If there’s any point about AI that’s critical, not just important, it’s whether the “real” AI applications evolve and take hold before the hype wears off. If there’s a single company we’d need to watch to answer this, it’s IBM, and they reported earnings last week. IBM has, from the first, represented the AI viewpoint enterprises find can actually make a business case. If there are signs that this “hype-less” version of AI is slow in evolving, then there’s a good chance that there will be a slump in AI spending, and stock prices. So what does the call show? Here’s my analysis.
The opening comments from IBM’s president and CEO Arvind Krishna set the tone; “Technology remains a key driver of growth and competitive advantage. AI adoption is accelerating, and hybrid cloud remains the foundation of enterprise IT”. Over the last five years, the tech-heavy NASDAQ 100 returned over 125%, the S&P 500 99%, and the Dow industrials 65%. Tech has been, and continues to be, the key driver of the economy. Hybrid cloud has actually dominated cloud computing, other than for OTTs, from the inception. AI adoption is in fact accelerating, though just how that’s happening and why is a bit harder to unpack owing to the contrast between the publicized forms of AI (which largely don’t pay off) and the sort of AI businesses in general, and enterprises in particular, are increasingly adopting.
Krishna followed up with this: “Software growth accelerated to 9%, led by strength in Automation. Automation was up 22%, highlighting our end-to-end portfolio of leading solutions that optimize operations, automate infrastructure and workflows, build resiliency, and drive cost efficiency for clients.” He adds that “Many of our automation products are infused with AI enhancing their capabilities.” I focus on this because it’s important, as I hope to show.
In cloud computing, IBM has been exceptionally influential, not just because they didn’t follow the “everything moves to the cloud” story, but because they explicitly ducked it. Hybrid cloud is a cloud model that admits that cost optimization and governance constraints will dictate that only the front-end piece of applications can logically be hosted in the cloud, because the cloud’s benefits accrue to applications that have highly variable and distributed workloads. Transaction processing, the inevitable conclusion of any profitable activity, runs at the pace of the business and demands exceptional data security, so it stays internal. IBM was practical.
In AI, IBM from the first presented it in a business context. Analytics, meaning business tools designed to draw conclusions from data patterns over time, was a logical focus, so IBM saw AI as an extension of business analytics, and they’ve been right. So why not push that story here in the earnings calls? Same reason why IBM didn’t push hybrid cloud after the opening. Been there, done that. IBM is sending a message that the next frontier is real-time automation, and that like cloud computing, this mission is anchored in data that’s almost surely going to stay hosted on premises.
IBM’s AI story is built primarily on “AI agents”, but IBM implicitly adopts the enterprise view of what an agent is, which is that it’s an application component that performs a specialized task with oversight rather than continued direction from people. The agent model builds AI into things, rather than using AI as a kind of work partner that you can ask questions of. It’s a tool, not an artificial person.
Despite the fact that IBM beat in revenues and earnings, and despite the fact that software was up 9%, with “automation” up 22%, the Street seemed to think that the software story was a bad sign. The reason I’d offer for this is that hedge funds wanted a reason to sell the stock short, hoping to drive it down so they could buy into an expected continued successful execution at a better price. The stock did fall from about $290 per share to a low just above $270, but it closed last week at $307, showing that most traders (and hedge funds make up most of the Street trades and volume) believe the story IBM is telling.
This is absolutely critical for AI overall. The cloud was overhyped, like pretty much everything in tech is these days. 5G was overhyped, and AI was and is overhyped. 5G shows that in some cases, when hype evaporates it leaves behind an insipid reality. The cloud has demonstrated that while the remaining reality may be less than the hype wave suggested, enough will remain to be a good business. The question for AI, then, is whether it’s a 5G or a cloud, in terms of hype/reality relationship. What IBM seems to be showing is that it’s going to be the latter. It’s still too early to say definitively that pragmatic AI will take up the slack as AI hype fades, but it’s not falling behind at this point, and that’s important.
You could argue that IBM’s endorsement of the notion that AI agents are a means of implementing both business intelligence and automation isn’t decisive, but IBM’s strategic influence over, and symbiosis with, key enterprises is real. In addition, there are other indicators. For example, one AI startup has secured what are effectively both business intelligence and automation missions for AI agent technology, even though the company doesn’t call it that, isn’t connected to the AI hype wave, and frankly doesn’t have great marketing. Their wins are with major companies, and few have ever heard about them.
A recent post on “X” says that Nvidia has been artificially promoting the giant-cloud-chatbot model of AI, providing substance for the AI hype. I can’t validate this, though I think it’s surely possible. In any event, I don’t think that an artificial GPU market can be sustained, so you’d have to assume that Nvidia is working toward supporting a real AI market, which of course it is. Their position in the symbiosis between AI software (really, AI agents) and digital twins is something I’ve blogged about in the past. So if the “X” post is true, then Nvidia is covering the transition period when hype becomes reality.
This also demonstrates the stakes HPE is playing for in their own shift, the AI push that accompanies their Juniper acquisition. The IBM quarter demonstrates that its move is smart, but also that it is likely not decisive enough. If IBM is already seeing success, if AI startups can make a go of the automation agent story, then the market is already developing there. You can’t afford to be behind an emerging tech market.
