Isn’t it remarkable how investors always say “this time is different” when it comes to a new technology or a bubble? This month such a rallying cry emerged in an essay-length blog post that proclaimed: “Something big is happening.”
It was a hyperbolic love letter to AI that’s gone viral, as the term originally meant. How many times were you sent it?
There was a lot of “gee whizz” and “nobody else knows what’s coming” that I found unpalatable. What did resonate is something I’ve been telling people for a while. If you’re using the free services, which most people are, you’re not seeing the full potential.
“The free version is over a year behind what paying users have access to,” Matt Shumer wrote. “Judging AI based on free-tier ChatGPT is like evaluating the state of smartphones by using a flip phone. The people paying for the best tools, and actually using them daily for real work, know what’s coming.”
Against this background, the week before the so-called SaaS-pocalypse wiped $500bn off US stock markets in firms that provide software-as-a-service or are invested in them.
If you missed it, SaaS was the tech world’s darling before we were swept up in the launch of ChatGPT in November 2023. Let’s call that AI-Hysteria 1.0. Shumer’s impassioned love letter — he’s the CEO of an AI firm, so reach for your salt — is AI-Hysteria 2.0.
There are enough smart, rational and non-hyperbolic people who think the new agentic AI services are indeed a new AI moment. I have many geek friends who swear by the efficiencies of AI services. But they pay $100 (R1,600) a month to be able to do that. Claude Code costs $200 (R3,200) a month.
The reason I feel such reservation about Shumer’s over-the-top enthusiasm is because we’ve all seen this overhyped movie before.
What happened last month was the release of Anthropic’s Claude Code agent. Much more sophisticated than the first versions of AI chatbots, it can automate tasks and plugins, such as accounting and HR software. It earned Anthropic $1bn in the first six months after its launch in May 2025.
Culmination
Last week saw the months-long culmination of fears over new AI services usurping the rich profit margins of SaaS firms. They have legacy systems that tie users to their proprietary way of working, which until now was a barrier to competition.
It may be a web-based service but lock-in is still a key problem, with major software vendors trying to grow their own businesses by adding in other SaaS offerings.
Workday, an HR specialist, has been marketing itself as the gateway to other services such as financial and accounting. Its global CEO resigned after its stock plunged this year.
In a single day last month the market capitalisations of big firms like Salesforce, Adobe, Workday and ServiceNow dropped 7%, while financial services provider Intuit plunged 11%.
But, as Forbes points out, “markets do not erase that much value unless a core assumption breaks”. And that assumption was the “durability of legacy SaaS growth”.
Forbes adds that “the damage had been building for months” as the IGV software index was already down 30% from its peak in September. “What changed last week was not the direction, but the speed.”
Not surprisingly, “short sellers have already made over $20bn in 2026 betting against legacy SaaS, and are doubling down”.
Why the SaaS panic?
AI has made a significant shift from a feature in other software to a direct competitor to that software. HR, accounting and customer relationship software has traditionally been standalone, including multibillion businesses like Salesforce, which was the poster child of its own SaaS revolution.
It was then referred to as Web 2.0 and was a genuine revolution in web-based services. The subsequent Web 3, an attempt to tie up all the crypto and next-gen payment systems, has been eclipsed by the ChatGPT hype cycle.
A more cogent, fact-based essay by another software executive, Paul Ford, made a far more coherent argument that sounded a lot less like a “guess what I uncovered” secret than Shumer’s sounds like. He describes his own so-called vibe-coding (where you make your own website or plug-in using text prompts to a chatbot), which he completed on his subway trip home. Using Claude Code, he rebuilt his own website (which would have cost $25,000) and made a database for a friend ($350,000).
“Is the software I’m making for myself on my phone as good as handcrafted, bespoke code?” he wrote in The New York Times. “No. But it’s immediate and cheap. And the quantities, measured in lines of text, are large. It might fail a company’s quality test, but it would meet every deadline. That is what makes AI coding such a shock to the system.”
His essay is appropriately titled “The AI disruption we’ve been waiting for has arrived”.
“No matter where you work, my hunch is this is coming for you. Have you noticed the software you use every day adding ‘AI features’? That’s the top of the slippery slope. Whatever unifying principle equates to ship risk in your industry, people are trying to mitigate it with AI. Insurance, finance, architecture, manufacturing, textiles, every kind of project management — they want to automate it all through AI.”
Indeed, as Shumer points out, something big is happening. But something big happens every decade. So do bubbles. It’s only the size that matters.
• Shapshak is editor-in-chief of Stuff.co.za.
Would you like to comment on this article?
Sign up (it's quick and free) or sign in now.
Please read our Comment Policy before commenting.