Who Is Anthropic Killing Next?
Anthropic wiped $285 billion in market cap in February. From product announcements. Thomson Reuters had its worst day on record (-16%). LegalZoom -20%. ServiceNow -23%. JFrog -25%. CrowdStrike -8%. Three product cycles, three weeks. If you hold any software or IT services name you’re probably wondering: who’s next? I’m going to show you a framework to predict which companies Anthropic disrupts next, using AI and their own public data. I ran this in Claude Cowork, Anthropic’s desktop app (available on Mac + Windows) for delegating research to AI agents. Three prompts, three Excels, a scored watchlist:
{{ snippet.was-this-email-forwarded }} Prompt 1: The Disruption TimelineI couldn’t keep track of what Anthropic was shipping in February and what it was breaking. I told Cowork to research every Anthropic release from the last 6 months, cross-reference stock reactions, and save a formatted Excel. Prompt 1: https://www.finprompter.com/share/847d2b8f-1bcd-45b1-aa55-73cb8bef460c The Result: Full output: anthropic_disruption_timeline.xlsx Four announcements moved markets: Here’s what jumped out. The damage isn’t random. There’s a repeatable signal chain and I’m surprised more people aren’t talking about it. Go back to cybersecurity. Anthropic published an “AI for Cyber Defenders” research blog in September 2025. Nobody cared. It was a technical paper, no product announcement. February 5th, Opus 4.6 Red Team research found 500+ zero-days. This was the capability demo, proving the model could do what the blog described. Still nobody connected the dots. Fifteen days later, Claude Code Security shipped as a full product. JFrog -25%. Six cybersecurity names hit in one session. Research blog → capability demo → product launch → stock crash. Same sequence with legal and coding too. The blog is the early warning. The capability demo is confirmation. The product launch is when everyone panics. That gap between steps 2 and 3 has been about 15 days. Can we get ahead of it? Prompt 2: The Kill ScoreI found what I believe is one of the most underutilized public datasets in investing right now. Anthropic publishes an Economic Index on HuggingFace. Updated quarterly. It covers task usage, automation vs augmentation patterns, and collaboration data across countries, all mapped to employment categories. Basically, Anthropic is telling you which tasks their models handle autonomously at scale. And once the model can do something well, the product follows. I got this prompt to analyze all the raw data and put it into Excels. Cowork downloaded the dataset, calculated a Kill Score for every task cluster, and mapped each to its sector. The Kill Score combines how often people are already using AI for a task with how much of that task AI handles autonomously. High on both? A named product is coming. Prompt 2: https://www.finprompter.com/share/3a10cbb5-75a3-4ba5-ab78-9f3e82297f8b The Result: Full output: Kill_Score_Analysis.xlsx Software dev alone scores 12.51. More than the next five sectors combined. Backward validation is what sold me. Security scanning grew +279% in API data before Claude Code Security shipped. Legal extraction was top-5 before Cowork plugins dropped. The data was quietly telling us what was coming. Now we know what tasks are getting automated. Prompt 3 maps those to the actual companies holding the bag. Prompt 3: The WatchlistWhich companies does this map to? And how does AI affect them? Not every company in the blast radius dies. Take Goldman Sachs. Anthropic is never becoming an investment bank. No banking license, no prime brokerage. But Goldman’s junior analysts spend 80+ hours a week on tasks Claude already does: modeling, document review, pitch decks. Fewer bodies, same revenue. Costs down, margins up. The firm gets stronger. Now compare that to Upwork or Fiverr. Their entire business is connecting clients with freelancers who do tasks Claude handles natively. There’s no moat. The product itself is what AI replaces. I built a prompt that scores companies on distribution moat, Kill Score exposure, and revenue concentration. Three archetypes emerged: Prompt: https://www.finprompter.com/share/bb1ba66c-eb4f-4f9d-8750-fc878c620c50 The Result: Full output: Anthropic_Kill_List_Watchlist.xlsx Top 10 most vulnerable: Top 10 are ALL Firm Killers. The bottom of the list tells a different story: Appian (8.43) is the most vulnerable name on this list with no Anthropic announcement targeting it yet. The Bottom LineHere’s what I keep coming back to. Anthropic is publishing the data that tells you which sectors they’re coming for. It’s sitting on HuggingFace. Updated quarterly. In plain sight. The signal chain from Prompt 1 tells you when to pay attention. The Kill Scores from Prompt 2 tell you what they’re building next. And the watchlist from Prompt 3 tells you who is exposed. To sharpen this further, layer on Anthropic job postings (hiring in a new vertical = leading indicator), blog alerts on their research page, and partnership announcements. Most investors will see the next product launch and panic. You’ll have seen it coming. Updates: Claude Code for Finance LIVE WEBINARWant something even more powerful? Everyone is talking about Claude Code. It's the most powerful AI tool I've used for finance. I'm putting together a live webinar course designed specifically for investors and buyside professionals. Here's what we'll cover:
This will be a capped live webinar course. Spots will fill fast once we open registration. Reply to this email or click the waitlist link below to get notified as soon as it drops. Hope to see you then.
|
Enjoyed this article?
Get more AI for finance content delivered to your inbox.