It’s fitting that after the worst real winter I can remember, tech would have its own nasty winter - SaaSpocalypse. I knew AI would come calling for tech companies, but even old school CRE brokerages weren’t spared.
The fire was initially built on January 12th, when Anthropic released Cowork (CoWork) in beta. This is a really cool tool that runs everything on your computer like a personal data butler. I wrote about this one last month in the MCP Server newsletter.
But then on January 30th, gasoline was thrown on that fire. Anthropic released a Claude Legal plugin (Legal) to “to review contracts clause-by-clause against your configured negotiation playbook, with green/yellow/red flags and specific redline suggestions.” The hamster inside the investors’ brains started slowly trotting on the wheel, pondering the potential ripple effects…
That announcement came on a Friday. The news started really circulating over the weekend and Monday saw investors wondering about the broader ramifications and then BAM!
Tuesday was total carnage - $285B in tech market cap evaporated in a single day (Bloomberg Tech Selloff). Over the course of one week, more than $900B in market cap was lost. Somewhere, Michael Stipe is nodding and thinking, “On a long enough timeline, I am always right” (there’s like 2 readers that will get that).
Last week, it bled over into CRE, where the publicly traded shops got hammered (Bloomberg Brokerages).
What started as an indictment of SaaS providers spilled over into high margin service providers. If an AI agent can replace lawyers, why not brokers?
But I don’t think the trade is really about replacing lawyers/brokers/tech. Instead, it’s the change to expected future cash flows. The SaaSpocalypse selloff wasn’t a funeral for SaaS - it was a repricing of the per-seat model.
And per seat only works when the seat is doing the work.
That’s the real problem: AI doesn’t buy seats, it buys outcomes. If an agent can crank through 80% of a workflow, the “$99/user/month for a prettier checkbox” tax gets harder to justify.
So the market did what markets do. It didn’t ask, “Is this company good?” It asked, “Is this company’s future math less good than I thought last week?”
The SaaS pricing you pay may change. Probably some messy blend of usage-based, outcome-based, and enterprise pricing - with a lot of awkward hybrids in between.
Is anyone really about to rip out Yardi and bet the farm on a weekend “vibe-coded” homebrew app some 26-year-old associate hacked together? Please. And nobody is firing their broker because ChatGPT can write an email.
When you pay for software or a brokerage, you’re paying for an ecosystem: expertise, accuracy, trust, security, support - and the fact that when something breaks, you can call someone whose whole job is fixing it.
Nothing has ever stopped a firm creating an internal team to replace brokers. Nothing has ever stopped a firm from building its own tech. And yet there is still a market for both. That's because you pay for the output.
AI lowers the barrier to building tools, sure. But “we built it ourselves” comes with a hidden subscription cost anyway: maintenance, support, upgrades, and the eventual realization that part-time software is rarely better than full-time software.
I know I’m talking my book here, but I have to say simma down SaaSpocalypse. Spring is coming. AI may change the pricing model, but not the underlying need for the outputs.
