娇色导航

[ ] Keith: Hi everybody.

Welcome to Global Tech Tales, where we talk with editors from around the world, around the latest technology and leadership topics, and hear stories about from IT leaders about what they are looking for. I'm Keith Shaw co-hosting along with Matt Egan.

He is the Global content and editorial director at Foundry. [

] Keith: He represents the uk and this month we are joined by Lucas Mi and he is a senior writer at Computer World. He is in an undisclosed location. Welcome everyone. Hey, howdy. All right. Uh, so this month, uh, the topic is new, thinking about cloud computing.

Um, it's a, it's a pretty big year for cloud and, uh, basically, uh, costs and a lot of other things are affecting enterprise companies and basically how they're operating with, um, cloud. [

] Keith: So what we like to do when we start off this show, we do start off with a bunch of statistics. So just bear with me as I run through a lot of the stats about the current state of the market.

So, uh, this was the, uh, from the IDC Cloud Pulse 2023 survey, which. Latest data that they had available and they said, uh, close to half of cloud buyers spend more on the cloud than they expected in 2023 with 59% anticipating similar overruns in 2024. [

] Keith: Uh, obviously the costs are, are basically just starting to go outta control. And we got another report. The Cloud zero, uh, cloud zero State of the Cloud cost said 58% of companies feel that their cloud costs are too high.

With 14% describing them as way too high, uh, which was an 11% increase from 2022. And what's interesting is that 66% of them cannot accurately report unit costs. [

] Keith: And 42% can only provide estimates when attributing cloud spend to different aspects of their. Businesses, but it's not just costs that are, that are affected, um, the world.

The issue of data sovereignty, there was a Polaris report suggesting that the sovereign cloud market is projected to grow from 9.83 billion in 2025 to 102.7 billion by 2020, right by 2034. [

] Keith: And again, this is driven by increasing regulatory requirements and the need for data control across several industries. And in, uh, and then, and then even the A IPC space is growing immensely. Uh.

The estimates are that approximately 48 million AI capable PCs were expected to ship globally, which accounted for 18% of total shipments. [

] Keith: And by 2028, this is predicted to reach 205 million units. Uh, and this was according to, um, canals. And so we're seeing a lot of different, uh, market forces that are making it leaders, um, rethink the cloud. Um, whether it's security, privacy, uh, costs.

So much so that I think consultants are now saying to, uh, enterprise IT leaders, let's rethink this. [

] Keith: The whole all in on public cloud. So I wanted to ask, um, let's get Matt on and say, you know, like, what is the state of the, the market from, from the IT leaders that you're talking with? [

] Matt: Yeah, no thanks Keith. I think you, you know, in hitting all those different data points, you, you really addressed something, which is it's complex, right? [

] Matt: And a question I like to ask successful IT leaders and maybe the unsuccessful ones too, um, is, is what do you now know not to be true that you used to believe? Right? You often list elicit quite interesting responses.

And I was talking to the senior leader of a European based global, uh. [

] Matt: Asked him, asked him that question. He said he had always previously believed that fundamentally cloud would be more efficient, right? That in a globalized economy, pooling resources on infrastructure would always work out to be cheaper in the end, and he now no longer believes that, right?

In part, that's due to economics and business. [

] Matt: Cloud providers wanna make money. So once vendors have you, they wanna lock you in to a deal that's advantageous to them. Um, and if cloud was always more efficient, finops wouldn't exist. Right?

But to be fair to those vendors and providers, the other issue is exactly what you hit on. And, and I'm really glad Lucas is here because this is something he's pretty expert in, right? [

] Matt: It's the demands of ai. So of the modern workspace a. [

] Matt: Um, so they wanna like repatriate and, and go to more kind of local on-prem, um, or to your point a I pc. So it's kind of at the edge, although I think that's a use case that's for the future rather than for today.

Um, others, other, you know, CIOs, IT leaders you speak to, they're signing up to long-term contracts with vendor groups because they're seeing an offer of, you know, uh, um, the software vendor, the cloud vendor, the hardware provider. [

] Matt: Coming together and saying, we'll give you next gen technology in perpetuity. And that allow, that will allow you to develop your AI projects without having to worry about all the infrastructure. So they're going kind of more into that sort of cloud space.

So I, I think when you throw in the complexities of managing data to support an ai, any organization and all of the complexities of the modern world, to your point, Keith Geopolitics kind of unstable around, um. [

] Matt: Cloud repatriations are happening. It's not about going all in on cloud, but also some other folks are saying, well, actually our future does lie in that space. It's just more complex than it used to be. I think. [

] Keith: Yeah, it Do you get a sense of why the, the cloud costs have gone up? It can't just be that the cloud vendors are trying to make more money or, or anything like that? [

] Keith: There are, there are a lot of reasons why. The, the costs have gone up and, and maybe gone out of control. I think a lot of CFOs were, um, spitting their, their, their coffee when they would get the bill on some of these things, right. [

] Lucas: I, I think it's difficult to manage those costs. It's difficult to know where all the data storage is, well, where all the data is going. [

] Lucas: That's part of the problem. It's hard to monitor that. I think if there was like one word that comes to mind now, it's, uh, with, uh, CIOs, CTOs, CEOs, and uh, and other leaders, it's, they're looking to optimize. The cloud infrastructure.

So they're looking to balance between on-prem and hybrid public cloud. So you've got years of skyrocketing adoption. [

] Lucas: And now folks kind of laser focus on optimizing the spend, cutting waste, getting resources to the right size, um, leveraging, uh, things like spot pricing where you can buy the extra spare compute discount. [

] Lucas: AWS uh, Azure, um, Google Cloud, they have a lot of unused server capacity and they want used at any given time.

But the catch, of course, is that the cloud provider provider can reclaim that or interrupt the service and capacity at any time with little or no notice when they need to back. So it's, it's not entirely reliable. [

] Lucas: But I think the more proactive, uh, IT teams, they're also looking at hybrid multi-cloud optimization, create more efficient workloads across multiple clouds and on-premise systems. So basically get the best mix of performance, um, compliance and cost.

And, uh, automation and AI driven tools are also making this a lot easier. [

] Lucas: They're being adopted to optimize, uh, cloud operations from monitoring, to scaling to security, got services like. Um, uh, cloud Health by VMware. Um, I think it offers a multi-cloud cost management with AI driven recommendations, that's a big one.

You're gonna get a lot of AI driven recommendations because it now can look at your usage of cloud and recommend what, where you can make cuts, where can optimize that. [

] Lucas: AWS uh, offers something called cost. Um, that offers AI driven rightsizing suggestions for its own services, AWS's own services, Google Cloud recommender, uh, that is, uh, AI to suggest cost saving optimizations. And then NetApp has.

That uses AI to optimize workloads as well in uh, instances, like I was just talking about in reserve capacity. [ ] Keith: Right.

So, so, so Lucas, nobody's actually saying, you're not hearing for anybody saying, we're gonna take all of our stuff from the cloud and move it back to on-prem. Maybe with the exception of ai, maybe some AI workloads.

Um, because I, I think this optimization thing is, is dead on, like, that's. People, you know, people wanna know where they're spending on it, why it's going up, and then if they can optimize it to save money. [

] Keith: I don't think anybody is saying right now, oh, let's just bring everything back. But, but maybe except with the exception of, of some AI workloads, Matt. [

] Matt: Yeah, I, I, I completely agree with that characterization. I was really interested to hear, um, what, uh, Lucas was saying, like, like. Are people looking to save? [

] Matt: Are people looking to move from the cloud? The answer is yes, but it's all in the context of a bigger picture. And this is a time of great change, right?

Every organization is seeking to be more data driven and to apply generative and particularly AI inside to drive efficiency and outside to generate new products and services. [

] Matt: And underpinning all of this, we've talked about this so many times, Keith, right? Is the need to. [

] Matt: But also able to scale with the organization's AI strategy that may not be fully fleshed out. So the old days of chucking everything in the cloud and managing costs is becoming outmoded, I think. Right? Cost is a driver, definitely. But so is scalability.

So is the ability to accelerate and grow and capability is a big part of it. [ ] Matt: Right?

And, and as Lucas so beautifully articulated, there are so many different. Options and opportunities here that that is where the IT leaders, not just the CIOs, but definitely architects are having to think through like their whole infrastructure and the, the part of cloud in it.

[ ] Keith: Right.

I think one of the clues that, that we were starting to hear about, uh, finops and the trying to reduce the costs were the number of companies that were pitching me for the demo show. [ ] Keith: Yeah.

Of, uh, you know, you know, their first pitch was. Cloud spending is out of control. And, and, you know, uh, CFOs are freaking out about all this.

And, and a lot of the CIOs and IT leaders were like, we, we have no idea where this, this, these increased costs are coming, which is like visibility on your spend. [

] Keith: And, and that was a, a, a big growth pitch for, for demo companies. So there's a lot of companies out there. We had a few on the show. Definitely. So, Matt, you brought up, uh, genic ai.

Uh, you know, we, I, it feels like we can't get around an episode of this show without talking about genic ai. Is, is that something that's gonna, um, help the situation in the cloud? [

] Keith: Uh, you know, these, obviously, these AI agents, when they're performing, their workflows are gonna probably need to go out on the cloud. So you better have your house in order before you start. Doing too. Right? [ ] Matt: Definitely.

And, and like I say, I, I would fold this into, and I really wanna fold this, your organization. [

] Matt: Think about it from the vendor side, right? This is a huge times of change are times of opportunity and threats, right?

If, if you're a vendor, the opportunity is to build long-term partners with organizations as they develop their AI strategies, offering future approved AI capabilities, technology that's evolving the partnership. [

] Matt: Basically saying to IT leaders, we can. Kind of outsource your architecture and we can provide it for you. And we've definitely seen some vendors see success in this space, particularly when they group together.

Like I said earlier, um, the cloud provision with the front end software and the hardware itself. [ ] Matt: Right?

Um, and we've seen vendors group together offer these deals because they can almost guarantee long term partnerships in. [

] Matt: If you can predict costs and if you can manage costs over the future, that's a really valuable thing. The threat from a vendor perspective is, again, it's a time of churn, so we are seeing increased cloud repatriation.

We are seeing organizations taking the architecture much more in-house. Building their own future proof infrastructure from the various different elements that we've described. [

] Matt: Cloud on-prem, some kind of hybrid even at the edge. And that's the whole A IPC promise, which again, I don't think is fully fleshed out at this point, but is definitely gonna be part of the the solutions.

And so you're also seeing organizations just wanting to keep their options open. Like, like, like that can be a strength as well to say, okay, you know, we're using cloud, we we're happy with the way we're using it, but we're not committed, uh, for the long term. [

] Matt: And it is driven by that, um, AI because of course the hidden piece about AI is it requires enormous amounts of compute capability.

Um, and that's not always recognized on the business side of the house, which is where some of this kind of confusion around, around cloud happens. I. [

] Keith: So, so let's assume then that, that a company maybe can get a handle on the, uh, uh, the issue around costs and optimizations, but it, that's not the only reason that they might wanna re-look at their cloud spend. [ ] Keith: Correct.

Um, um, in addition to that, you also got, um, the data sovereignty issue. And, uh, security and privacy.

I think a lot of people that are, um, talking about moving back to the cloud or moving into a different situation than, than all in on public is because of those two issues.

Um, Lucas are, are you hearing more about the data sovereignty issue or are they, or is it more about. [

] Keith: Companies wanna keep some of their, their, their AI workloads may be private because they're worried about leakage or, or agents or things like that getting out there. [

] Lucas: I, I think it's both. It's a little bit of both, but right now, as far as regulatory requirements, it's, it's up in the air. You've got President Trump trying to nix state regulations that are already in place. [

] Lucas: At the same time, a lot of these corporations, uh, the cio, CTOs, CEOs, whom we talk with are multinational. And so they have to deal with things like the eus AI regulation. Yep. Um, g uh, um, um, um, I'm sorry. Just went out on my head.

Um. Uh, GD Oh, GDPR, right. And then, uh, I was talking about the state regulations, the California Consumer Privacy Act, which is, I, I don't think the President Trump is, is going to be, uh, successful in, uh, putting a ban on state regulations. [

] Lucas: It's not constitutionally capable, but So they're still gonna have to deal with that. Yeah. Sovety solves part of that problem keeping, but. Multinational. So they still have to deal with that. How do you separate the two? We over. [

] Lucas: And one of the things he did was he looks at how they, he was looking at how they're using public cloud and they were basically using a public cloud for everybody within the organization. And he said, well, we don't really need to do that.

So what they started to do was they, they were using, um, um. Cloud for researchers who weren't part of their internal organization, which they use a lot of third party researchers, right? [

] Lucas: And then they're using on-prem supercomputers for the researchers who needed the data close to the supercomputer for the, for the faster results. And that cut their costs. Basically, they managing the transfer between clouds.

So we just looked at the data, where it was, who was using it and how it was being transferred. [

] Lucas: And that's one way that they can cut the costs. Um, and I, I'm not sure if this was part of the conversation you wanted to have, but cloud providers are also embedding generative AI and agent to AI into their platforms.

And that's gonna make it easier and faster and more secure to adopt. Uh, it also keeps cloud providers more relevant, uh, as this competition continues to increase with AI being a priority among IT and business leaders. [ ] Keith: Yeah.

[

] Lucas: So you got stuff like Azure Open AI, secure, uh, service, which integrates ai, uh, o Open AI's, LLMs, uh, large language models like, uh, GPT and then Codex and Dolly, you know, that's right to Microsoft Azure's, uh, Azure Azure's ecosystem.

You've got a W S's code whispers, that's a big one right now is jive coating. [

] Lucas: I love that name. [

] Keith: I thought it was Vibe. Vibe co. You mean vibe coating, right? Not Jive. [

] Lucas: I, I call it jive coating, but vibe coating. Yeah. It's sort like it's, that's okay. Night fever thing, you know. There [ ] Keith: you go.

There you go. [ ] Lucas: But the [

] Keith: Great, you, you're the first guest that has mentioned Saturday Night Fever? I think so. [

] Keith: You know, you get bonus points for that one, [ ] Lucas: but.

And, uh, folks are predicting that within five years, AI is gonna be producing 90% of code. And that doesn't mean it's gonna get rid of developers or peers. What it's gonna do is gonna offer them more ability to be creative.

So it's gonna take care of the base code, but the other thing it's gonna do is gonna allow business people. [

] Lucas: To get involved in coding. Yeah. Use natural language processing to do this. That's all part of cloud services. So stuff like, uh, AWS's Code Whisperer, that's an AI powered code generation tool to help developers and business leaders.

So if they get an idea for an application, they don't necessarily have to go to the developer team. [

] Lucas: To get done, they're gonna learn how to do it themselves. [

] Keith: You brought up, you brought up a lot of different issues, Lucas, around that. Like, again, not just the data sovereignty, because we've had, uh, guests on the show before, um, especially from India and some of the Southeast Asia, uh, areas.

They, they are really big on the, uh, in the Middle East as well. [

] Keith: That was, uh, a data sovereignty issue. But you know, when you bring up the coding part of it too, there's a lot of companies that don't want their code out in the cloud. And that brings up the security issue.

So maybe instead, uh, we, we, we keep all of our coding, you know, being done and all of the AI coding being done on premises, or even on an A IPC, I know we've done a bunch of demo episodes with people that were showing off the, uh, the a IPCs and the chips.

[

] Keith: And that was the, one of their big selling points was like, there's a lot of companies out there that don't want a lot of that stuff leaked. So, um, you know. Excellent points on, on, on all of those.

So what, what are you hearing from the security side of things too? So I, I [

] Matt: would like just rolling back a little bit on that because I think this is, this is really important issue and Luc has hit on this when he talked about organizations, the, the hospital you spoke to like. [

] Matt: Looking at use of data and, and Keith, in this instance, we can include, um, development and coding in that, right? And, and creating a, a British idiom, but for courses for each, um, for each use case. Like what do we want? Be OnPrem, uh, cloud.

And within the cloud, right? We can have a cloud that is, that is, uh, sovereign to one particular nation or one particular task. [

] Matt: Um, industry specific clouds, region specific clouds like that can be critical if we're building infrastructure in that way. So it's about architecting the right solutions for the right tasks, even within an organization. Right.

We're talking actually then, um, kind of a level of discipline around data and. [

] Matt: Which the 娇色导航of a large enterprise was talking about how, in their view, confidential computing had to become table stakes for their organization to remain in the cloud. Right.

Again, same sort of thing like the use cases they were talking about had to be completely sandboxed, um, in order for it to be viable for them. [

] Matt: So for them, data security and their customers. And the trust that comes with it was just the start.

Um, because there is, as we've covered, there's regulatory compliance, but that organization wants to be able to work in collaboration with suppliers and partnerships on AI projects and development projects. [

] Matt: But it needs to be able to process data. And it needs to be able to do creative activities, IP related activities like development in a completely encrypted and anonymous way, whilst also being collaborative, right?

So that can mean on-prem, but it can also mean to Lucas's point, the vendors getting really smart about the cloud services they offer in a way that they are kind of. [

] Matt: Sandbox, um, in, in a way that you can be really confident that you're utilizing cloud capability and technology, but doing it in a way that is completely private to you and your organiz. [

] Keith: Yeah, that, that, that's a great point. I keep always in, in my head when I start thinking about cloud, I'm always assuming public cloud, um, versus on-prem, which is then private, but you could do private clouds, so I always keep forgetting about the private cloud option.

[ ] Keith: So.

You know, there's companies that could still have a cloud infrastructure, but it's all completely private. [ ] Matt: Right?

And the vendors are savvy to this, right? Lucas was talking about it, so I shouldn't talk over, but the vendors are getting savvy to this because that in, in many cases, that is a key selling point for an organization to be able to use the cloud. [

] Lucas: So, uh, right sizing, we, we wanna talk about this as well.

Um, we talked to a CTO at the Reinsurance Group of America, RGA, and he's working on right sizing his company's public cloud footprint by looking at processes around work, workload intake and distribution criteria and implementation practices across the private and public cloud. [

] Lucas: And what he's saying is expense optimization, defining workload selection criteria is gonna determine which. Public cloud, public cloud in which things go to the private cloud.

As, as you know, vendors charge not just for, uh, ingest but also uh, egress fees for data leaving, uh, uh, premises as well. So they're trying to manage their daily, their daily, um, life cycle of the data and minimize transfer between clouds. [

] Lucas: 'cause you're charged for both. Um, so that's something to look at. And between availability zones, regions, and cloud. You're charged. Well, when you leave a uh, avail availability zone, you're charged extra for that as well.

So again, they're looking at those things that Matt was just talking about, um, in terms of sovereignty and, uh, and, and, and transfers between different zones and how much more it's gonna cost. [

] Lucas: These are all playing into how, how they determine whether to use public and private cloud as, as far as PCs, uh, ai, PCs, those are a very real thing.

There are a lot of, a lot of data out there that's showing it's adoption rates right now, well put it this way, so shipments alone this year are gonna be about 40% of all PCs. [

] Keith: That's ama, that's a, that's an amazing number. Like I, yeah. Do you even, does, does, does that feel right to you or is it just because I have a Mac and I never really follow the PC marketplace? [

] Lucas: Oh, I think it is because I think the, the, uh, the companies are getting ahead of the curve, especially with their, yeah. [

] Lucas: With, uh, the chips, the AI chips that are coming out now, that's gonna dominate the industries by 2028, IPCs could come for 80% of the PC market. We're talking 600 million computers over the next four years.

And that is part of this is because of their concern about privacy. [

] Keith: I mean, it's also the natural refresh cycle for a lot of computers too. [

] Keith: And plus it like with Microsoft too. Yeah. They're sunset setting Windows 10. Right. So it, I, I think they're gonna, I think most companies, you know, probably within five years you're gonna have an IPC at some point. [

] Matt: I think so I'm gonna throw an element of skeptic because I'm not, I'm not disputing anything Lucas said, but I, but I think one of the drivers behind it is exactly what you're just talking about, Keith. [

] Matt: And we're talking about shipments here, which is like the hardware vendors are definitely driving this cycle. Of course they are, because they've got differentiation opportunity. They're able to say the hardware you've got, you now. It's out of date because it's not an AI pc.

The technology is growing amazingly, and that's tied to the, the Windows 10 retirement piece is like the, the intention here is I used to edit a, uh, a PC magazine many, many years ago, and we were so happy whenever new windows came out, right? [

] Matt: Because we knew we would desktop PCs to people and, and it is that kind of a, kind of a vibe that from a marketing perspective, I [ ] Matt: ISAs.

Those IPCs at the edge are gonna do a lot of, a lot of the heavy lifting around this compute. They're gonna bring great scalability to the ability to do things. Privacy, like clearly beyond compare from the cloud perspective. I just personally, from talking to analysts and.

Um, IT leaders. I'm not hearing of many actual use cases where it's happening now. [

] Matt: In real May, maybe I'm speaking to the wrong people. Right. Um, because it, because it definitely, that modular scalability, that increased security definitely is a use case. It's really exciting to have it.

Um, I still believe, to be honest with you, that yes, I think as um, organizations refresh their PC stock, they're gonna GOCs. [

] Matt: It's gonna be the right thing to do. In part, it's gonna be driven by the fact that that's the only option they've got. Right. Because, because they're not being sold anything else.

Um, but I, and I also still believe that cloud will be part of the solution. Right? They'll be, they'll, yeah. Absolutely.

At the edge, if you can put in your hands of your end users, the capability to do AI immediately right there, but they're still giving a connective tissue in the infrastructure and that's still gimme a mixture of cloud and on-prem. [

] Matt: That's just my understanding. You [

] Keith: mentioned in the earlier part of the show, Matt, that you were a little bit skeptical about the, the growth of the A IPC market.

And again, you, you, you hit the nail on the head with the, the, this is being pushed by the hardware vendors, not just the, the computer makers, but also the chip makers too. [ ] Keith: Um.

Again, we've, we've done some demo episodes where we've gotten some demonstrations of, of, of the, a IPCs from, um, I think we did Lenovo, uh, Qualcomm, Dell there. There's a whole bunch of people out there that, that did some demos with us.

Um, what are, what are the reasons that you think that, that, why are you being so skeptical? [

] Keith: And if you don't have an answer, I do have one on my own. If you wanna just [

] Matt: purely speaking to it, decision potential. Actually, I was speaking to an analyst recently in an event who, whose job is to coach and work with it and was saying. [

] Matt: Just like right now, very few IT leaders are seeing immediate kind of return on investment on IPCs. But to be fair, Lucas's point was it's about future proofing, right? And so that is definitely a viable thing.

'cause I think if you are expecting AI to be baked into everything you do, it definitely makes sense for end users to have the devices. [

] Matt: I just, I suppose my skepticism such as it is, is based around. It is being driven by the hardware manufacturers, obviously. Right? Because, because they really want people to riff.

Like, like with, um, we did another demo, uh, episode recently about, you know, using uh, uh, a cloud-based operating system to, um, to give another 10 years to, to hardware, right? [

] Matt: It is, with modern hardware is completely conceivable. You can keep it running for if doing basic computational tasks. I think most organizations are gonna be doing more than just basic computational tasks, but that is the kind of balance that it decision makers have to make.

So, yeah, I'm, I'm not skeptical on the concept ofs. [

] Matt: I think it's super exciting. I'm just. Right here, right now. But that's based on speaking to it. Decision makers. I'm interested, Keith, right? Your skepticism. Where's your skepticism? [

] Keith: Well, I'm gonna throw, I'm gonna throw it to Lucas to see if he wants to, to, to slam down the basketball right now, or, or if you want to pass it back to me [

] Lucas: talking about hardware vendors, but let's not forget that. [

] Lucas: Microsoft PC concept, I mean 365, you know, all [

] Lucas: computering units function these features, so. They're pushing it pretty hard too, 'cause they see the future, uh, of AI as well. They were all on board. Remember? They, they were doing like 49% of, uh, open ai.

Uh, so they dove in right from the beginning, even, even before 2022. When, uh, was 23 that, uh, Chatt, PT came on board. [ ] Lucas: Late 2022.

Yeah. 20. Yeah, late 20 October. Yeah, I think it was November. They, they've been in on this from the, from the get go, and they see this as the future.

So the software vendors are pushing this hard too, because it's all gonna be part of the software and they're gonna tell you, you're not gonna be able to access these features, or they're not gonna function, uh, efficiently without an A IPC. [

] Lucas: The other thing you're gonna see is a lot of IOT devices. Running these AI chips for the same automation. A lot of things are gonna be automated and we could talk about future of work there as well, if you like to.

But yeah, so a lot of, a lot of the devices for security reasons are going to, uh, use AI on-prem. [

] Lucas: But at the same time, if your staff, if your workforce is not educated on how to use these devices. You're right, you're not gonna see a return on investment and right now folks are still trying to figure out exactly where the IR ROI is with ai.

That's so hard to say. [ ] Lucas: Alright, [

] Keith: Lucas.

Lucas, you're really close, but you didn't have the answer that I had. There's. Oh, you wanna hear what you wanna hear? What my, my thought is in my gut, there's no apps yet. They're the, the software is not there yet.

Um, I know Microsoft has been doing a lot, they spent a lot of money. They, they're throwing everything at the wall. [

] Keith: They're throwing the pot at the wall, see what sticks. Um, the demonstrations that I've seen, they look really good, but they're not completely, um, they're not there yet. Once people start seeing that, that there might be a killer app usage for doing, doing an A IPC.

Rather than just going to the cloud and using all of the, the, the AI services, this is just from the end user side too that I'm seeing. [

] Keith: Um, un until you start seeing those apps with doing this on the PC without needing a, a network connection, then I think then I think they'll start to take off. I.

And then, and then one thing that I didn't mention was the battery life on the a I PCs.

Maybe if you've got end users that really just don't wanna carry around that charger anymore, or you know, the power pack or, you know, try to fight at the airport for that last plug in the seating area. [

] Keith: Um, you know. Once you see a 12 hour battery life, that actually is a 12 hour battery life.

Um, 'cause again, someone told me at what during one of the, one of my travels was like, yeah, battery life is great until then Microsoft gets involved and then they stack all of their apps on it and then all of a sudden you go from 12 hours to like three.

[

] Keith: So again, maybe that's just my, my my, my Microsoft, um, bias showing. [ ] Lucas: So [

] Lucas: for a second, let's just face it, most workers are using chat. I mean, everybody's playing with chatbots right now, and that's, you're not seeing a real, I just wrote an article about this. A study showed that they're not finding an ROI.

They're not even seeing a lot of productivity gains. There's some efficien efficiencies happening, but they're not using the additional time that they're getting from this to do additional work. [

] Keith: I think you and I need, need to do this on another today and tech show because I, I wanna, I have a story about, about how I tried to, uh, to actually get a work process done through AI and chatt.

Bt and I spent more time trying to get the, the AI to do what I wanted it to than what my original goal should have been. [

] Matt: That's a, that's an issue for another show that probably, and that's definitely the. [

] Matt: Research with developers that shows it, it, it has proven possible to this point to show like an increase in productivity from using like coding, uh, AI driven coding. And yet you speak to individual developers, they're all like, oh, it's brilliant. It makes me do more. Right?

Because it is taking away toil. [

] Matt: It is removing in the same, there's a brilliant article on Info World recently about the rise and fall of Stuck Overflow. And, and I remember my roommate at college, he was a, he was a developer. He's a developer.

Um, and it, this, this was in the dim and distant pre, uh, pre-internet age. Like when he was learning coding, he was still copying and pasting from other people, right? [

] Matt: 'cause that's how it, it worked. And AI. [

] Matt: From scratch and he will tell you. Yeah, yeah, absolutely. Like chat, GPT, whatever, all these tools, they're really helping me be more efficient. But, but are we seeing the return on investment on the, the outside of it? I don't know. There's less toil.

Maybe there's more creativity. I don't know, but we'll definitely get there for sure. [

] Matt: Those killer use cases will come, but, but I, in my work in life, and I'm using these things all the time, I don't see those killer use cases at this point, but I'm sure they will come. [ ] Keith: All right.

I, I got one more, I got one more question to ask or one more topic around the world of cloud, just to get us back to cloud for a second. [

] Keith: Um, a lot of the, the discussions with cloud computing was around the, the idea of being a green cloud and sustainability and all of that kind of stuff. Has that just gone away?

Um, basically because AI came around and, and the, the power requirements for ai, um. It's, it's very, very hard to be a green data center or, or a Green Cloud provider. [

] Keith: Um, if you've got all of these other requests, is that one of those things where money took over, um, and, and, and, or are you still seeing a lot of these, these efforts? Or is it, is that a hot potato?

Should we not even bring up the sustainability part of it? I go back to Iron Man and Tony Stark. [

] Lucas: No, no, no. Please explain. Well, I. [

] Lucas: So I, I can't remember the, the protagonist, uh, the oh and the antagonist in that, of what his name was. But the, uh, they, they go into the, uh, uh, the center where they have a, uh, a fusion, uh, uh, a, a fusion generator.

And he looks at it and, and he, and he points at and he goes, we just made that to please the environmentalist, you know, that. [

] Lucas: It was never cost effective. And it kind of feels like that a little bit. Like we just did that to please the people who, you know, who, who thought AI was the next, next big thing.

I don't think it's that exactly, but right now they're still rolling out in, in small measure. So is everybody rolling this out? [ ] Lucas: Yes.

In some form there's a lot of fear of missing on out there, but is it, is it really giving, are they, are they rolling it out widely within the organization? No. So they're, they have small, every.

You know, like 80% of companies more, they're rolling it out, but not 80% within the company, if that makes sense to you. [

] Lucas: So it's, there's small instances of it. They're testing of it and they're all different types of, of, uh, of, of, of, um, chatbots and, uh, agents that they're rolling out. But again, small instances, they're not seeing the ROI yet, they're still testing it.

They don't this, we're gonna efficiencies. It's gonna happen, especially in the cloud because AI has those monitoring capabilities. [

] Lucas: It has the capabilities to look at data well, once we get past the errors and the hallucinations. So Matt, [

] Keith: this may be our US bias showing through where, where Lucas and I roll our eyes when it comes to a lot of the sustainability you're in the uk, are, are they take, are they trying to take, you know, green Cloud computing more seriously? [

] Matt: I don't think this. [

] Matt: I, but I do think there's potentially even a generational difference. And I think, so, I think sustainability has always been the hidden cost of computing and connected computing, right?

Like, like think about receiving an email from someone that says at the bottom, please don't print this email. Right? As if, think as if posting that information in the cloud takes no energy, right? [

] Matt: Like its madness. But that's how, how we kind of. [

] Matt: Is most top of mind when IT leaders talk about driving efficiency and talk about efficiency through the cloud.

And that's also at the root of a lot of talk that and security or at the root of a lot of talk about moving to the edge and using AI infused end user devices. But I do expect sustainability have become increasingly important factor in cloud and an infrastructure.

[ ] Matt: In general.

Partly that's to do with cost savings. Partly. I think geopolitical instability is like a big factor in that, right?

We've all got used to, in our lifetime, just being completely comfortable with where our energy's coming from and that it's sustainable, that that may not be the case like forever. [ ] Matt: Right.

I think the cost of energy, I mean a big driver in the uk you asked about the UK like, like.

Like between the, uh, Ukraine conflicts, um, the UK leaving the eu like energy prices at household level in this country have spiraled over the past few years and that really changes people's behavior, right? [

] Matt: Like, like when, when you are worried about paying your next month's gas and electricity bill, like people start to act in amongst more sustain with ai.

If it does, I mean, Lucas is right this statement with the situation today, but like if it does drive extreme need for more computational power, I think yes, you'll see people looking for more sustainability. [

] Matt: I also think there's a whole other conversation around compute itself.

Quantum's part of this, like, you know, my, my 10-year-old came home from school telling her that her teacher had told her that the future was quantum computing, and that was gonna be great because you wouldn't need the same battery capacity and you wouldn't see, need the same power to drive a computer.

[

] Matt: I don't know if that is the case, right? But that's definitely the vision for like, like future, um, future state computing. So, yeah, I, I think overall, I think sustainability is something that, like, what really care about is cost efficiency.

I do expect sustainability to become a much bigger issue in the, in the near to midterm. [

] Lucas: You got one more? [

] Lucas: They're still gonna be used out there, but we're going to smaller language models. Absolutely. Excellent point. Fewer parameters, less compute necessary, more efficient, more targeted at per yeah. Indu [ ] Keith: industry.

Specific LLMs will be, will, will be a lot less expensive to run. Right. Yeah. Yeah. Excellent point. [

] Keith: All right, so I wanna get to the end of show vote. Uh, this is where we usually do a yes no answer, but you know, everyone usually creeps in. Uh uh, depends.

Um, alright, so out of all of these different goals and trends, which will be the biggest mover for companies during the rest of 2025 and going into 2026, so biggest issue around cloud, will it be cloud spending, uh, data sovereignty, security, sustainability, or edge and IPCs. [

] Keith: I'm gonna go first 'cause it's the easiest answer. I think it's, it's cloud spending. I think we're at that point where, um, everyone needs to get their house in order, so I'm gonna go with cloud spend. If you could, you're allowed to pick. Whatever you want.

I [

] Lucas: mean, spending is one. I was, I had it as AI adoption with an enormous increase in attack surfaces, especially with IPCs edge devices becoming widespread, companies are facing new types of cyber threats linked to AI powered systems and cloud integrations. [

] Lucas: You've got new regulations related to ai, even if President Trump is trying to ban them, you still have like said earlier, state res, other. EU AI Act, GDPR. Um, those regulations, they're just unacceptable risks. Uh, such as, uh, real-time remote facial recognition.

We're already seeing that social scoring systems, we're already seeing that systems talking with vulnerable groups and the cost and risk associated with data breaches through, uh, AI model exploitation. [

] Lucas: So I could go on and on, but yeah, I think it's, I think it's AI combined with cost. Alright. And Matt. [

] Keith: Well, you, you need to go over the list again. [

] Matt: You know, my answer is always all of the above, but actually, um, the way the, the, the way Lucas framed [

] Keith: it, one of, one of these days, I'm gonna get you to commit to a yes or your answer. [ ] Keith: I'm, I'm [

] Matt: gonna commit, but I'm gonna commit to an all of the above kind of answer. Right. Which is okay.

I, don't disagree with Lucas, but I think to me it's about the future state of organization and like when they're thinking about cloud, it's not in isolation infrastructure, I mean, storage, compute. [

] Matt: And the security around that, right? And it's about what is our organization gonna be in future? And that's what AI is driving.

So, um, I think, I think it is about AI adoption to use Lucas's term, but I think from a specifically cloud perspective, it is what is our organization gonna look like in say, five years time? [

] Matt: What do we need our infrastructure to be? And that is driving real change, like more than we've seen for quite a long period of time. [

] Keith: Alright, thank you guys. I, you know, I know that we only had three guests on this show, this, this, this week, but we're still at that 41, 42 minute mark. So again, thanks, uh, thanks for joining us on the show and again, great, great topic. [

] Keith: So, uh, we're gonna be back next month with more global editors to talk about how AI is changing the data center, which is basically the, uh, you know, turn it around from cloud to data center. So that's all the time we have for the shows today.

Be feel free to add any comments you have below. Be sure to check out our other Tech talk shows such as today in Tech. [

] Keith: 娇色导航leadership, live and demo. If you are interested in seeing B2B product, demonstrate demonstrations. Uh, I'm Keith Shaw again. Matt, thanks again Lucas. Thank you. And, uh, we'll see you next time. [ ] Matt: Thank you.