Cloudflare sell‑off, AI split

Cloudflare plunged roughly 13% on fears that new AI entrants like Anthropic's Claude Mythos will trigger a 'SaaS‑pocalypse' and compress pricing, dragging some cybersecurity names down. ( ) At the same time, market commentary says hyperscalers are still spending heavily on AI infrastructure—Rob Black and quarterly recaps cited $166 billion in Q1 hyperscaler spend (up 13% YoY) and massive compute deals—so investors are weighing AI capex tailwinds against pricing risk. (youtube.com)

Cloudflare closed at $166.99 on Friday, April 10, after dropping 13.5% in one session, and the sell-off spread across software and cybersecurity names in the same two-day stretch. The trigger was not a weak Cloudflare quarter. The trigger was a new round of fear that artificial intelligence tools are getting good enough to do jobs that software companies used to package into subscriptions. That fear got sharper on April 7, when Anthropic published details on Claude Mythos Preview and said the model could identify and exploit previously unknown software flaws across every major operating system and every major web browser in its testing. Anthropic also said more than 99% of the vulnerabilities it found were still unpatched, which is why it limited disclosure. Investors read that like this: if one model can inspect code, find bugs, write fixes, and automate security work, then some software vendors lose pricing power the way taxi dispatchers lost pricing power when ride-hailing apps arrived. The market hit Cloudflare, Okta, CrowdStrike, SentinelOne, Zscaler, Atlassian, Workday, Adobe, Salesforce, and Intuit in the same sweep. Cloudflare got caught in that basket even though its own business is not a simple seat-license software story. In its latest reported quarter, Cloudflare posted fourth-quarter 2025 revenue of $614.5 million, up 34% from a year earlier, and said it closed its largest annual contract value deal ever at $42.5 million per year. Matthew Prince used that same earnings release to argue the opposite of the panic trade. He said agents are becoming “the new users of the web,” and that more agents should drive more traffic through Cloudflare’s network, more code onto Cloudflare Workers, and more demand for security and networking products. That is the split investors are trying to price in. One camp thinks artificial intelligence turns software into a cheap commodity, while the other thinks artificial intelligence sends a flood of new traffic, code, and security demand to the companies that own the pipes and compute layers underneath it. The spending data still favors the second camp for now. Axios reported in February that Amazon, Alphabet, Meta, and Microsoft were on track to spend about $610 billion in 2026, roughly triple the level from two years earlier, as the artificial intelligence buildout keeps getting larger. So the market is no longer asking one broad question like “Is artificial intelligence good for tech.” The market is asking a narrower one: who captures the money when the models get stronger, the software gets cheaper, and the infrastructure bill keeps rising. Friday’s Cloudflare drop says traders think pricing risk can hit before infrastructure benefits show up in reported numbers. Cloudflare’s next scheduled earnings date is May 7, 2026, which is when investors will get a cleaner read on whether this was panic selling or an early warning.

Get your own daily briefing

Scout delivers personalized news, insights, and conversations tailored to your role and industry.

Download on the App Store

Shared from Scout - Be the smartest in the room.