Artificial Intelligence is changing the world. But behind the excitement lies a frightening question. Is AI the future of technology? Or is it the biggest financial bubble in modern history?
The numbers are staggering.
In 2020, the world's biggest technology companies spent about $90 billion on AI infrastructure. By 2026, that figure is expected to soar to $725 billion. Tech giants like Amazon, Microsoft, Google and Meta are investing at an unprecedented pace, building giant AI data centres across the world.
But there is one alarming problem.
The AI industry is earning only about $75 billion a year. That leaves a massive $650 billion gap between spending and revenue. Companies are burning money much faster than they are making it.
Investors are beginning to worry. Venture capitalist David Cahn has called it "the AI industry's $600 billion question," referring to the huge revenue needed every year just to justify today's massive investments.
The biggest concern is that AI is not delivering the promised profits.
Studies show that nearly 73% of enterprise AI projects fail to achieve their expected return on investment. Another report found that 95% of generative AI pilot projects fail to produce measurable financial benefits. Businesses are discovering that AI often costs more than the value it creates.
Even industry leaders are sounding the alarm.
Alex Karp, CEO of Palantir, says many companies are "paying for tokens that create no value." He has also warned that "these models have been completely, irresponsibly, oversold."
His remarks reflect growing frustration among businesses that expected AI to transform their operations but are struggling to see meaningful financial returns.
Building AI itself is extraordinarily expensive.
A single large AI data centre can house nearly 100,000 Nvidia GPUs. Each chip costs between $30,000 and $40,000, making the hardware bill alone $3 billion to $4 billion. Once cooling systems, electricity, networking, security, fire protection and construction are added, one AI data centre can cost anywhere between $10 billion and $25 billion.
The impact is already reaching ordinary consumers.
The enormous demand for AI chips has pushed up memory prices across the electronics industry. Consumer devices have become more expensive as manufacturers compete with AI companies for critical components. Even major technology firms have been forced to increase prices because of rising hardware costs.
History offers a chilling warning.
During the 1990s internet boom, companies invested billions in fibre-optic networks. Demand failed to keep pace with expectations.
Much of that infrastructure remained unused, many firms went bankrupt and investors lost fortunes. The internet eventually succeeded, but countless companies did not survive the bubble.
Many experts fear AI could repeat that painful cycle.
If companies continue spending without earning enough revenue, investment could suddenly dry up. That could trigger a sharp market correction, bankrupt AI startups and lead to widespread job losses across the technology and service sectors.
There is another possibility. To recover their enormous investments, AI companies may dramatically raise prices. Today's affordable AI tools could become premium services that only large corporations and wealthy customers can afford.
The technology behind AI is real, and its long-term potential is undeniable. But the current investment frenzy is raising serious questions.
Unless profits catch up with spending, the AI revolution that promises to reshape the future could instead become one of the biggest financial shocks the world has ever witnessed.
