reddit.com - Ibms oneday crash the mechanism customers is a good pre-existing discussion which I'm linking to not rehash the timeline and numbers.

18x FCF is not bad with a clean balance sheet.

I'm really surprised by the degree of the reaction, and they called out the hardware shortfall (Z series mainframes) revenue because of customers prioritizing AI related hardware and panic buying capacity constrained dual use components like memory.

The maximally negative case - purely my take - is that IBM benefits to a large extent from a closed system: buy their hardware, run their software on it, and use their consulting to keep it running. A collapse in hardware has a ripple effect and could weaken all their business. If the purchases are cancelled and AI spend is replacing mainframe spend durably, then IBM is in trouble.

The positive case is that AI is in a pronounced bubble, and the purchases are just deferred. And IBM is a player in AI with Watson, and a major player in quantum computing, which is barely ramping yet compared to AI. And they have diversified with open source revenue streams with Red Hat, Hashicorp, DataStax, and Confluent.

The answer's probably in the middle. I don't think Watson's really a player in the way that Anthropic and OpenAI are. But the market is frothy, IBM is a real player, and companies like Palantir are grouped in the investing cohort without foundational models of their own. Replace Watson ai with a foundational model provider and there's sill probably a valuable core. I don't think AI and mainframe workloads overlap and compete that much. I am certain there are near-panics in procuring memory in competition with foundational model provider datacenter buildouts. I am also certain companies are wary of relying on foundational model providers for both pricing, capacity, and ensuring critical IP stays inside their security boundaries; many may be building out their own small scale inference infra.

A linger doubt is that I don't know why IBM's consulting is holding up as well compared to Accenture (ACN), both in the reported numbers and impact to the company valuation.

I'm not buying the drop, but I am sniffing around. I'd rather buy IBM than Micron.