The AI hype bubble is almost over. When investors refused to fund new AI startups because they know they’re going to fail, then the AI hype bubble is starting to fail. Even the most clueless, Investors now know that this is a bubble waiting to burst like the metaverse bubble and the dotcom bubble. And nobody wants to be left, holding the back on this bubble. The era of free money iis over with interest rates sky high. When big tech companies are losing money on AI and the bigger AI startups are demonstrating that they are all dead and struggling to even have revenue, then no one is interested in funding, smaller AI startups that they don’t believe will be profitable anytime soon. People aren’t waiting for an AI unicorn to break out. They are waiting for Microsoft, Google, Amazon, and similar to start buying distressed AI startups. Collapsing companies and minimal paths to profitability doesn’t help a space that was already overhyped. Don’t forget in all this mess that the unfortunate reality is that engagement and retention don’t match installs and downloads. What I mean by this is the most basic of mobile tracking were ignored. You don’t judge success in the mobile market based on downloads or installs, but rather based on retention and engagement. What this means is, it’s not actually impressive for OpenAI to get to 100 million users quickly if 2% of them retain. The question is not how many people signed up for a free account. The question is how many monthly active users and daily active users do AI startups like OpenAI, Anthropic, Perplexity, Stability, and others have. And how many are paying customers? And what we are finding is that retention is in low single digits. The reality is that AI startups are getting users the way that mobile free to play did. And they haven’t found a way to actually make money on that. Don’t ask an AI startup how many users signed up. Ask them how many DAUs and MAUs they have. Ask them what their ARPDAU is. What you will find is that every number is low. AI startups have no idea how to do monetization, and they don’t have regular active users in any volume. What this points to is that they’re all going to fold or get bought.
Venture capitalists at this week’s Collision tech startup and investor conference in “Toronto approached the next wave of [AI] startups with increasing skepticism.” “Fears of investing in the [AI] startup boom—and feeding a bubble like the one that devastated firms in the dot-com era—have investors wary of writing checks with the fervor of prior years. “The dot-com boom of the late 1990s got messy because every venture-capital firm needed its bet in the space—leading to inflation in the costs of expenses such as hiring and office space, according to one VC. A similar dynamic is playing out now with the AI boom, investors say.” The ChatGPT boom started last year and it has broadened this year. “ChatGPT fever hit something of a peak last year,” and the “aftermath, and a sense of déjà vu, are giving them a clearer sense of what’s investible.” “Of the 1,623 startups that exhibited at Collision this year—the highest number of any Collision event—20% are building AI products,” and “that doesn’t include a large proportion of #startups that now have “AI components” in their business.” Investors put $21.8 billion into generative AI deals last year, up fivefold from the prior year. Mistral AI ($650 million), Anthropic ($2.75 billion), and CoreWeave ($7.5 billion) were the biggest recipients. Referring to the startups presenting at the conference, investors said that “Only a small number of them will survive and break through the AI #hype.” “Some said they’re increasingly looking for startups with business models that have long-term viability, products that solve corporate business problems and those with access to stores of private or unique data to train AI models.” Another said: “training large language models at the scale of OpenAI requires millions of dollars for computing and AI chips, so that’s not an area in which new startups can be competitive.” My take: The article doesn’t mention that startups in general haven’t done that well over the last 10 years. Almost 90% of America’s publicly traded unicorn startups are still unprofitable despite an average age of 15 years, and that doesn’t count the dozens that were liquidated or acquired at fire sale prices over the last two years. And if those are mostly unprofitable, an even higher percentage of privately held unicorn startups are likely unprofitable. Investors should be worried about the AI bubble, but also about past startup bubbles. They should be asking themselves why they keep investing in #startups that don’t become profitable? Are they using the wrong criteria for investing? This is one of the subjects of my forthcoming book in October. Unicorns, Hype and Bubbles: A guide to spotting, avoiding and exploiting investment bubbles in tech #technology #innovation #artificialintelligence https://lnkd.in/gwGcNXsS
M&A is looking like the most viable exit for many AI startups, to be honest. There’s nothing inherently wrong with that, but as you’re saying, Rafael Brown, without achieving retention or “stickiness,” any AI startup (or any startup, really) is bound to struggle. Who wants to buy a product/startup that can’t keep its users engaged or provide features that users actually want to use regularly?
Thony Doerga
2w
I think this is a real great analysis. I do believe ‘AI’ has an amazing future ahead in the coming decade. But it is definitely overhyped and in a bubble right now. I was honestly surprised that investors looked at downloads and installs again instead of real metrics like retention. It is as if with every tech hype they throw all former lessons out of window and have to relearn that same lesson again and again. I am excited because I believe that when the bubble bursts and the attention goes away again, that eventually the technology will get good enough to cater to real use cases.
Ashley Allen
3w
I sold my NVIDIA shares after the split (well, most) - when even the shovel makers are massively overvalued during the gold rush, it’s time to cash out. Will see if I regret this in 5 years 😆
Mike Hall
2w
The article doesn't support your hypothesis.
Paul Burchard, PhD the cracks are showing already. :)
Founder, CEO, CTO in the interim @ NEXTGEN.LX, Senior Referentin für Erwachsenenbildung @ MyGatekeeper Hannover, Gastberaterin @ BDU/ Fachbereich Personalentwicklung & Coaching
2wMy brilliant former college and data-scientist, Ben Cowell, wrote a fantastic piece, reviewed a few hours ago by Prime https://www.youtube.com/watch?v=k0XuoK132z4 Here is the original script btw: https://ludic.mataroa.blog/blog/i-will-fucking-piledrive-you-if-you-mention-ai-again/ AI is still developing (a little), but is largely seems to have plateau-ed and we may have to wait for another 10 years to witness some substantial increment. Another chatbot might not prepare you for the future 😊