The AI Shockwave: Why Tech Is Riskier Now Than Ever Before
Why You Have to Rethink Tech in the Age of AI
The voiceover above is a custom-made, slightly adapted version of the blog post, edited for a smoother and more engaging listening experience. It’s one of the perks available to paid subscribers, as I’m always focused on adding as much value to my subscribers as possible. Enjoy!
If you're like most investors, chances are your portfolio has some degree of tech exposure. It might even be heavily tilted toward it.
For years – maybe decades – this tilt made a lot of sense. Tech was the growth engine, the compounder, the place where network effects, sticky products, and high-margin software businesses lived. It was where you could relatively safely stretch for 20% annualized returns by just betting on the Nasdaq Composite Index and still feel like a long-term, fundamentals-driven investor.
But 2025 feels… different. More chaotic. More fragile. And maybe even more dangerous.
The rise of generative AI isn’t just the next big thing – it’s an earthquake, making many investors feel excited (due to potential innovative breakthroughs) and uncomfortable (due to increased uncertainty) at the same time. ChatGPT has exploded from zero to 800 million users in two and a half years, clocking over a billion searches per day. That’s faster than Google, faster than TikTok, faster than anything we've seen before.
Fifty percent of S&P 500 companies now mention AI on earnings calls – a number that was close to zero just a few years ago. And yet, beneath all that excitement, there’s a gnawing sense that the tech landscape is shifting under our feet. Maybe violently.
This shift isn’t happening in a vacuum. The average lifespan of public companies in the US has been declining for decades – from above 60 years in the 1950s to 15 years today.
“The average lifespan of a US S&P 500 company used to be 67 years. Now it’s 15.“ - EY
And with AI accelerating the pace of disruption, I expect that number to keep falling – maybe even sharply? What used to take decades to play out may now happen in just a few years. Structural advantages evaporate faster. Obsolescence sets in sooner. And the margin for strategic error keeps shrinking.
My perception is that many public market investors today underestimate how AI may undermine business models that not too long ago were considered “the best business models the world has even seen.”
The irony is that we’re living through a time when tech feels both more ubiquitous but may be less defensible than ever.
Moats are shrinking. Differentiation is collapsing.
In 2022, building a new product might have required a dedicated team and a 12-month development cycle. By the late 2020s, that same product – at least in prototype or concept form – could be generated with just a few well-crafted AI prompts. The gap between idea and execution is collapsing.
The very speed at which innovation is happening – from AI generated movie productions (see video below), to AI-powered drug discovery platforms in healthcare, to humanoid robots navigating private households, to autonomous coding agents that debug and ship software with minimal human input – might be the thing that makes long-term advantage in tech investing increasingly elusive.
This blog post is about risk – not volatility, but the kind of risk that actually matters: permanent capital loss. And specifically, the new kind of risk that investors with “tech exposure” face in a world where software might be turning into a commodity, and the barriers that protected incumbents for the last 10–15 years are suddenly being breached at light speed.
I’m not sounding the alarm just for the sake of it. I’m sounding it because it might be time to step back and ask:
What are you really exposed to?
What assumptions about tech are you still clinging to that might no longer hold?
And how should we think about portfolio construction in a world where even the most entrenched players – Google, Microsoft, Amazon – are throwing tens of billions at a new technology they themselves don’t fully understand?
This post should serve as some food for thought!
Here’s what I’ll explore in this piece:
How the explosive rise of AI is compressing competitive advantages across tech
Why traditional tech moats like network effects, switching costs, and proprietary data are being eroded
Why the current moment might be more structurally dangerous than the dotcom bubble
How even Big Tech’s massive AI investments come with highly uncertain returns
The debate over whether AI is just a feature or a full-blown paradigm shift
What Mark Leonard’s take reveals about which businesses are more insulated from AI threats
Why the physical economy still matters – and might actually be a source of hidden defensibility
How to rethink tech exposure and portfolio construction for a world that’s changing faster than ever
Let’s unpack all of this.
AI Mania and the Fragility Beneath the Hype
AI is having its Cambrian explosion moment. Everyone sees it. Everyone feels it. We’re witnessing fundamental change in real time! It feels nuts – often overwhelming.
The metrics are staggering. ChatGPT’s user base has grown 8x since launch, now hitting 800 million monthly users. India – not the US – is currently the largest national user base, accounting for over 14%. Daily (!) searches via ChatGPT recently passed one billion, doing so more than five times faster than Google ever did.
In parallel, AI is dominating corporate communication. As shown above, over half of the companies in the S&P 500 now mention AI on their earnings calls – a figure that was practically nonexistent just a couple of years ago. From tech to healthcare to energy, everyone’s trying to convince investors (and perhaps themselves) that they’ve got an AI strategy.
But here’s the thing: just because a business embraces AI, this business isn’t in a defensible position. In fact, the insanely high rate of change we are witnessing – the speed with which things are changing – might be what’s making this moment so uniquely risky. Not for consumers – they’re getting better tools every day – but for businesses with “AI exposure” (whether they know it or not) and, of course, investors too, particularly those anchored in the traditional tech playbook.
Because what AI is doing, almost invisibly, is compressing the time window in which competitive advantages matter. Features are cloned overnight. Products that once had 12-24-month leads now get replicated with a weekend hackathon. The friction to build, ship, and distribute new tools has collapsed. In some niches – CRMs, note-taking, developer tools – it’s already a free-for-all. Differentiation is vanishing at a rate that’s hard to mentally model.
And that has real consequences for investors. Because both the magnitude and the durability of return on invested capital are what ultimately drive long-term shareholder value. That’s what moats were supposed to protect. They bought companies time – time to reinvest at high rates, time to scale without being undercut, time to defend pricing power and margins while the rest of the market played catch-up.
But if that time window is shrinking across the board, then the kind of businesses that used to be considered rare and structurally advantaged may become even rarer – or, in some industries, may not exist at all in the same form.
That’s the fragility I’m talking about. AI doesn’t just accelerate innovation – it accelerates obsolescence. And while it might be too early to say who the long-term winners and losers are, it’s not too early to acknowledge that the rules of the game are changing, fast.
It’s a strange paradox: The tech sector has never been more dynamic. Yet as investors, we may never have had less clarity about what actually endures. And without clarity, visibility, or predictability, you cannot invest (because you cannot value assets with a reasonably high degree of certainty).
Before we dive back in, a quick note…
If this post already gave you something to think about, consider this: paid subscriptions don’t just keep this blog alive – they help me produce the best-possible content I can think of. No ads, no fluff. Just high-quality, independent research for long-term thinkers. Your support helps me go deeper, write better, and serve smarter investors like you. If you value clarity in a noisy world, subscribe.
Defining “Tech” – And Why It’s Getting Complicated
“Tech” has always been a loose label. Investors have long lumped everything from chipmakers and cloud platforms to consumer hardware and SaaS billing software into one oversized bucket. Consider my portfolio’s sector exposure shared above. Arguably, my entire portfolio consists of “tech,” but only 7.7% of my holdings are labeled as “Information Technology” companies – are Meta, IBKR, Evolution Gaming, and Wise not “tech” names? I certainly think of them as “tech.”
So arguably, the term has been more of a shorthand than a precise definition – and for a while, that ambiguity didn’t really matter. Whatever sat under the “tech” umbrella generally shared a familiar profile: scalability, high margins (or high margin potential), low incremental costs, asset-light business models, sometimes platform potential, and optionality baked into the narrative. It was less about the sector classification and more about the economics – tech was where operating leverage lived.
But in today’s AI-driven market, that old shorthand is starting to break down. The differences between so-called tech companies are becoming more important than their similarities, as the risks and pace of AI-driven disruption vary widely across business models and sectors.
Apple and Paycom aren’t facing the same threats. Amazon Web Services and a niche vertical SaaS provider don’t enjoy the same moats. NVIDIA and Salesforce may both be “tech,” but the underlying economics, risk profiles, and defensibility of their models couldn’t be further apart.
And now, generative AI is amplifying those differences. It’s pushing us to rethink what “technology exposure” even means from a portfolio perspective.
Take software. For the past decade, it was widely viewed as the holy grail of business models. Recurring revenue. High gross margins. Low incremental cost. “Software is eating the world” became gospel – and rightly so, in that era.
But what happens when software itself becomes easier to generate, easier to replicate, and, in many cases, easier to give away?
The more AI levels the playing field, the more exposed many “tech” companies become. It’s no longer enough to simply be digital, cloud-based, or subscription-driven. The market doesn’t care that your UI is slick or that you’ve got an elegant backend. If someone can prompt-engineer a clone of your product in a few days and distribute it for free or at a significantly lower price, your clients may switch and your economic castle may no longer have any form of “moat.”
This is especially true for companies operating at the shallow end of the complexity pool – low-touch SaaS tools, lightweight consumer apps, etc. These are increasingly becoming commodities. And in a world where code is free, where hosting is cheap, and where AI can write 80% of what used to take teams of developers, software as a category is losing its privileged position.
To be clear: not all tech companies are equally vulnerable. But that’s precisely the point! The broad-strokes label of “tech” is no longer useful. Investors have to go granular – to look not just at sector, but each company’s business model, at management execution speed, general product defensibility features, market structure, and how deeply embedded a company is in its customers’ workflows or supply chains (consider my Monday.com deep dive as a case study).
Monday.com: A Quiet Powerhouse in a Crowded SaaS World
When Monday.com went public in 2021, many investors dismissed it as just another project management tool competing with the likes of Asana, Smartsheet, and Trello. In a sector flooded with flashy SaaS IPOs, it was easy to overlook a company that presented itself not with bold visions of "changing the world," but with a rather modest mission: making work…
That’s where the real differentiation is now. Not in the label, but in the specifics.
The best part is just ahead – unlock the full post to continue.
Investing is the one domain where better thinking compounds. If this post up to this point gave you new insights, the full piece goes even deeper. If you’re serious about sharpening your investing edge, the full post (and all my previous premium content) is just a click away. Consider a paid subscription is an investment in your decision-making process – and your returns.