This past week, IBM Corp. hosted an annual analyst forum in New York, where CEO Arvind Krishna and top executives unveiled key developments for the company.

The AI-first strategy was one of the big subjects for theCUBE Research founding analysts John Furrier (pictured, left) and Dave Vellante (right) to dive into on the latest episode of the CUBE Pod.

“Krishna is basically saying, ‘I would do more acquisitions if I didn’t have regulators on my back,’” Furrier said. “He’s itching, you can see in his eyes. He’s got a new culture change. They’re betting on an open ecosystem, they’re streamlining the product portfolio.”

The IBM AI-first strategy in key focus

Under Krishna’s leadership, IBM has more focus these days, according to Vellante, citing the widespread effect of their 2018 acquisition of Red Hat Inc.  for $34 billion.

“That has enabled IBM to execute on a hybrid cloud strategy, and it’s given them access to OpenShift, the leading container platform. They’ve grown that business,” Vellante said, citing Krishna’s statement from the forum where “he said they grew it from, I want to say $150 million, to now it’s well over $1 billion, headed towards $2 billion, so it’s a 10x revenue business for them.”

The IBM AI-first strategy has also come into focus. Watson 2.0 or watsonx is much more focused than before, according to Vellante.

“It’s not trying to be everything to everybody,” Vellante said. “And of course, he mentioned that with Watson 1.0, they tried to automate everything, and he said, ‘That type of strategy didn’t work. So now we’re really out targeting real problems.’”

IBM’s stock is also at an all-time high. All of this suggests that IBM is back, according to Vellante.

“I wrote a post on LinkedIn … nine months ago, [that] I haven’t been this excited about IBM in a long, long time,” Vellante said. “I’m happy that the stock has performed subsequent to that post because it’s always good to be right. And I think the vibe was, we got our mojo back and Arvind Krishna is going for it.”

There were also additional indicators as to what the IBM AI-first strategy may look like. When a company is approaching a hundred billion dollars, they can take a lot of cost out, according to Vellante.

“[Krishna] said that IBM has realized $1.6 billion in cost efficiencies to date, with approximately half of that attributed to AI deployment. This is around automation, this is this year,” Vellante said. “He said that he aims to achieve $3 billion in efficiencies in 2025. And he also said that 75% of their service queries to IBM are now addressed through AI-powered self-service platforms, and that’s been a culmination of what he’s called a five-year journey. So that was pretty interesting.”

The war of words between Snowflake and Databricks

This week, SiliconANGLE Managing Editor Paul Gillin delves into why two AI giants are at war over who is more open. The post explores Apache Iceberg and how it has become the subject of an escalating war of words between Snowflake Inc. and Databricks Inc.

“I think that kind of journalism does good for the industry, because it’s not so much we want to create a war between Snowflake and Databricks, it’s more of they’re both going to be very important companies,” Furrier said. “I mean, both companies don’t lose in this growing market. I think you may have an opinion on Snowflake, but my opinion of Snowflake is they’re too big to lose at this point, unless they really screw up.”

The ongoing power problem

Other big news in the industry currently involves hyperscalers looking at nuclear reactors. When it comes to pressure from the industry around the performance needed to power AI, that’s spawned a renaissance of nuclear energy, according to Furrier.

“There’s a lot of questions there. One, how fast does it go and is that good or bad? We don’t know,” Furrier said. “Of course it’s good for AI if we get nuclear power, but do we want another Three Mile Island? You saw Chernobyl. So there’s a little bit of an interesting discussion there.”

If power can come online, then the limitations around power limitations for Nvidia Corp., for instance, goes away, according to Furrier. That’s interesting given recent discussions around whether Nvidia has a moat.

“And you talk about Foundry. Who else is doing Foundry? Everyone’s designing chips. So you have chip development demand, power demand, two huge constraints right now in the industry,” Furrier said. “And if Nvidia, which everyone says draws a lot of power in Blackwell and the future version people are talking about to us off the record, it’s going to even require more power.”

Here’s the full theCUBE Pod episode:

Don’t miss out on the latest episodes of “theCUBE Pod.” Join us by subscribing to our RSS feed. You can also listen to us on Apple Podcasts or on Spotify. And for those who prefer to watch, check out our YouTube playlist. Tune in now, and be part of the ongoing conversation.

Photo: SiliconANGLE