Palantir is down 28% in 2026. Michael Burry's "AI is eating the middleman" thesis just got proven by Anthropic's newest product — and it's coming for every SaaS tool on your company's spend.
On April 8, Michael Burry — the hedge fund manager chronicled in The Big Short for calling the 2008 housing crash — posted on X that Anthropic was "eating Palantir's lunch." He deleted the post within hours. Palantir's stock fell 7% anyway. Two weeks later, the stock is down roughly 28% for the year.
That's the part you read in the news. Here's the part that matters: later the same day Burry posted, Anthropic launched the product that proves his thesis. It costs eight cents an hour. And it's coming for the software your company is paying for right now.
How Jennifer Aniston’s LolaVie brand grew sales 40% with CTV ads
The DTC beauty category is crowded. To break through, Jennifer Aniston’s brand LolaVie, worked with Roku Ads Manager to easily set up, test, and optimize CTV ad creatives. The campaign helped drive a big lift in sales and customer growth, helping LolaVie break through in the crowded beauty category.
TLDR: Michael Burry is shorting Palantir because he believes enterprise AI is replacing the software middleman. On the same day his post went viral, Anthropic shipped Claude Managed Agents — autonomous AI workers at $0.08/hour already running in production at Notion, Rakuten, Sentry, and Allianz. Roughly $1.4 trillion in SaaS market cap has evaporated in 2026. The software bill your finance team approved last quarter is being repriced in real time.
Burry's Bet Is Working
Burry disclosed his short position on Palantir last fall — long-dated put options expiring in 2027, covering roughly 5 million shares. His deleted post cited a single data point from Ramp's enterprise spend tracker: Anthropic is now capturing 73% of all new enterprise AI spending.
His core argument isn't really about Palantir. It's about the business model. Palantir sends its own engineers into client offices for months at a time to keep its software running — in its own 10-K, the company books this work as "professional services." Burry calls it a low-margin consulting shop dressed up as a high-growth AI company. Anthropic, meanwhile, offers a plug-and-play API that drops into any workflow. No engineers on-site. No six-month integration.
The growth trajectories back him up. Anthropic went from $9 billion in annual recurring revenue at the end of 2025 to $30 billion by April 2026. Palantir took 20 years to reach $5 billion. Wedbush analyst Dan Ives called Burry's thesis a "fictional narrative." The chart has a different opinion.
Your meetings just got a lot more useful
Granola connects your meeting notes to tools like Claude and ChatGPT, so your AI actually understands your work context. Update your CRM, extract tasks, or write follow-ups without copy-pasting a thing.
The Product That Sealed It
Hours after Burry's post went viral, Anthropic launched Claude Managed Agents — a production platform for deploying autonomous AI agents without any of the infrastructure work that usually eats months of engineering time. Price: $0.08 per session hour on top of standard API tokens. Notion, Rakuten, Sentry, and insurance giant Allianz were already live at launch.
If you're a CIO deciding between a multi-million-dollar Palantir deployment and a Claude Managed Agent at cents per hour, the math doesn't require an MBA. This is the product enterprise buyers have been waiting for — and the product Burry said would make the middleman obsolete. Both things shipped the same afternoon.
The $1.4 Trillion Already Gone
The market already believed Burry before he posted. Roughly $1.4 trillion in SaaS market capitalization has evaporated in 2026. The broader software stock index is down 25.5% year to date. Salesforce, ServiceNow, Workday, HubSpot, Atlassian — all down more than 20% off their peaks. Analysts started calling it the "SaaSpocalypse" without irony.
The common thread across every selloff: these are companies whose core product is a workflow wrapper. Ticketing. Scheduling. Routing. Dashboards. Forms. Reports. The valuations were built on the belief that the wrapper itself was the moat. Claude Managed Agents is a direct argument that the wrapper is actually the bottleneck.
Why This Reaches Your Line Item
Here's the uncomfortable part for any department head reading this. Seat-based software licensing is built on a simple equation: you have 100 employees, so you buy 100 seats at $120/month. If 10 autonomous AI agents at $0.08/hour can do the work those 100 seats enabled, the math breaks. And it's already breaking.
Ramp's April AI Index shows business AI adoption crossed 50% for the first time — up from roughly 1-in-25 companies a year ago. VC-backed firms are at 80% adoption. Nearly every category of mid-tier SaaS is exposed: marketing automation, HR screening, customer support tooling, analytics dashboards, document workflows. The moat isn't the software anymore. The moat is regulated data, deep enterprise lock-in, or being the API itself.
The reason Burry's bet keeps working: his thesis was never really about Palantir. It was about every company whose revenue depends on being the wrapper around an AI the customer can now just buy direct.
The Prompt (Copy This)
This prompt audits your company's software stack for AI-replacement risk. Paste it into Claude or ChatGPT.
You are an enterprise software consultant with 20 years of experience in
AI automation. Before giving me any analysis, interview me with these
questions one at a time:
1. What's my role and industry?
2. What's my company size (headcount and revenue range)?
3. What are the 3-4 SaaS tools I use most at work (name them)?
4. For each tool, what's the main workflow I use it for?
5. What's the rough annual spend per seat for each (or "not sure")?
After I've answered all five, analyze my stack and return:
A. Per-tool risk assessment (HIGH / MEDIUM / LOW exposure to replacement
by an autonomous AI agent within 18 months), with the specific reason
for each rating.
B. For any HIGH-risk tool, the exact workflow an AI agent could replicate
today, and a rough hourly cost estimate to run that workflow
autonomously.
C. A 90-day action plan: which tool to audit first, which to renegotiate,
which to keep and why.
Be direct. Don't soften the findings. I want to know where I'm over-paying.
Prompt Proof Table
| Reader | Main Stack | AI Risk |
| Marketing Director, SaaS startup |
HubSpot, Mixpanel, Canva Pro | HIGH |
| HR Leader, 2,000-person enterprise |
Workday, Greenhouse, BambooHR | MEDIUM |
| Finance Analyst, mid-size manufacturer |
Bloomberg, Excel, Power BI | MEDIUM |
| Ops Manager, mid-size law firm |
Clio, DocuSign, Slack | LOW |
RED = most workflows replaceable today · AMBER = partially replaceable, regulated or integration moat · GREEN = regulated data, legal authentication, or platform lock-in.
Same prompt. Different stacks. Different exposure. Run it against your line items.
Burry bet that enterprise software is being repriced one plug-and-play API at a time. He's winning.
Your company's next renewal cycle will find out how much.
About This Newsletter
AI Super Simplified is where busy professionals learn to use artificial intelligence without the noise, hype, or tech-speak. Each issue unpacks one powerful idea and turns it into something you can put to work right away.
From smarter marketing to faster workflows, we show real ways to save hours, boost results, and make AI a genuine edge — not another buzzword.
Get every new issue at AISuperSimplified.com — free, fast, and focused on what actually moves the needle.
If you enjoyed this issue and want more like it, subscribe to the newsletter.
Brought to you by Stoneyard.com • Subscribe • Forward • Archive