AI Bubble Debate: Are We Repeating the Dot-Com Era?
Recently, Sam Altman, CEO of OpenAI, stated that he believes we are currently in an AI bubble, comparing today’s hype to the dot-com bubble of the late 1990s (1995–2000). That period resulted in losses of up to 80% among high-value tech companies as excitement outpaced realistic business models. Web start-ups launched rapidly, often with infeasible or unsustainable strategies.
Are we in a phase where investors as a whole are overexcited about AI? My opinion is yes. Is AI the most important thing to happen in a very long time? My opinion is also yes.
Sam Altman
Arguments for the AI Bubble Theory
Overinvestment
Large amounts of capital are being poured into AI start-ups with little or no revenue, with valuations driven by hype and potential rather than fundamentals. OpenAI has been valued at $300 billion (March 2025), while smaller AI companies such as the French firm Mistral are valued at roughly £13 billion despite offering similar capabilities.
Try Mistral AI
Overpromising
The release of GPT-5 was perceived by some as disappointing after significant anticipation on social media. OpenAI acknowledged issues such as glitches and routing problems that caused weaker performance on certain tasks, with earlier models sometimes outperforming it.
History repeating itself
Many companies are entering the AI space without clear long-term, sustainable business models — a pattern reminiscent of the dot-com era.
More about the dot-com bubble
Arguments Against the AI Bubble Theory
Widespread adoption
Around 52% of UK businesses now use AI, with 92% reporting increased revenue according to an AWS report. Consider everyday online services — most now include some form of AI functionality.
Infrastructure investment
Not only big tech firms but also a wide range of organisations are investing heavily in AI-ready infrastructure. Some build their own systems, while others rent infrastructure from providers such as Amazon or Google.
Real gains
Unlike the dot-com era, where many companies lacked functional products, modern AI tools already deliver measurable improvements. Industries such as software engineering, healthcare, finance, and logistics are seeing tangible productivity gains, and by 2025 these tools are widely accessible on personal devices.
Conclusion
While there are signs of hype and potential overvaluation reminiscent of the dot-com bubble, AI development is producing measurable real-world benefits. Widespread internet adoption and ongoing infrastructure investment suggest genuine market readiness, indicating that AI may represent a lasting structural shift rather than a temporary speculative trend.