NVIDIA Unveils Groundbreaking AI Technologies at CES 2026 | Utilizing AI Ep. 8

We're kicking off 2026 with a look at the first announcements from Nvidia and a MD at CES. Vera Rubbin is shipping Helios here and a MD and Intel launched embedded processors with AI capabilities for mobile and Edge. We're also seeing more AI in the physical world with it NVIDIA's, um, all pyo, uh, Osmo and Cosmos.

Nvidia is also Aqua hired, uh, grok, which suggests an increased focus on inferencing in 2026. And we'll look at Accenture's acquisition of faculty AI and predictions for the rest of the year. Welcome to utilizing ai, the podcast focused on practical applications of artificial intelligence from the Futurum group.

Each episode brings together diverse perspectives to explore news and use cases of the ways in which AI is transforming enterprise IT and the industries it serves. I'm your host Steven FoST, president of the Tech Field Day Business Unit here at the Futurum Group. And joining me to kick off 2026 are two fantastic panelists.

Let's meet who they are. Hi everyone. Brad Shiman.

I'm VP Practice Lead for Data Intelligence, analytics and Infrastructure at the Futurum Group. And hi, I'm Nick Patients on the AI plat platform's practice lead at at rum. And as I said, I'm Steven Foskett.

I've been covering AI here on utilizing tech and utilizing AI for a few years. As I said, we are recording this, uh, January 6th, uh, right away here as, as 2026 has started. And also today is the start of CES, which is one of the big industry groups, uh, industry expos that happens every single year.

Now, CES is not an enterprise tech conference, but there certainly is a lot of AI happening. Uh, so although we can't dive too, too deep because not all of it's happened yet, uh, what's caught your eye coming out of CES so far, Nick? I guess the, the inevitably the NVIDIA's announcements around Vera Rubin, although, yeah, not ne, not necessarily news talking about, you know, production in volume of shipping later this year.

Um, and then, um, AMD's kind of counter with, its with its Helios, um, you know, rack system. Um, and it's just interesting contrast the way, you know, the two companies, um, you know, are operating in the, in the space. Obviously, Nvidia, um, has a, has a dominant market share as we know in it in GPUs.

And so a lot of what it's talking about is increased vertical integration. Um, so not only hardware and not only chips, but also env, ENV link. And then of course the crucial, the software element, um, including a bunch of new frameworks, um, that it put on GitHub.

And then a AMD obviously has to, you know, reply with a more multi-vendor, um, ecosystem approach, and it has that UA link alternative to ENV link. So I do think it's an interesting, um, you kind of rail of what happens down at that, that chip layer in, in ai. And I think it's interesting to, to think about the Vera Rubin, um, not as a chip like we did when we thought about, or we heard about Blackwell for the first time, for instance.

Um, but it's more of an architecture. It's more of a platform, like it's a, its own little supercomputer. Like, like you said Nick, you know, it incorporates memory, networking, storage and compute.

And, uh, you know, that alone makes it something that's, that's going to, I I would say, appeal to a lot of where investment is heading right now within the AI space for data centers, uh, for companies looking to maximize that vertical integration that they, that they see within in, you know, innovations like the Vera Rubin platform. But for me, I wanna say that what struck out, struck my, uh, or tickled my fancy with was cs, uh, was the return of the Dell XPS line. Okay.

Zero ai, but, but a beloved brand. Uh, hopefully everyone listening is old enough to know what that is. Oh, yeah.

To, yeah. So these, these are like gorgeous laptops built for gaming, built for, for high performance. And, and I, I applaud them, especially in this time period wherein Ram is so difficult to get in rolling out, you know, this 32 gigs for under two grand with a, with that name as, uh, a backer.

Hmm. Love it. Yeah, absolutely.

And, uh, I, you know, if anything can, uh, should be competing with, uh, with the, uh, the prevalent, uh, MacBook Pro, uh, contingent out there, it should be Dell's XPS. I know, uh, our friend Keith Townsend is a big, uh, fan of the XPS line as well. Uh, oh, is he just a great group overall.

Uh, you know, and it's interesting on that note too, in that we're seeing a little bit more diversity from these companies in terms of these, these platforms. Um, a MD is here with, uh, rise and embedded processors. Intel's Panther Lake is here, which has, uh, AI capabilities as well.

So, um, on the client side, um, you know, if if 2025 was the year of the AI pc, I think 2026 is also the year of the AI PC And the Edge too, right? Yeah, Steven and, and Nick, it's, uh, when you look at Nvidia rolling out their Jetson 7,000, I think that was at CES, and you get basically a Blackwell architecture, uh, if for under, in under 70 watts, uh, power consumption, um, that, that's pretty amazing. For what, what, two, 2000 again, what a price 200 bucks, uh, assuming you buy a lot of them.

It's, uh, it's, it's an interesting time for hardware right now, and I, I we're seeing a lot more diversity than I would've expected in this market. Yeah, It's interesting kind of as we, you know, interesting sort of base level for, for the robotics, uh, and the physical ai, um, you know, explosion that we're expecting to happen this year and, and, and next year, I guess. And, uh, obviously those companies we mentioned are, are gonna be at the forefront of all that.

Yeah. Let's talk a little bit about that physical AI concept, because of course, uh, Nvidia is here with, um, le let's call it Alpa Mayo. Uh, I wasn't at CES, so I don't know how to say this.

Um, ALPA Mayo, uh, Brad, what is this? None of us know. Uh, it, it's, it's actually their set of models are family models and supportive, um, you know, frameworks, uh, all geared around robotics.

And so, um, you've got specializations for things like Alpha Sim and you have, uh, a bunch of frameworks for designing, um, uh, training runs for robotics. Like with, you have a visual visual model for a robot that's a, a restaurant robot, let's say that there are very specialized training runs that you must do in order to train the robot to flip an egg properly. And with Alpa Mayo, um, what we see from, from Nvidia right now is this like massive specialization, uh, of, you know, helping their customers and their partners build out this, the robotic solutions on top of their own, uh, set of models like, uh, the Nitron, uh, three family, and which includes grot, which caught my attention in particular because, uh, you know, Nvidia released at the same time a full collection of like thousands and thousands of, of data, uh, data points for training.

Sorry, that's the wrong word to use, but you know what I mean. They, they've actually opened up, uh, training data that's completely open source. It's on GitHub, I'm sorry, not GitHub, but hugging face.

You can download it and use it, uh, however you see fit. And they, and the Osmo, what was it called? Osmo Robotics, um, workflow development cycle thing, um, for orchestrating, you know, synthetic data generation, um, using their, uh, cosmos, uh, world models as, as, uh, NVIDIA calls calls them.

And so I guess that that's, uh, you know, that's, and then I think they also announced the cs, you know, more cosmos, um, foundation models, um, transfer, predict reason, um, and otherwise other ones like that. So yeah, it's a lot of, um, you know, as, as we know with, with NVIDIA's announcements, uh, although everybody kind of focuses on the chips and love to see gen standing there holding something physical, um, yeah, in order for, you know, all these things to actually, uh, make a difference in the world. This, the software is increasingly is, it's so important.

It's absolutely crucial. Um, and, uh, you know, NVIDIA's extremely well placed, um, there, and a lot of it's open source, um, but there's also crucial layers are not. And so, you know, that's how, uh, that's how they play, play, play.

Well, Coda. Yeah. Yeah.

Coda the Ultimate. Absolutely. Yeah.

Totally. Well, uh, you know, I mean, who, who can blame 'em? I mean, they are a, a, a for-profit company, and they're certain, certainly raking in the profit right now.

Um, we also heard of an acquisition, or sort of an acquisition by am or by, uh, Nvidia, uh, over the, the holiday. Um, one of the companies I've been following closely for a long time is Grock with a Q, not the other Grok that also made news this week for a reason that I'd rather not talk about. Uh, again, um, grock with a q, they, they were, uh, an early hardware startup in the AI space.

A bunch of Google engineers got together, um, lately they've been, um, using their, uh, inferencing hardware as a, uh, platform, uh, d you know, basically developing, uh, enterprise applications around ai. It's great stuff. Uh, apparently, uh, the company has been acqui hired by Nvidia, um, uh, even though it still kind of exists.

Um, we talked about that on the Textron gang. Uh, what's your take on Nvidia plus grok with the QI mean, I, yeah, it's, it's an, it's a weird, um, well, from what we know, it's a slightly weird deal though, isn't it? It's, it's kind of, what was it, $20 billion to acquire assets and, and IP and, and, and people.

Um, but there's also kind of non-exclusive nature to it. So I think, you know, because CRO was building these, uh, language processing units, LPs, um, you know, and so I think, I guess that's what they were. And they were, and they were extremely, um, performant and in real time, uh, token generation.

And I guess that's what they're, uh, you know, looking, looking, looking to get. Um, so it's, it's, again, I, one of these, you know, NVIDIA's model or not models have shifted, but its business model has evolved from, you know, training, training and training to inference and grew up with very much an inference shop. Um, and allegedly still is gonna be one.

Um, but it's not entirely clear what it's, what it's gonna do with its key people not there. Um, and, uh, some of it's a license, um, it's IP now in, uh, in hands of Nvidia. Yeah, I would, I would imagine that it's gonna play a, a nice, you know, balancing role within NVIDIA's NIMS architecture, for instance.

'cause that's all about containerized, you know, instances where you can stand up a fully working stack and just start inferencing off of it. And if grok is, is, you know, an alternative to a third party, like open router, uh, for instance, that, you know, Nvidia users would, I'm sure appreciate that. And as you know, Steven mentioned they've been around for a little while.

Grock has in a little while being less than five years. Um, but, uh, nonetheless, they, um, you know, I, I think we're very early to, to tackle what we see right now as such a huge area of investment in terms of, you know, highly specialized chips for inferencing. And you just look no further than AWS reinvent we had in early in December last year, where, uh, all, all roads pointed toward train, uh, and how important that had become to, uh, AWS not for training, but for inferencing experiencing.

Boy, they named that one wrong. Um, no, I I what's In their name. I do think that, that you're onto the, onto it there, Brad.

Uh, when I thought of, of grok and when I thought of Nvidia, I, I, my first thought was exactly what you mentioned, what we heard at reinvent, what we heard from Google Cloud, uh, what we've been hearing from Microsoft and others about, um, using specialized hardware to do inferencing, to do token generation. Uh, that's really groks strong point. That's certainly what Google and Microsoft and AWS have been leaning into, along with the software platform aspect of it, which is something that Grok has been leaning into.

And so I, I do caution listeners that, uh, this is, this is very new. We don't yet know which components are going where and what the angle is gonna be here, but I could certainly see a situation where, you know, Nvidia is part of that conversation as well, just like Google Cloud, AWS and Azure, right? Yeah.

But they also partnered deeply with, with all of those, it's, uh, of course they do a highly cooperative, competitive marketplace. Yep. Yep.

Now, another acquisition, um, we just heard about is Accenture. So, uh, not that Accenture was purchased, but that Accenture is buying, uh, faculty AI and, um, promoting, uh, the head of an AI company to become the CTO there. Uh, I know that there's been a lot of, uh, fear, uncertainty, doubt, and mischief around, uh, ai, uh, replacing people.

Uh, Accenture is the people company. Uh, what's, uh, your take, Brad on the Accenture, uh, faculty story? Yeah, I, I don't know enough about faculty to say technologically what impact they're going to have.

I think instead, what this points to is, um, you know, a, a sort of raising of a, of a banner at Accenture to say that we are an AI company. We are not a people company anymore. And I guess you could call them accelerationist for that.

Uh, and, and for better or worse, um, but, uh, you know, I recall visiting them in New York City, uh, two years ago, and having them demonstrate their in-house AI that they were building to, to help optimize and, you know, 10 x uh, their employees to, to be better, uh, at their jobs. And, uh, I'm sure nothing has changed since then. I'm sure that that is still what their objective is here, is to become a, a much more efficient company at helping their customers become more efficient companies through the adoption and use of ai.

Yeah. I, I mean, here in London, uh, faculty AI is, is, is quite prominent. And, um, they're quite well known for government work.

Um, mm-hmm. Yeah, I heard, I read something this morning trying to make out the, you know, that the UK's equivalent of Palantir, which is, um, you know, size wise, not really comparable, um, but it did have this kind of, you know, they called it a decision intelligence platform, um, called Faculty Frontier. And again, I think it's kind of these four deployed, um, engineers into, into customers.

But they, they had a lot of government contracts going back many years. I think they're only, well, they're about, uh, 12 years old as a company. Um, but yeah, that CEO Mark Warner, if I, you know, I go to a lot of, you know, a lot of London AI events, and either he or somebody, somebody else from that, that company is often there.

They usually keep themselves, um, to themselves. They're quite quiet about what they do, so it gives you an idea of what some of the work is, um, that, that, that they've done. But yeah, with him becoming Accenture's CTO, it does, uh, it does kind of reinforce that realignment, they said back in September, October, I think it was of last year, um, realign around AI laying off, um, um, up to 11,000 employees even over time.

Um, not in one go, obviously, um, attrition and all the other kind of usual words that are used, put it, put it in perspective. Um, Accenture has 800,000 employees. Um, so it's, it's, it's, it's a, it's a small drop, but the, but faculty does a lot of work with OpenAI, with anthropic, with the, what was the AI UK AI Safety Institute, ai, UU AI Security Institute, um, and the stuff like that.

So there's a lot of kind of AI safety work. While I, I'm very skeptical about some of that emphasis around AI safety, I think is a little bit, um, um, kind of trying to induce some sort of, you know, you know, panic around what AI does. They, they, they were pretty, um, heavily involved in that.

So I think it's, uh, I mean, I, it's, it's an interesting, it's an interesting, um, time and I, I, I sort of wrote a little note about it, um, on very little note on LinkedIn. And my last, my last three words were kind of re-skill or exit. So it's kind of, it's, it's that kind of thing.

Like if you are, if you work for Accenture or you work for any kind of company where, you know, knowledge work is, is, is key, um, you've gotta, you've gotta adopt ai. Um, and I think that that's, that's pretty blindingly obvious statement. But it was interesting, I noticed when they announced their last results, which were announced in December, um, they put a, they kind of, they have a, a series of charts, you know, QR Q1, I think it was advanced AI bookings, advanced AI revenues and that kind of thing.

And then at the bottom, it had a note saying, this would be the last quarter in which we share, um, advanced AI bookings and revenues. We've reached a point, I'm reading it now, where advanced AI is in being embedded some way across nearly everything we do. Um, and so, et cetera, et cetera.

Um, we're not gonna then split it out. So, you know, when when does, you know, AI is, uh, the old, the old gag is AI is whatever, it doesn't work yet. Um, and so, you know, that yeah, they kind of, they're seeing it as so in, so embedded in e everything they do, um, that, you know, it is just, it is business.

Um, it's not a, it's not a necessarily a peculiar thing. Uh, obviously as the leader of the AI platforms practice, I would, I would, uh, counter, there are some unique things to ai, um, and will be for, for, for many, many years to come. Um, but for a company like that, to make that kind of statement, uh, and bring in somebody like that, I think it's, uh, you know, it's a, it's a, it's an interesting marker for the rest of the industry.

Yeah, it does reflect the nature of business, um, that a company of Accenture's status. I mean, as you said, this is a company with, you know, almost 800,000 employees, you know, billions of, of revenue. Uh, I think it says here, 120 locations.

I mean, Accenture is everywhere. They're working with basically every major company. And to have a company of that status say, effectively, AI is just part of what we do, and part of what everybody does, I guess, that reflects, uh, where we're at in 2026.

Frankly, it sounds a little bit like Futurum. I mean, that's what we're doing here as well, right? Yeah.

And I think we're gonna move from the away from the ai ai experimentation years of 2023 and 24, 25 was, uh, a new set of experiments, um, with, with Agen. Um, and I'm not saying in 2026, ag agentic is gonna be completely, you know, um, mainstream, which it isn't. Um, but other elements of ai, um, you know, maybe more on the kind of, you know, predictive model side, um, are, are just, um, you know, the meat and potatoes of, uh, of, of, of how, uh, technology gets adopted.

I would say, you know, with Accenture, um, in particular, what this makes me think of is that, uh, these companies, futurum included, are right now not endeavoring to, you know, enter, you know, into an opportunity where they're making ai, but instead they're using AI to do their job, to do what they do as a company. 'cause that, that early meeting, I, I mentioned it was all about, you know, we're, we're gonna have our own, you know, chat bot model with its own anthropomorphized name, and it's gonna, you know, be so and so to, to do what you get from, you know, OpenAI and others. And I think we're seeing a shift where companies, uh, are starting to look at this as just how they do business, not the business they do.

And you could see that reflected in how they, you know, talked about and positioned their earning statements. I think that that's very telling. And, and can I get a hallelujah on that?

Um, I, I, I think I should remind, uh, listeners that, uh, this is called utilizing ai. I named it utilizing AI in 2020, because I was waiting for the time when we were making practical use of this technology, not when it's a science project, not when it's something cool, not, not when we're doing it for the sake of doing it, which frankly sounds like a lot of AI out there right now, is that we're just sort of playing with it. Um, you know, I mean, now we're actually using it for productive purposes.

And I think that that's what I want to see in 2026. I want to see this being used for productive purposes that let us do other things, the things we always do, let us do them better. Um, you know, Brad, I've said that, uh, you know, the, the, the futurum use of AI and the, and the, the, the specifically, the, the signal report that I know that you've been deeply involved in, it's not that it's ai, it's that it's, it's, it's that technology that gives you and Nick and the rest of the folks a superpower when it comes to analyzing data.

And that's, I think what Accenture is saying is that AI gives their clients, um, extraordinary abilities to do the things they already do. And that's what I'd like to see more of in 2026. So, uh, let's talk a little bit about what we're looking forward to, Nick.

I know that you've had some questions recently about, uh, what are the big momentous things that are gonna come out in 2026. Could you maybe, uh, reiterate that for our audience? Uh, what, um, you know, what business moves, what IPOs, what actions do you think are gonna happen in 26 that are gonna really shake the foundations here?

Sure. So on, yeah, on the financial side of the, rather than the sort of technology side of it, I think the, um, we're reaching that stage of maturity where, um, exits happen. Um, companies looking, you know, looking for looking for exits either, either through m and a or through IPOs, although the, although the IPO market for enterprise AI is being pretty quiet, um, from what I hear and what I see out there, you know, there's a lot of optimism that it will pick up.

I mean, mind you, that that always happens in January. There's a lot of optimism is gonna pick up, which, you know, you have to see what the state of the actual market is, whether it's ready for it. But you've got open ai, um, and anthropic, um, really as the two, the, as the two kind of, uh, poster children as it were, of, of, uh, not only enterprise ai, but also consumer AI in terms of open, open ai, a lot of ai.

Um, but you know, those, those are both, um, potential IPOs this year. I think they're the, they're interesting contrasts in valuations, uh, to say at least. So OpenAI, maybe rumor is maybe looking for a $1 trillion valuation.

Its most recent round was at 500 billion. So there's a, there's a bit of a gap, um, there, which it, which it might. Um, you know, it's, that's kind of a price for perfection scenario.

Um, you know, everything has to work for it to be worth that much money. I mean, not say it won't get it, um, but it, and then there's also the interest around Microsoft's stake. Obviously Microsoft was incredibly prescient in 2019 with what it did.

Um, but there also, that's, it does create a little bit of a messy, um, kind of scenario from a governance point of view. Um, so that, that will be interesting. Andro is a more kind of enterprise safety hedge, and the kind of more, you know, it's kind of the, the more sensible looking, you know, enterprise AI company at and at a low evaluation, um, maybe, you know, 300 billion, um, some something like that, who knows.

Um, but it's, but it's, you know, so I think it's, if if investors open AI should be the bell, you know, would be the kind of bellwether AI stock. If it went, if it went, um, but philanthropic could be the kind of more conservative, um, uh, alternative. And then if you think about, you know, what, what's the purpose of IPOs is to raise money and to provide liquidity and all that kind of stuff.

So when they raise the money, where does that money go? What is the use of the proceeds? Now, you would've said maybe if this was in a fantasy land, this was happening 18 months ago, you know, Nvidia would be, you know, vacuum 'em up, all that, that extra new cash.

Now it might be in, it's obviously with, with these companies both working on custom silicon, um, maybe that's not gonna be the case. So, you know, where do the, where does the proceeds go? Um, so I think, you know, a lot of it will go towards custom silicon development.

Um, I also think it will go towards energy infrastructure, um, and obviously data centers and, and, and all that kind of stuff. So I think it's, um, you know, those, that, those that have the, those stocks or those companies that have the kind of clearest path or the clearest picture of owning a stack and the stack these days doesn't start with the computer start, you know, essentially with power, um, and, and land and things like that. Um, you know, those have the biggest, you know, the clearest path for that would be, um, probably get the most rewarded and would, in theory at least be, uh, can take on hyperscalers, at least in terms of how the public market sees it.

So yeah, just from a financial point of view, there's many other things we think is gonna happen this year, but then that's, that's gonna be, um, an interesting one to look at. Yeah. And for me, I, I feel like if I were to try to see what's, what's coming for the entirety of the year for the space that I look at, uh, I would say that this is definitely, you know, the morning coffee after the hangover that I think we've been going through a little bit with AI and what it can and can't do.

And, um, I think we're gonna see a, a tremendous focus on, um, you know, the science projects are over. It's all about if it doesn't scale, if it doesn't make money, it gets cut. And this goes to what Nick is talking about with this emphasis on, on ships and vertical integration all the way down to the power coming into your data center.

Um, but it's also about, you know, how that software stack actually works, um, in supporting that optimization. Because as we all know, you know, we, we've gone through a bunch of phases in the industry where we've tried to have these sort of, you know, do everything platforms best of breed that, you know, have a sense of lockin, but also have a sense of performance to them because of that lockin. Um, and what we're trying to understand right now is, well, can we have our cake and eat it too?

Can we have a highly composable stack that still emphasizes speed? So you see, um, companies like, uh, Microsoft with their, their Cosmos db for instance, rolling out, um, you know, uh, basically caching, um, semantic caching within that database to support agentic workflows, uh, directly, you see the sort of death of, you know, ideas like data meshes and even data fabrics to be replaced by a, a more composable, um, landscape or data estate, as we like to say, that's, that's built on top of, uh, a semantic layer, which, uh, is, you know, some something we all, we all really need to get behind, and if, if we aren't hiring people that understand what the word epistemology means, we're already losing. So I, I think it's, it's gonna be a very interesting, um, you know, vendor community and hiring system situation in 2026 in terms of what we hire for and how we build our solutions.

It's, uh, not anywhere we've been before, Just to add to a few other things that, that, um, my practice will be looking at this year. And, and I think maybe if there's some things we'll have, there'll be future topics of, uh, discussion on this, on this podcast, um, sovereign AI and kind of the corporate con nature of corporate control, um, that's, that's gonna be a big, uh, focus this year. Um, it already was for, you know, towards the back end of last year, but I think, you know, increasingly that will be, um, similarly related is kind of global AI regulation and compliance.

The, um, EU AI act really does, does come into force properly in, in, in 2026. Um, and obviously we've seen in the us, although there are various state things going on, um, you know, Trump has made it very clear that, you know, wants a federal, um, AI regulation. So there's not as, say, 50 different regimes, um, to deal with, um, obviously Ag ai, if not at scale, you know, rolling out you more fully.

Um, and, uh, I guess down at the infrastructure layer, we talked about it just now really, but the energy and cooling as a kind of bottleneck, a bottleneck for data center built, build out, bottleneck, bottleneck for all, you know, the stuff that goes in data centers. And I think that will, um, that will become a major, major issue. And as we kind of said earlier, would probably result in some sort in m and a, uh, we've already seen it to a certain extent, and I expect we'll see more of that.

And then the kind of fragmentation of models, which I, I don't mean breaking up of models. I mean the, the landscape, you know, LLMs are really important. Um, they're not the only game in town.

Small lang small language models, um, will become more and more, um, prominent. So yeah, those are some of the things that we're looking at. I does, when I'm sort of, when I was drawing up the list, I was thinking this is getting quite infrastructure heavy.

Um, AgTech is, is kind of where the rubber, but where the business value will be created. Um, and, but, but there's a lot of stuff going on underneath, um, that we, that we'll be, uh, spending time in 2026 and say, we, we should probably, we dig into one or two of those, uh, in the coming weeks. Yeah.

And another point I'd like to make too is that, uh, the AI infrastructure market is by, by no means dead. Uh, there's still a lot of money flowing into AI infrastructure, networking hardware, uh, you know, I mean, you look at companies like Cisco, um, you know, broad Broadcom and, and companies like that, that are doing so much, uh, chip work. Um, and, and for me, one of the, one of the dark horse, um, heroes of AI that I'm looking at is Google Cloud.

Um, I'm hearing such good things from enterprise users about, uh, Gemini, uh, as an alternative. And, you know, you, you, you talk about, uh, like you did Nick about these, these, you know, vertically integrated stacks, uh, don't count the hyperscalers out, I would say. Um, they, they know how to run data centers.

They know how to run platforms, they know how to work with enterprise companies. Uh, they've, they've been learning it the hard way for 20 years now. And I think that, um, you know, companies like AWS are really going to roar forward, um, Microsoft, Google, uh, in 2026 as, uh, applications become more practical.

Uh, that would be my prediction. Don't count the hyperscalers out. Oh, yeah, I totally agree.

Steven. I just wanna add really quickly to that, that, uh, it is a, has amazed me as, as both a practitioner and an analyst in looking at, um, Google Cloud in particular in terms of how they're investing in their APIs, uh, around Gem Gemini. And it seems like almost every couple weeks they reinvent what we took for granted before.

So we used to like scrape web pages to, to turn them into, you know, to embed them in a rag pipeline. And they're like, no, we own the search index, so why don't we make it so that you don't have to scrape webpages, you just query the webpage and get all the data from it. And they're, they're doing the same thing again.

Uh, and again with different, you know, aspects of what was expensive, uh, and, and time consuming in terms of, you know, optimizing the inferencing cycle. So it is, the money that they're spending is all about the gravity still has been, always will be about data gravity on their platform. But my goodness, you know, with it, like you said, AWS and Microsoft and Google are seriously investing in, in making the, you know, giving, giving enterprises the tool set they need to, to do this the right way.

Final words, Nick? Yeah, just on that, I would say, you know, I think of, of the three hyperscalers we're talking about, I would say Google probably had the best year of 2020, um, in terms of improvement of, of its, of its position in the market. Um, but we've reset now we're all back to all back to zero in the kind of analyst ranking game, and, uh, and we go again.

And so it'd be really interesting to watch, uh, all three of those and, and many others. Um, maybe not quite hyperscalers, but other cloud providers around the world. And as we say, the on-premises, um, shift and, and push because, you know, the, the, the likes of, um, Broadcom with VMware or Red Hat with its software stack, um, selling through other cloud providers around the world, there's an awful lot of interest in, in that stuff as well.

So, uh, yeah, it's, um, it's not all about infrastructure. There's a lot of infrastructure stuff, um, to watch, um, to enable all this good ai ai stuff to happen. Yep, absolutely.

And, and I'll just, uh, throw in a little thing. We're, uh, gonna be holding a, an AI infrastructure field day event, uh, end of January. Uh, we're gonna be hearing from some of the companies that are building, uh, AI infrastructure stacks.

Uh, we'll also be doing another AI Field Day event. Um, and as we joke, every field day event is AI Field day. Um, so thank you very much for joining us, both of you.

Before we go, um, give us a hint of what you're gonna be working on in the, the coming quarter in terms of research at futurum. So why don't we start with Nick this time? Um, so yeah, we're gonna be redoing our, um, our AI platform signal, um, and we're gonna look at some, some other, um, sort of, uh, sub-sectors of, of that, um, potentially around, um, sovereign ai, maybe I haven't quite made my mind up on that.

Um, I'm redoing my, um, decision maker survey. Um, so that's gonna be going into the field this quarter. So that will be a, you know, a, a re-up of that.

We've only just published our, our, um, market forecast in, in December, um, our five year market forecast, which is, uh, which is really interesting. Um, so those are the, those are some of the, some of the things I'm, I'm gonna be working on, Brad, it's similar to me. For me, uh, we're, we're twins obviously.

Um, so I'll, I actually have a second forecast that, uh, we're doing for data intelligence, analytics and infrastructure coming out. That'll be five years as well. And also fielding a, a new survey, uh, for decision maker, uh, we call it decision Maker Survey.

And I also will be updating my, uh, data intelligence platform signal, um, actually really quickly. So that'll be something we'll publish in early February for that. And, uh, I have a new one that I'll be working on, I'm really excited about.

It's gonna be about Semantic bi. So, uh, if for, for all of you guys who love your love a good dashboard, uh, I would invite you to tune in for that one. Well, that, that's really the thing about Signal.

And the thing that is exciting about what Futureum Research is doing is the fact that you can refresh this data much, much more frequently than ever before, uh, thanks to the, the tools that we're using on the backend. And it just means that stuff is more valid because, as you said, there, there are things being introduced every single day in this market. And so, you know, you cannot rely on a year old report when making decisions on pretty much anything in enterprise tech these days.

So thank you both for joining us. com. Uh, please do subscribe to the utilizing AI podcast.

Uh, you can also catch our other podcasts with the security, uh, Boulevard Podcast, uh, agents of Dev, uh, as well as, of course, the Tech Field Day podcast for more. Um, if you enjoyed this discussion, please do subscribe. Uh, you can find us on YouTube, uh, as well as in your favorite podcast application.

Uh, please do give us a rating review and a comment we would love to hear from you. Uh, this podcast is brought to you by the analysts and experts from the Futurum Group, where Insight meets ai. For show notes and more episodes, head over to Text Strong Do ai, which is our sister news site, uh, the utilizing AI YouTube channel, or of course, the Textron TV app, where a great many people apparently are watching this according to our, um, our stats.

Uh, the majority of people seem to watch this on the, uh, on, uh, TVs and uh, in, in, in formats like that. Thank you for listening, and we will catch you next week.