From Dotcom Mania to the AI Gold Rush: Lessons for the Next Tech Cycle

In 1999, a sock puppet became the mascot of Pets.com — a company that would go public, burn through $300 million, and collapse within 268 days. It was a symbol of the era's unbridled optimism about the internet.

Fast forward to 2025, and AI has become the new dotcom. Capital is flooding into generative AI, valuations are skyrocketing, and companies are scrambling to brand themselves as "AI-first." But history is whispering a warning: technology revolutions may be inevitable, but their early stages are often messy, overhyped, and painful for unprepared businesses.

When the Past Rhymes: Parallels Between 2000 and 2025

The similarities between the dotcom bubble and today's AI boom are striking:

• Valuations Based on Potential, Not Profits: Many AI start-ups are valued in the billions despite no clear path to monetisation.
• Hype Everywhere: Just as every company once became an "e-business," today every company is "AI-powered."
• Fear of Missing Out: Boards feel pressured to announce AI strategies quickly — sometimes before they even know what problems AI will solve for them.
• Infrastructure Catch-Up: Just as the early internet lacked broadband, today's AI revolution is constrained by compute shortages, high energy demands, and limited guardrails.

The Talent Arms Race

Unlike 2000, today's AI boom is accompanied by an unprecedented competition for talent.

• Skyrocketing Salaries: Senior machine learning engineers now command packages that rival hedge fund traders.
• Concentration Risk: A handful of big tech firms (OpenAI/Microsoft, Anthropic, Google DeepMind) are hoovering up top talent, leaving start-ups struggling to compete.
• Brain Drain Concerns: Smaller markets — including those in Africa and the Middle East — risk losing their best engineers to global AI giants, deepening dependency rather than building local capacity.

For business leaders, this means AI strategy isn't just about capital — it's about access to the right people. Companies that can't attract and retain AI talent will struggle to turn hype into real capability.

Regulatory Readiness: A Critical Difference

Another key difference from the dotcom era is that regulators are moving much faster.

• The EU AI Act introduces risk categories and obligations for high-risk AI systems.
• The US SEC is considering AI risk disclosures for public companies.
• Countries like Brazil, Singapore, and the UAE are drafting frameworks to govern responsible AI adoption.

While this won't eliminate overinvestment, it does mean the next AI "correction" may play out in a more structured environment. Compliance costs, audits, and liability will increasingly shape which companies survive — favouring those with robust governance in place.

Avoiding the Next AI Winter: Lessons for Leaders

The dotcom crash didn't kill the internet — it cleared the field for the winners that would go on to define the next two decades (Amazon, Google). Leaders today can avoid painful lessons by focusing on three areas:

1. Anchor on Business Fundamentals Invest in AI where ROI is clear — customer experience, productivity gains, or risk reduction. Avoid shiny-object projects that look good in a press release but deliver little measurable impact.

2. Build Defensible Moats Proprietary data, domain expertise, and regulatory trust will separate winners from copycats. Don't just integrate someone else's API — build something competitors can't easily replicate.

3. Invest in Governance Early Bias audits, data provenance, and explainability aren't just ethical considerations — they will soon be regulatory requirements. Companies that implement these now will save headaches later.

The Bottom Line

AI is not a passing trend. It will transform industries, just as the internet did — but its early hype cycle will produce both winners and spectacular failures.

Leaders have a choice: either chase the wave blindly and risk getting caught in the undertow, or build the infrastructure, talent pipeline, and governance frameworks that will let them emerge stronger when the hype inevitably cools.

Are we in an AI bubble, or are we just at the beginning of a much bigger wave? How is your organisation preparing for an AI future that could include both explosive growth and painful corrections?