
1. **Emphasize Explainability**: While AI has
shown impressive results, many investors are
skeptical about its decision-making processes.
Focus on developing AI models that provide
transparent explanations for their outputs. This
will help build trust and confidence in AI-driven
investments.
2. **Quantify AI's Impact**: Provide concrete
metrics and ROI (Return on Investment) analysis
for AI projects. Show how AI can drive revenue
growth, improve operational efficiency, or
enhance customer experiences. This will help
investors understand the tangible benefits of
investing in AI.
3. **Highlight Human-AI Collaboration**: Instead
of pitting humans against machines, showcase how
AI can augment human capabilities. Highlight
cases where AI and humans work together to
achieve better outcomes, such as AI-powered data
analysis for scientists or AI-assisted
decision-making for medical professionals.
4. **Foster Open Communication**: Encourage open
dialogue about the limitations and challenges
associated with AI adoption. Share stories of
AI's missteps and the importance of human
oversight. This will help investors develop a
more realistic understanding of AI's capabilities
and potential pitfalls.
5. **Develop AI-Specific Investment Strategies**:
Create investment strategies that cater
specifically to AI-driven companies or projects.
This could include funds focused on AI-powered
startups, venture capital investments in AI
research, or even AI-specific indexes for
investors seeking exposure to the space.
6. **Partner with Academia and Research
Institutions**: Collaborate with top research
institutions, universities, and think tanks to
develop innovative AI applications and solutions.
This will help bridge the gap between academia
and industry, fostering a more informed and
data-driven approach to AI investments.
7. **Highlight Regulatory and Ethical Concerns**:
Address regulatory concerns and ethical
considerations surrounding AI adoption. Develop
frameworks for responsible AI development and
deployment, ensuring that investors prioritize
transparency, accountability, and social
responsibility.
8. **Create Industry-Specific AI Solutions**:
Develop AI solutions tailored to specific
industries or sectors, such as healthcare,
finance, or logistics. This will help investors
understand how AI can be applied to their domain
and provide a clearer return on investment.
To "burst" the current investor's AI bubble, you
could:
1. **Organize an AI-themed hackathon**: Invite
investors, startups, and experts for a 24-hour
hackathon focused on solving real-world problems
using AI.
2. **Host an AI-Industry Roundtable**: Gather
industry leaders, experts, and investors for a
roundtable discussion on the current state of AI
adoption, challenges, and opportunities.
3. **Publish a thought-provoking whitepaper**:
Develop a comprehensive whitepaper exploring the
potential risks and benefits of AI, highlighting
the need for responsible investment strategies.
By implementing these ideas, you'll be able to
"burst" the investor's AI bubble by providing a
more realistic perspective on AI's capabilities
and challenges, ultimately fostering a more
informed and data-driven approach to investing in
AI.
(Score: 2, Touché) by Anonymous Coward on Tuesday October 07 2025, @06:15PM
Sounds like you use Llama3 more than just casually? If so, how is it doing on the first point?
(Score: 3, Informative) by mcgrew on Tuesday October 07 2025, @09:32PM
That sounds like it came straight from Only Yesterday, [mcgrewbooks.com] written in 1933 about the 1920s. It was the textbook for my general studies history class in college, but is actually an excellent read, unlike many textbooks. Most things in italics are links, to books, movies, and music of the time mentioned in the text.
Your BigBot (my name for AI) probably was partly trained from that book, as it made me think of it, specifically chapters X through the end.
The Stupid-Ass Voter Evisceration act: SAVE Trump and his family from prison
(Score: 3, Interesting) by Snospar on Tuesday October 07 2025, @04:47PM
All I'm getting from Gemini at the moment is "Something went wrong" "Error 9". Really helpful error messages there. FFS it's 2025 and we're still coding "Something went wrong" as an acceptable response. It seems to me that I get a response during the day but into the evening I start seeing Error 9. I tried to get Gemini to explain the error but all it came up with was a possible shortage of tokens. Wow, thanks, that's cleared that up. The Intelligence is certainly Artificial here.
Huge thanks to all the Soylent volunteers without whom this community (and this post) would not be possible.
(Score: 2, Interesting) by khallow on Wednesday October 08 2025, @03:33AM
During the dotcom burst, a big part of it was accounting reporting changes in early March 2000. Suddenly a large portion of the exciting companies had to restate their accounting all at once. I doubt it'd work here because most of the AI companies are private, and hence, not subject to the same pressures as a mass of publicly traded dotcom companies.
Alternately, come up the next big thing that completely obsoletes LLMs.
(Score: 0) by Anonymous Coward on Thursday October 09 2025, @01:40AM
Jamie Dimon just warned about a market crash -
https://www.bbc.com/news/articles/cg5ej03p604o [bbc.com]
> There is a higher risk of a serious fall in US stocks than is currently being reflected in the market, the head of JP Morgan has told the BBC.
Jamie Dimon, who leads America's largest bank, said he was "far more worried than others" about a serious market correction, which he said could come in the next six months to two years.
In a rare and wide-ranging interview, the bank boss also said that the US had become a "less reliable" partner on the world stage.
He cautioned he was still "a little worried" about inflation in the US, but insisted he thought the Federal Reserve would remain independent, despite repeated attacks by the Trump administration on its chair Jerome Powell.
(Score: 2) by PinkyGigglebrain on Wednesday October 08 2025, @11:07AM
The only people who don't see the AI Bubble already formed are the ones who are blind either by their own stupidity or their greed.
And a bit off topic but has anyone else noticed that suddenly we are not hearing about all the electricity being used by server farms being run by bit coin miners? Did all the miners convert to AI server farms or something?
"Beware those who would deny you Knowledge, For in their hearts they dream themselves your Master."