Morningstar doesn't take away moat ratings lightly. It just took away a lot of them.
The investment research giant reviewed 132 companies across its global coverage to stress-test whether their competitive advantages can hold up against AI. In Morningstar's framework, a company's "economic moat" is whatever keeps competitors at bay — brand loyalty, switching costs, proprietary data. A "wide moat" means that advantage should last 20 years or more.
The results triggered the largest reclassification in the firm's 20-year history, bigger than the reshuffles caused by the 2008 financial crisis, COVID, or the oil price collapse. Twenty-two wide-moat companies got downgraded. Another 18 narrow-moat companies were knocked down further. Only two were upgraded.
"We felt it was no longer sufficient to assess moats without an explicit AI lens," Eric Compton, Morningstar's Director of Equity Research for Technology, said.
The headliners among the downgrades: six wide-moat software companies — Adobe, Salesforce, ServiceNow, Shopify, Descartes and Manhattan Associates — all cut to narrow moat. Oracle and Thomson Reuters also lost their wide-moat status. FactSet fell all the way to no moat.
What ties the losers together is switching costs. Their competitive edge relied on the idea that once a business builds its operations around Salesforce or Adobe, moving off the platform is too painful to bother with. AI coding tools are chipping away at that logic. When a team can use Claude Code or Codex to build a custom alternative in days, the calculus for staying on an expensive platform starts to shift.
The companies that survived tell a different story. Morningstar maintained wide-moat ratings for Broadridge, Moody's, S&P Global, MSCI, Verisk and Fair Isaac — firms sitting on decades of proprietary financial data that an AI model can't generate on its own. Microsoft also held its rating, with senior analyst Dan Romanoff writing that "the firm should thrive regardless of AI."
Two companies actually came out ahead. CrowdStrike was upgraded from narrow to wide moat, with its fair value estimate raised to $460. Cloudflare earned the same upgrade. The reasoning: as AI agents spread across more systems, the attack surface grows, and cybersecurity firms with massive network effects become more essential. Cloudflare processes more than 20% of global internet traffic. That kind of scale is hard to replicate.
Morningstar isn't the only firm rethinking its playbook. William Blair, the Chicago-based investment bank, recently overhauled its own rating framework with AI as the core variable. "Against a backdrop where AI changes everything, shouldn't we change the lens through which we evaluate the companies under our coverage?" the firm's analysts wrote. Their review flagged N-able, an IT management company, as one that doesn't pass the test, citing weak execution and "questionable prospects for organic revenue growth acceleration."
Compton was clear this isn't a blanket doomsday call. "AI is certainly increasing uncertainty," he said. "We don't buy into the thesis that AI is just going to destroy everything … You have to try to sift through the winners and losers."
Even among the downgraded companies, Morningstar noted that many still have large user bases, popular products and important customer relationships. Some may even be undervalued after the selloff. But the net result is hard to ignore: there are now 169 US stocks with wide-moat ratings, down from 177 before the review.
Valley View
The dividing line from this review is clean: switching costs are the moat AI erodes first, and proprietary data is the moat it erodes last. If your business advantage is that customers find it too painful to leave, that's weakening by the quarter. If it's data nobody else can replicate, AI may actually make it more valuable. With Morningstar and William Blair independently reaching the same conclusion, this isn't one firm's opinion — it's becoming how Wall Street evaluates who survives the AI era.
