From conversational chatbots to recommendation systems and content creation tools, artificial intelligence is everywhere, especially since the release of ChatGPT in late 2022.
Investments in AI are exploding, with record-breaking fundraising for companies like OpenAI, Perplexity, and Anthropic. But this frenzy reminds many of the dot-com bubble of the 2000s, which culminated in a resounding financial crisis.
What if AI followed the same path?
The Yiaho team had some fun comparing the two periods and analyzing the risks and opportunities, while keeping an eye on the players that might survive a potential storm.
The 2000 bubble: a historical precedent
In the late 1990s, the internet was the new technological frontier, much like AI in 2022.
Startups like Pets.com or Webvan raised hundreds of millions of dollars on the mere promise of a digital future, often without a viable business model. Investors, driven by euphoria, injected massive capital into companies with stratospheric valuations.
In 2000, the bubble burst: thousands of startups went bankrupt, stock markets plummeted, and only a handful of solid players, like Amazon, Google, or Cisco, survived to become today’s giants.
The parallels with current AI are striking.
Just like back then, expectations around AI are immense: it’s presented as the solution to every problem, from business productivity to scientific discovery. But this hype is accompanied by valuations that are sometimes disconnected from economic reality. Many AI startups promise revolutions without yet demonstrating clear profitability—a direct echo of the excesses of 2000.
Massive fundraising in AI: a risky craze?
The numbers are staggering:
- OpenAI, creator of the latest ChatGPT 5, raised $6.6 billion in October 2024, reaching a valuation of $157 billion, surpassing many established companies.
- Perplexity, a startup focused on AI-powered search, raised $250 million in 2024, with a valuation of $3 billion.
- Anthropic, founded by former OpenAI researchers, secured an additional $750 million that same year.
- Other emerging players, like xAI (creator of Grok) or Mistral AI in Europe, are also attracting massive investments.
This fundraising reflects overflowing confidence in AI’s potential, but it raises questions. Many of these companies depend on models that are expensive to train and maintain, with infrastructure requiring billions of dollars in hardware (notably Nvidia’s GPUs).
Yet, the revenue generated by AI often remains modest compared to the investments.
OpenAI, for example, forecasts losses of $5 billion in 2024, despite its breakthroughs. As in 2000, the gap between promises and economic reality could weaken some startups.
The survivors of 2000: a lesson for AI?
The dot-com bubble showed that the companies best prepared to survive are those with solid foundations: diversified revenue, a long-term vision, and an ability to pivot.
In 2000, companies like Microsoft, Google, and Amazon weathered the crisis thanks to their financial resilience and adaptability.
Today, these same giants dominate the AI landscape. Microsoft, with its strategic partnership with OpenAI and Azure, is a leader in enterprise AI. Google, with models like Gemini, is investing heavily while relying on its advertising revenue.
Meta, which survived the dot-com bubble as a young company, is betting on AI to strengthen its social networks and its ambitions in the metaverse.
These well-established players have a key advantage: they can absorb the shocks of a potential bubble burst. Their diversification and cash reserves allow them to continue investing in AI, even in the event of an economic slowdown.
Conversely, recent startups like Perplexity or Anthropic, though promising, could be more vulnerable if investors become more cautious. And the slightest mistake or bad buzz, such as the use of Perplexity for Donald Trump’s social network, could permanently damage their growth.
The risks of an AI bubble
Several warning signs are reminiscent of 2000.
- Hyper-dependence on a few key players, like Nvidia for chips or AWS for cloud infrastructure, creates bottlenecks.
- Fierce competition between startups and tech giants could lead to an unsustainable race for innovation, where companies burn through capital just to stay in the running.
- Unrealistic expectations from investors, who anticipate quick returns, could clash with reality: the development of generative AI or Artificial General Intelligence (AGI) will take time and colossal resources.
An AI bubble burst could have significant repercussions. A drop in startup valuations could slow innovation, while layoffs in the tech sector would worsen an economic crisis.
But just like the post-2000 era, such a crisis could also purge the market of less viable players, making room for companies capable of creating real value.
What if the AI bubble doesn’t burst?
It’s also possible that AI isn’t a bubble, or at least… not in the classic sense!
Unlike the internet of the 1990s, which was in its infancy, AI relies on technologies already integrated into many sectors: healthcare, logistics, finance, etc.
Companies like Tesla use AI for vehicle autonomy, while pharmaceutical giants use it to accelerate drug discovery. These concrete applications could anchor AI in the real economy, reducing the risk of a total collapse.
Furthermore, AI benefits from institutional and governmental support that the internet didn’t have in 2000.
The United States, China (with DeepSeek, for example), and Europe are investing heavily in AI for strategic reasons, which could cushion a crisis. Finally, the growing demand for AI solutions in businesses suggests that the market could continue to grow, even if some startups disappear.
Who will survive the AI era?
Our platform Yiaho, obviously! But not just us… The history of the dot-com bubble teaches us a clear lesson: only companies with solid fundamentals and a long-term vision thrive after a crisis. In the current AI landscape, giants like Microsoft, Google, and Meta seem well-positioned to dominate, thanks to their resources and diversification.
Startups like OpenAI, Perplexity, and Anthropic will have to prove their ability to generate sustainable revenue to avoid the fate of the Pets.coms of 2000.
Is AI a bubble? Maybe. But even if a correction comes, it could be the catalyst for a more mature industry, stripped of speculative excesses. For now, enthusiasm around AI remains at its peak. It remains to be seen whether this craze will lead to a lasting revolution or a painful return to reality. One thing is certain: the survivors of this frantic race will be those who can combine innovation with pragmatism!
