It’s been another busy week of events for theCUBE, with full coverage of UiPath Forward and Celosphere 24.

The trends emerging out of those events, including around AI production readiness, were big subjects for theCUBE Research industry analysts John Furrier (pictured, left) and Dave Vellante (right) during the latest episode of the CUBE Podcast.

There were a number of revelations that emerged out of those events. Among them was the sense that there’s no bubble popping in AI at all, according to Furrier.

“There’ll be air being let out of the balloon, but there’ll be no bubble popping,” he said. “There’s too much value being unlocked on easy use cases. Get on base, we say. Just get the first base. Get hit by a pitch, take the walk, whatever you can.”

AI production readiness is key

This week, businessman and author Andy Kessler wrote an opinion article for the Wall Street Journal that suggested that AI can’t teach AI new tricks. It might involve a good argument on paper, but there’s more at play here, including questions of AI production readiness, according to Furrier.

“The reality is most people use synthetic data for training simulations, not so much production,” Furrier said. “That was another thing that came out of this week … in these certain verticals where there’s zero tolerance for failure or hallucinations or drift or context poisoning or prompt injections, they will not put any in production at all until it’s vetted.”

That’s why this hardening of loops comes in handy, according to Furrier. Small language models get a lot of attention because people are having success with use cases that are domain-specific, where data can actually be vetted, meaning AI production readiness is key.

“Broad, sweeping AI has not worked for the enterprise. We know that. We talked about it last week at length,” Furrier said. “But Anthropic releasing improved Claude models is a good sign that the industry is doing its job of lowering the cost of tokens, increasing context windows and also increasing capabilities.”

IBM’s revenue miss and Intel’s bigger miss

Last week, IBM Corp. kicked off the quarterly earnings season and saw revenue misses. At the same time, executives said that AI would fuel future growth.

“The good news is the software business is humming. It’s now 45% of IBM’s business, which is great,” Vellante said.

Meanwhile, a recent New York Times article covers how Intel Corp.’s former CEO proposed buying Nvidia Corp. for $20 billion back in 2005. That’s a huge miss, and the company is now sideways, according to Furrier.

“It’s a case study, in my opinion, of a fall of an American institution,” he said. “Had they done that, it could have been a whole different Intel. We might not have seen that Nvidia boom.”

Perplexity’s big funding round and what’s next

Elsewhere in the market, Perplexity AI Inc. is reportedly discussing a funding round with investors that could increase its valuation to more than $8 billion. It’s a notable turnaround, according to Vellante.

“In January, this company was valued at $520 million, and this summer the valuation went to $3 billion, and now it’s hovering around $8 billion if they do this round,” he said. “I mean, that is insane how fast that valuation has grown in a year.”

It’s a notable supercycle, according to Furrier. They are also unique in history.

“If you go back to 1986 and look at that 10-year period from 1986 to 96 … we went from proprietary client-server and mainframes to open systems,” Furrier said. “That created an inflection point that drove massive amounts of wealth creation, massive amounts of transformation. That set the agenda for what would be the next run, up until the web, that next transformation cycle.”

That was a fast supercycle, according to Furrier. Generative AI is the latest supercycle.

“To me, if you’re not a winner in the first three years, you’re not going to be the winner. The winners are determined early on,” he said.

Over the coming weeks, theCUBE will continue to track the latest developments as part of its ongoing event coverage. There are a couple of big events on the radar.

“KubeCon. You’ve got SuperComputing. I’ve got the Dell analyst event,” Vellante said. “Obviously, we’ve got re:Invent at the end of the year.”

Meanwhile, lots of principal research has dropped on theCUBE in recent weeks that continues to draw significant attention, according to Furrier. That includes coverage on agentic AI.

“The agentic stuff is on fire,” Furrier said. “Getting a lot of good feedback on the work you guys have done there, and the team is doing.”

Watch the full podcast below to find out why these industry pros were mentioned:

Jim Cramer, TV personality and author
Peter Tuchman, stock trader
Brian J. Baumann, director of Capital Markets at NYSE
Andy Kessler, American investor and author
Jim Kavanaugh, CEO of World Wide Technology
Ginni Rometty, co-chairman of OneTen, former CEO of IBM
Arvind Krishna, chairman and CEO of IBM
A.M. (Toni) Sacconaghi, Jr., managing director and senior research analyst, US IT hardware and electric vehicles at Bernstein Research
Mohamad Ali, SVP and head of IBM Consulting at IBM
Paul Otellini, former president and CEO of Intel
Bill McDermott, chairman and CEO of ServiceNow
Steve Lohr, reporter
Don Clark, journalist
Andy Grove, former president and COO Intel
Ray Wang, principal analyst, founder and chairman at Constellation Research
Joe Kernan, director at RK Marketing
Tracy Kidder, American writer
Steve Jobs, co-founder and former CEO and chairman of Apple
Nicole Petallides, American journalist
George Gilbert, senior analyst at theCUBE Research
Scott Hebner, principal analyst for AI at theCUBE Research
Rob Strechay, managing director and principal analyst at theCUBE Research
Christophe Bertrand, principal analyst for theCUBE Research
Bob Laliberte, principal analyst at theCUBE Research

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