Today the market’s up!
Tomorrow it’s down 1 percent.
Nvidia’s earnings are this week ... The market is down another 1 percent in anticipation.
They beat expectations ... but the market still drops in fear that the AI tech market is becoming overvalued.
Does any of this make sense? Is this the beginning of a bubble bursting or just the normal gyrations of a volatile market in a crazy geopolitical world? That’s the question I keep asking myself.
I’ve lived through several stock market corrections and crashes. I remember the day the dot-com bubble burst. I was standing around the Bloomberg terminal at the public relations consulting firm I was running, watching companies like Balls.com (yes, that was a real one) skyrocket one day and disappear the next.
It was a time of wild speculation, easy money and “get big fast” business plans that had little to do with actual revenue. But as history shows, the internet didn’t fail — bad business models did.
Today, as we watch artificial intelligence dominate the headlines and draw staggering amounts of capital, I can’t help but feel a sense of déjà vu. The question is, are we headed for another tech bubble, or is this simply the natural turbulence of high speed innovation?
Lessons From the Dot-Com Bust
- Underlying technology endures: The internet survived the crash; AI infrastructure should do the same.
- Bad business models fail: Companies like Pets.com failed because applications lacked a sound revenue model or real-world problem-solving capability.
- The litmus test for applied AI: Can the developing company clearly articulate the problem that exists in the workplace and that it is going to solve? If a company can’t answer “What specific business process does this improve?” in one sentence, it’s a red flag.
What Exactly Is a Tech Bubble?
Before we decide whether we are on our way to one, let’s remind ourselves what a tech bubble actually is.
A bubble happens when excitement about a new technology — and the money chasing it — outpaces reality. Investors pour cash into companies not because they’re generating revenue or solving real problems, but because they might do so someday. Valuations soar, business plans get thinner and everyone convinces themselves that “this time is different.”
We’ve seen this movie before. The dot-com boom and bust of the late 1990s, the social media mania of the 2000s, the crypto and NFT frenzy just a few years ago. Each time, a rush of innovation was followed by over-investment, then a correction occurred. But the underlying technologies, whether the internet, social platforms or blockchain, didn’t vanish. They matured. The money invested to establish the infrastructure for the technology was well spent, even if some of the companies involved are no longer here to enjoy the success or reap the benefits.
This is why I’m not convinced that what we’re seeing today is a bubble. Yes, trillions of dollars are being invested in AI data centers. The Wall Street Journal recently reported that global AI infrastructure buildout could hit $3 trillion by 2030. As a result, volatility is likely. A correction is a distinct possibility. But, if history teaches us anything, a full-blown collapse is hopefully not in the offing.
AGI Infrastructure vs. Applied AI
As I’ve been reading, listening and talking to people across the communications technology industry, I’ve realized that, when we talk about AI, we’re actually talking about two different things happening at the same time.
On one side, there’s the AI infrastructure build-out — the race to build the data centers, chips, models and computing power necessary for achieving the goal of the behemoth tech companies. That goal is artificial general intelligence, basically computers that can think and do what humans do — scary! This is the ambitious moonshot, the big, world-changing vision that we’re hearing about weekly on podcasts.
But running alongside this is something different. It’s the application of AI both for consumers and for the workplace. Let’s call this “applied” artificial intelligence. This is the practical layer — the AI tools companies are building to solve real workplace problems. Think HR automation, employee communications, customer service reduction, workflow augmentation, analytics and productivity tools. This part of the AI economy is about making people more efficient in their work and driving measurable business outcomes, not visionary AGI milestones.
The thing is, both carry bubble risks, both independently and together. When the dot-com bubble burst, both types of investment crashed. That means it affected the visionary companies like WorldCom, Global Crossing and 360networks that built the backbone that supports today’s internet successfully. Likewise, numerous companies like Pets.com, Kozmo.com and Napster that were building applications to be applied across the newly established internet foundation went under.
I think it’s fair to say that the internet was successful. But what lessons can we learn from the applied internet companies that failed? What should companies developing applications that will take advantage of the evolving AI infrastructure learn from their internet economy predecessors? What should companies thinking about using these applied solutions be worried about?
AI in the Workplace – What to Look Out For
When starting my company in the early 2010s, I remember telling my team that for our solution to be worth anything (and for companies to want to pay for it), it needed to address a real problem that exists in the workplace.
The same logic applies to what’s going on today with AI. Many so-called AI companies out there are just hoping to take advantage of all of the hype. They’re aiming to capture some of the massive amount of money that’s being invested in the AI companies that are actually laying the groundwork and infrastructure for the future.
When I look at AI in this moment, I can’t help but feel echoes of my days watching the Bloomberg terminal. Swap out “mobile” or “internet” for “AI,” and it’s the same movie: a lot of excitement, a lot of capital and only a handful of teams who truly understand the business problems they’re solving.
Some will win. Many will not. The ultimate litmus test, however, should be whether the technology addresses a clear, measurable workplace need. Can the developing company clearly articulate the problem that exists in the workplace and that it is going to solve? If a company can’t answer the question, “What specific business process does this improve?” in one sentence, that’s a red flag.
Of course, there are other things to look out for. Things like the following:
- Is the new solution asking users to change the way they work or does it simply enhance productivity? Solutions that force people to change established workflows will face adoption challenges unlike tools that integrate smoothly and support existing processes.
- Does it seamlessly integrate with existing workplace platforms like Microsoft, ServiceNow and Workday, or is it looking to replace them? Companies that think now is the time to upset the status quo, as if it hasn’t been thrown into question already, are probably out of touch and unlikely to succeed.
- Is the company talking about data ethics, trust, transparency and governance? AI presents serious issues when it comes to a company’s proprietary data, who has access to it and how it is used. Technology companies looking to take advantage of the AI opportunity must offer believable assurances and empathy towards the organizations they want to do business with.
To Survive, Look to the Past
So, are we in a bubble or is one approaching? The answer is it’s still early. The infrastructure is under development. The likes of Nvidia, Meta, Google, Apple and OpenAI are spending to create the next technological generation. This is the same stage that the internet was at during the mid 90s; the plumbing was being built, long before anyone fully understood what the applications would become.
Meanwhile, lesser-known companies, some well funded and some bootstrapped, are racing to apply AI in the workplace. Some of these teams understand their industries deeply. They know the workflows, the challenges and the measurable outcomes that matter. Others are simply shouting “AI” as loudly as possible and hoping the market rewards their volume.
History tells us what happens next. Companies that pair technological capability with a real business problem and define, clearly and simply, the value they bring to the workplace will become winners. And there will be those that fade away just as quickly as their dot-com predecessors did. But the underlying AI transformation presently underway will continue.
Just as the internet survived its crash and went on to reshape our lives, artificial intelligence will continue to evolve, mature and become foundational to how work gets done as well as the way in which we live our lives. This time, that hopefully won’t include a bubble bursting.
