The Complex Relationship Between AI Investments and Human Needs in a Rapidly Evolving World

rajeev385
Dec 9, 2025
4 min read

The technology has already woven itself into our lives. The only question is which companies survive the journey.

Is AI in a bubble? Probably. Will it crash? Quite possibly. Does any of that matter for AI's long-term trajectory? Not really.

Even Sam Altman, the CEO of OpenAI, has acknowledged that investors are "overexcited" about AI. The Bank of England has warned about the risks of a global market correction. Apollo Global Management's chief economist has stated that the current AI bubble is bigger than the internet bubble of the 1990s. And in a recent Bank of America survey, 53% of fund managers believe AI stocks are already in bubble territory.

So why am I saying AI is here to stay regardless? Because we've seen this film before. And we know how it ends.

The Numbers Are Genuinely Bonkers

Let me be clear about the scale we're talking about. Global AI spending is forecast to hit nearly $1.5 trillion in 2025, rising to over $2 trillion in 2026 (Gartner). Big Tech is collectively spending $405 billion on AI infrastructure this year alone. Microsoft's quarterly AI infrastructure spend hit $35 billion. In the first half of 2025, AI-related capital expenditure contributed 1.1% to US GDP growth, outpacing consumer spending as an economic driver (J.P. Morgan).

These figures would have seemed delusional five years ago. And yes, some of the financial engineering looks worrying. Nvidia investing $100 billion in OpenAI, which then spends that money on Nvidia chips, has all the hallmarks of circular financing that plagued the dotcom era. As investor Michael Burry (the "Big Short" bloke) has noted, many of these deals resemble the kind of structures we saw right before previous crashes.

But here's what the bubble-watchers are missing.

The Technology Is Already Embedded

Unlike the dotcom era, where companies were promising future transformation, AI is already transforming present-day operations at scale.

According to McKinsey's 2025 State of AI report, 78% of organisations now report using AI in at least one business function, up from 55% just a year ago. Generative AI adoption has more than doubled, with 71% of organisations using it in at least one area, compared to 33% in 2023.

In healthcare, 27% of health systems have already adopted AI, with ambient scribing tools reducing clinical documentation time by up to 80% in some deployments (Menlo Ventures). Manufacturing has seen 77% adoption, with companies reporting a 23% average reduction in downtime from AI-powered automation. Financial services firms are spending over $20 billion annually on AI, with 68% of hedge funds now using AI for market analysis.

And it's not just enterprise. Deloitte's 2025 Connected Consumer Survey found that 53% of US consumers are now either experimenting with or regularly using generative AI. Over 378 million people worldwide actively use AI tools. Nearly one in five American adults relies on AI daily.

This isn't speculation. This is infrastructure.

The Dotcom Parallel Actually Supports the "Here to Stay" Argument

Everyone remembers the dotcom crash. The Nasdaq fell 77% from its peak. Pets.com became a punchline. Thousands of companies went bankrupt.

What people forget is what happened next.

Amazon nearly died during the crash, watching its stock plummet from $107 to $7. It survived by diversifying and relentlessly focusing on customer experience. Google launched during the bubble and emerged stronger by building a sustainable business model around search advertising. eBay, PayPal, and Yahoo all weathered the storm.

More importantly, the internet didn't go away. The bubble's burst didn't undo the web's potential; it just brought rationality back to valuations. As one analyst put it, "The infrastructure born out of this boom propelled the revolution."

By 2022, twenty years after the crash, internet traffic had increased 1,000x and the price of data transit had fallen 1,000x. The companies that won weren't necessarily the ones getting the most press in 1999. They were the ones with solid business models that could exploit the technology effectively.

What This Means for the AI Era

AI is following a remarkably similar pattern. We have:

Massive infrastructure buildout (data centres, chips, power)
Circular financing arrangements that inflate apparent demand
Speculative valuations on companies without clear paths to profitability
Genuine underlying technology that's already being adopted at scale

The difference? AI valuations, while high, aren't as stretched as the dotcom era. The Nasdaq currently trades at about 28x forward earnings, compared to over 70x for tech leaders at the 2000 peak. More importantly, there's real revenue. OpenAI is projecting $13 billion in 2025 revenue. Anthropic is targeting $9 billion run-rate with plans to double in 2026. These aren't clicks and eyeballs; they're actual paying customers.

Will there be a correction? Almost certainly. As JP Morgan's Jamie Dimon recently said, some money invested now will be wasted. But he also noted that AI "will pay off... just like cars in total paid off, and TVs in total paid off, but most people involved in them didn't do well."

The technology survives. The infrastructure survives. Many of the current players won't.

The Real Question

The debate about whether AI is a bubble misses the point. The better question is: what does responsible adoption look like in an environment of investment exuberance?

For organisations implementing AI today, the lesson from the dotcom era is clear. Focus on sustainable business models, not hype. Build for actual user needs, not speculative future capabilities. And remember that the companies that emerged strongest from the dotcom crash weren't the ones that raised the most money; they were the ones that solved real problems efficiently.

AI has already crossed the threshold from experimental technology to operational infrastructure. Healthcare systems are using it to reduce clinical burnout. Manufacturers are using it to predict equipment failures. Financial institutions are using it to detect fraud in real-time. Consumers are using it to write emails, plan meals, and answer questions.

That genie isn't going back in the bottle.

Some of the sources I have referenced: