AI’s Soaring Power Needs Push Tech Giants to Build Private Energy Networks | Tech Field Day News Rundown: February 25, 2026

AI generated content is bad. Data centers are bad. Microsoft says MCP is good.

Las owing an AI agent, AWS says, AI coding isn't bad. Beam me up T-Mobile, and we're gonna look into private power networks in this week's episode of the Tech Field Day Rundown. Hey everyone, welcome to the Tech Field Day rundown.

It is February the 25th, and I hope that you're all staying warm with a nice big bowl of clam chowder. Yeah, that's what day it is. Clam chowder, who knew it needed its own day, but you need your own day to be able to keep up with the news.

And you know what day that is? It's Wednesday because that's the day that we come out with the rundown. And that is the day that I am joined by my co-host, Mr.

Alistair Cook. Alex, good to see you again. Always a pleasure to be here with you, Tom.

And, uh, as you know, I like my nuts covered in chocolate. And so today is National Chocolate Covered Nut Day as well. Yeah.

You know what? We are all about the food here on the rundown because it is the thing that we subsist on that allows us to bring you all of the great news that we are uncovering digging through, uh, by hand. Because, uh, you know, we, we wanna bring you the most curated, artisanal, handcrafted news that we can find.

And we're gonna start off with a recent story because on Monday this week, the 23rd of February, 2026, privacy regulators worldwide working through the Global Privacy Assembly, issued a joint warning about AI tools that create realistic images and videos of real people without asking for consent, while acknowledging AI's many benefits, they raised concern about harms like non-consensual intimate images, defamation and risks to children. The statement urges organizations to put strong safeguards in place, be transparent, remove harmful content, and work closely with regulators so that innovation does not undermine privacy, dignity, and basic rights. This sounds like a very wa audible goal.

Al, do you think anybody's gonna listen? You know, I doubt it. And that's disappointing because this absolutely is a, a very laudable goal because as we know, uh, deep fakes bad, lots of harm has come already to people from fake content being created.

And we're seeing increasing the use of fake AI generated content to push a narrative that is not necessarily truthful. And so there is absolutely a problem here, but I'm not sure that a joint statement from a variety of underfunded privacy, um, regime or privacy agencies within a variety of small and large companies around the world is gonna make a big difference. Uh, absolutely having some ideas that, uh, there may be legal issues with deep fakes around the world.

And we know lots of legal frameworks are really struggling with keeping up with the capabilities of ai. The frameworks that we have are based around humans taking action at the speed, speed of humans, not AI generating terrible fake content at the speed of ai. Uh, so there's some real challenges to be seen, and this is just a, a sort of beginning framework ideas of you need to have strong safeguards, you need to have meaningful transparency, uh, you need to have accessible mechanisms for, uh, takedowns of harmful content.

And well, once somebody think of the children, we must, uh, also be addressing some of the specific risks towards children. Age appropriate content being delivered for children, and also appropriate content being generated about children. Uh, I like that there is a global group of privacy organizations.

I particularly like that New Zealand's Privacy Commissioner is one of the signatories to this piece of work. Uh, but so far it's just saying, yes, there are bad things out there, and you should take steps to not do bad things rather than anything that'll make a difference. Because of course, the almighty dollar is what's driving people to do these, uh, deep fakes, create this content.

Uh, or also that maybe there's also bad people out there in the world and bad people won't follow good guidance, will they? Uh, yeah, this just feels like a lot of hand wringing and not a lot of action to me. I think, uh, it will be important as we move forward in life to consider that not everything that you read, not everything that you see is truthful.

Of course, here at the rundown, we bring you some handcrafted artisanal truth, truth, not some AI generated truth. As the 2028 presidential race takes shape, uh, Democrats are dialing back their strong support for AI after voter backlash over rising energy bills and job losses. Governors like JB Pritzker, uh, Josh Shapiro and WHI more are stepping away from generous incentives for power hungry AI data centers to bring money into their constituencies 'cause their public are concerned.

Strategists say that these projects, uh, now symbolize wider fears about ai, um, pushing candidates towards tougher rules, local approval and fair taxes with skepticism spreading across both parties, uh, and calls for worker protection, increasing the area of automatic political support for big tax ai. Uh, the expansions of maybe the support is coming to an end, Tom, is this really gonna put a crimp in the growth of AI data centers? It could provided enough of these governors get together and say, not in my backyard.

Because the companies that are looking to expand these data centers and, uh, you know, build out this infrastructure, they need access to specific kinds of things, right? They need high tech, uh, infrastructure, they need access to water. Uh, they need a clean power grid, which we'll talk about a little bit later.

And a lot of those are in places that right now are being governed by people from the Democratic party, uh, people on the left of, uh, the American political spectrum. Uh, that's not to say that they can't build some of these things in places that are on the right of the political spectrum, but one of the biggest things that's coming out of this is that people are starting to wake up and realize maybe it's not a good thing to build a giant data center in the middle of my, uh, constituency because the power is gonna go up. We know that, right?

The amount of power consumption is gonna increase significantly, and that means that everybody's electricity bills are gonna be going up. It also is, quite frankly, an eyesore. I don't want a giant warehouse just sitting in my backyard.

Um, we've seen this a lot with Amazon fulfillment centers. We've seen this with a number of other things that get built, and they don't really bring a lot to the table. Uh, now before anybody leaves a comment and says, well, it's, it's gonna bring a lot of jobs, how many?

And more importantly, what kind of jobs, because I think it's gonna be a lot of people who are working on the server farms. That's not something that your average, uh, hourly employee can really jump in and do. Which means either these companies are gonna have to import their people from other states, which is typically something that does happen, or they're gonna have to spend a lot of time spooling that up, which means there's going to be fewer jobs available for the amount of resources that are being consumed.

And that is a problem for people. That would be a problem for people, except for the fact that you're also supporting ai, which is purporting to take people's jobs away. So I think people are not only rejecting the data center itself, they're rejecting what's gonna be housed in the data center.

You know, this isn't a data center like Apple's building out to house, more iCloud storage. This is big, big stuff. So I think what we're gonna see is that towards the end of 2026 that there's gonna be a little bit of a pushback.

You know, maybe we're not just gonna give you all the, the energy and resources that you want, and then that will kind of determine what happens in 2028, whether this build out's gonna happen, whether the ai, um, balloon will deflate a little bit, or whether or not Lex Luther and Brainiac and the rest of their crew are finally gonna be able to launch data centers into space. Yeah, don't hold your breath on that one either. Engineers over at our friend Microsoft are building MCP servers and saying that it's very different than building an API gateway, sharing some of the lessons from their work on MCP servers for Visual Studio.

They explain that agents perform better when tools are simple, opinionated and designed around real, real workflows, not overloaded with API options by reducing context, switching limiting tool choices, and using built-in feedback to guide agents, Microsoft found that MCP servers can help developers complete tasks faster and more accurately. Their takeaway designed for how agents think and act and not how internal APIs are structured. Al this is just earth shattering news that sometimes you have to build things for the way that people think and not give them every option in a Coke freestyle machine, right?

Yeah, it's a, um, interesting perspective. I mean, I'm looking at it as an architectural precept. You know, the idea of having an API gateway historically has been we'll bring together all of these disparate sources of interaction with our applications, will bring them together into this gateway, and we'll provide access to everything, every possible way you could interact with our IT systems will come into this API gateway.

You can use it as a library of things that your application can talk to. Now, human developers are really good at ignoring the fact that they are provided access to 15 things they don't need to use and to retaining that. Whereas AI agents, and this is one of the things I took outta the, the Microsoft blog post on this.

AI agents will want to go and look at all 400 things they have access to and work out which one to use. And we'll typically do that really frequently. And so they get very overloaded with options.

They're kind of like a 2-year-old when you say, here are the 15 things that you could have for your snack. They get a bit lost and can't cope. When you zero it down and say, in fact, here are the two snacks you are allowed to have today.

Which one would you like? They can make much better decisions. And in this way, AI agents do seem to be like two year olds.

You need to direct them a lot. Reading a little bit along the way in here, uh, the Microsoft team built a specific way for AI agents to talk to their documentation. So rather than having, um, read it any way you'd like, what Microsoft did here for this MCP server was build an interface that allowed basically two options of getting to things, go search or retrieve a specific item.

And they very strictly controlled what the AI agents could do in order to access this information. So that's, that's a good thing. It's not that nice simplification.

But then we started reading a little more through this, and there's elements in here of their lessons of, uh, expect things to not work the way you you thought they would, and that you have to change things dynamically over time. And this is not some fixed piece. It's not gonna work all the same way, because if you change a little bit of the description in your MCP servers interface, the AI agents are going to come to a different conclusion about how it works, huh?

So I've now gotta write my descriptions to the agent that's gonna access this, but different agents are gonna, so now I need to have different MCP servers for each of my agents and hang on. And we just lost the whole point of having MCP when it gets this complicated. I think the promise of CCP is not quite as great as we thought.

And then there's, there's a, this whole idea that iterations through this that you can have problems with hard hardcoded, uh, configurations and, and agents, agents will remember the way things work a couple of days ago sometimes, and will continue to work that way. Mm, yeah. This does seem like a, uh, MCP is not the silver bullet to all of our problems with working with AI agents that maybe we've been led to believe still feels real early, much more to work to be done here to turn this into an easily usable product.

The so security is introduced to new capability one designed to track and analyze agent behavior and helping organizations apply stronger guardrails as a agent AI use accelerates. We've seen the need for more guardrails around agentic ai, particularly, uh, all of those claws that you might be thinking of, the open ones. The feature called Intent Deputy establishes behavioral guidelines that allow security teams to detect misconfiguration and drift and malicious activity in real time, according to do offer draw the tool analyzes, uh, full session histories to ensure AI agents act only within their authorized purpose while adding minimal performance overhead as AI agents, um, inherit broad permissions and are deployed faster than security controls can keep up.

But so argues that continuous behavioral monitoring will become essential to limit damage, meet compliance requirements, uh, and prepare for the increased regulatory and audit security as we hear of more and more security problems with agent AI being given excessive permissions. Uh, Tom, does this give you hope that we'll be able to put a lasso around our agents and keep them under control? Lasso deputy, hang on, partner.

If you wanna keep an eye on all those potential bad hombres, you need to know exactly what they're doing at all times and make sure it's to squared with what they used to be doing. Otherwise, you're going to have to lasso 'em. You're gonna have to throw 'em over there and not their jail and scene.

I, I kid a little bit, but like, this to me is kind of what we should be doing with AI agents, right? Is we should be analyzing what they're doing, setting baselines, and then looking for what happens when they drift. Normally, when we do this in security, we are looking for indicators of compromise, right?

I have taken over a process and I'm gonna use it to send information to um, uh, G-P-T-A-P-T-Y-P-T uh, system out there. Uh, but in this case, I'm looking for the agents to drift away from what their guidelines are, right? Because this ins is the problem that we've run into with some of the latest revisions of agents.

There was a story that just broke this week about how someone at Meta gave one of those bots access to her entire email, and it decided to clean things up by deleting everything, even after she specifically told it to stop multiple times. It's like, oh, yeah, I probably shouldn't have done that. Sorry, my bad.

Uh, with, with lasso security in their, their deputy system, basically what they're doing is they're saying, here, your guardrails, act within your guardrails, and we're gonna keep a close eye on you to make sure you don't drift away from it. Because what that does is it allows you to know almost instantaneously when there is a problem, and it allows you to make sure that you're always in compliance and you're probably thinking to yourself, well, that's great because, you know, if I'm gonna deploy these AI agents, I don't want them to do anything crazy. 'cause I don't wanna have the increased workload of having to clean up after them.

No, that's not what you need. You need to be able to show auditors and regulators that you didn't let the AI agent loose and do things that it really wasn't supposed to. Especially if you work in an organization that keeps things like medical records or financial transaction records or, you know, any kind of PII that would possibly be leaked out there to the internet and that would be bad, right?

Uh, 'cause uh, in the words of Steven FoST, you know, if you make a smoothie with strawberries, you can't reconstitute that by giving the strawberries back out of the smoothie. Uh, that wasn't related to an AI thing, it was something a a little bit older. But that's really what happens with these models.

Once you've put all those units in there, you can't extract them. And so you really don't want that data to leak in there, and you don't want those agents feeding things they're not supposed to, or, you know, acting in ways they weren't tasked to do. So I hope that lassos, uh, solution is one that gets adopted, and at the very least, I hope it gives people, uh, a moment to pause and think about what they really should be doing with their data as opposed to just letting those agents run wild like a gang of hooligans.

Our friends over at Amazon, you know, they, they've had their fair share of some issues, but they're disputing reports that internal AI coding tools were responsible for some of their recent service disruptions at AWS following coverage by the Financial Times. The company said that a December outage stemmed from good old fashioned human error and not access control issues, um, or AI autonomy. Despite involvement of internal coding assistance like Kiro and Amazon Q developer, Amazon emphasized, bolded, underlined, and italicized that the incident affected only a limited cost management service in one region and did not dispute core AWS infrastructure, still events of intensified debate over AI governance permissions and guardrails in complex enterprise systems as companies accelerate the use of autonomous agents in production environments.

So Al is it a good thing that this wasn't caused by code, but in fact, people who didn't know what they were doing? Well, maybe, but here's the thing. Before Kero was let loose to do its thing, automatically this fault occurred that the excessive permissions existed and they weren't causing an outage with the existing permissions.

So there'd been a mistake, made excessive permissions, and then Kero comes along and exploits those excessive permissions. Who's to blame? Both of them, right?

Both Kiro taking these excessive permissions and making changes in a production environment. Now it's a fairly common kind of behavior in a cloud native application. When you need to deploy an update, you tear down the old one and you build a new one.

Well, not quite. And that's what you do when you're in non-production. When you're in production, you build a new one, you switch the workload across and then tear down the old one.

And as I'm reading between the lines of what's described here, it sounds like the non-production methodology was used in the production environment here by Kera, the fact that KERA was given enough permissions to be able to do that. Absolutely. That's a, an error of excessive permissions along the way.

But Kira really exposed a couple of things here, not being sufficiently aware that it was working in the production environment to take production methodologies and just, but it was a multiplying effect of having an existing fault. And we know system outages sell them, have a single fo cause, right? We build our systems to cope with a single failure, and it's when there's multiple failures or multiple errors being made that we end up with system outage, its care exposed a fault, but it also was at fault itself.

So I think saying that it wasn't the coding assistant, uh, it was the excessive permissions is splitting his here is here. The, if the, uh, excessive permissions had had not been excessive, you know, not been in place, Keira might not have been able to do what it did. Uh, but if Keira hadn't been running there, there wouldn't have been so much of a consequence of the excessive permissions given to Keira.

Uh, the other use case they talk about was Amazon Q, which is the plugin that I keep getting asked to update in my visual studio code. Uh, and that, that in that case, there was no user visible outage that, uh, apparently Amazon Queue did some things somewhere and made an application not work quite right internally, but not to the extent that it caused an outage. What's the root in this?

Well, one of the things is that AI moves faster than humans do and has no idea of the context of what it's doing if you don't provide that context. And so it's a little bit like handing the intern the keys to whatever system they're working on. So you give them the keys to production, they can mess up in production.

Yeah, this shared responsibility for security continues to extend here. We know that excessive permissions is a normal part of operating an environment on cloud. This is just the way the reality works.

That doesn't quite match up to the architecture diagrams that we get really on. Uh, we do need to think about handing control of our wider systems to our AI tools and to put the same controls around those AI tools that we would've put around any of our, our other systems deploying changes out into production. Well, those changes should have gone through all of the staging and testing and the correct methodologies for deployment into production should have been followed.

KERA should have been provided the information about how you deploy out into production, not treated it like a deployment out, out into a test. Hey, maybe there was just a tag set on that particular region for that particular service that called production test. And so KERA was exactly following instructions.

It was mistaking, yeah, that had never happen, I'm sure. Uh, be careful what you let your AI do. It's the same as letting an intern do the same activity.

Be very careful. You should also be careful of quantum teleportation because it's now left the lab. It's not quite escape the lab in the way of a disease that's gonna infect everyone so much as having a field try Dou Telecom and connect successfully demonstrated quantum teleportation over 30 kilometers of live commercial fiber in Berlin, while it was still carrying normal internet traffic.

So commercially available hardware and commercial network connectivity provided a average teleportation fidelity of 90%. I'm not quite sure what fidelity means, uh, providing that quantum proving that, uh, quantum networking can operate in real world telecommunications environments. This milestone brings the quantum internet closer to reality, enabling future breakthroughs and secure communications distributed quantum computing and quantum data centers.

Tom, are you certain of your state or your position? Neither, both. Hey, was that a cat?

Okay. Yeah, I, quantum mechanics is, it's, it's like table reading at least for Natasha Romanoff. Uh, but the, the thing you need to know about this is that they were able to teleport data across a link 30 kilometers in length while that link was carrying other normal traffic.

So that fiber was noisy. One of the biggest problems about bringing quantum teleportation and communications networks outta the lab and into the real world is you're not in a lab anymore, so you have to contend with watts of issues. And that 90% fidelity rating basically means that they were able to get the data across the link 90% of the time.

Uh, and you're probably thinking to yourself, well, 90% sounds good. Yeah, it is. It's also two orders of magnitude more loss than we would expect on even the worst link that we would be, uh, using in, in any kind of, uh, production environment.

1% packet loss on your link, you'd be screaming at your ISP to fix it. So we're not there yet, but we're proving that it can be done. And this is not easy because in order to do this, you have to entangle particles on either side of the link and you have to introduce a filter in the middle to make sure that only those particles are traveling.

You have to make sure that you get the entangled particle over there. Now, when it works the way that it's supposed to, they instantaneously appear on either side of the network across any distance. And the 30 kilometer link is important because in a modern network that is carrying traditional data across a fiber at 50 kilometers, you have to have a signal booster or repeater in order to re uh, generate the signal to be able to send it to the other side.

So if we could have that happen without having to build those repeaters in the middle, that's big for people. And this is absolutely the forefront of secure communications. com and search for quantum networking.

And there is an amazing video from Tim z Getty of Cisco talking about this. He uses tennis balls to explain how you entangle particles, and he helps people understand why it is absolutely secure to be able to send entangled particles over a medium. Because to al's joke, once you observe the particle, you've collapsed the position and now everything is done.

So, uh, brush up on your quantum mechanics kids, because you're gonna need to know that for the future of computing. Alright, let's take a closer look at a story that involves, well, what people hope would be unlimited power. As artificial intelligence drives unprecedented electrical demand, major tech companies are increasingly bypassing traditional power grids by building data centers at their own onsite energy generation across the us.

Firms like Meta and Oracle are turning to natural gas powered behind the meter facilities to avoid long grid connection delays in rising utility costs. According to research from clean view, dozens of these projects are already in development. Supporters say that self-generation eases pressure on strained grids, but critics warn of things like higher emission, higher emissions, reliability risks and rising costs for public utilities signaling that AI's rapid growth is beginning to reshape the nation's energy system as much as it is the technology sector.

So al maybe it's a little bit different in New Zealand, um, but up here, uh, data centers seem to be wanting to gobble up all this power, but now that they can't get it as reliable as they want, they're gonna start building their own generation facilities next to them. Is that a good idea? Yes, yes, absolutely.

It's, uh, provided you have something that's gonna be the source of energy for that generation here, it's natural gas. So now you have to build your data center close to a natural gas pipeline or a a somewhere that's extracting natural gas. So it changes from being, you need to be on the, the wires that deliver the power to you need to be on the pipeline that the power doesn't, it take away the need to be close to somewhere that you've got the source of energy.

Uh, the thing that does shift is the idea of using those micro nuclear plants that we talked about in the rundown last year, and that don't seem to be any closer to happening. Modular nuclear, uh, is still the research project, but fundamentally, yes, this idea of doing your generation of all of your electricity onsite at the data center seems a great thing. Uh, one of the things that historically data centers have always had was a, a ability to continue to operate when the grid disconnect.

And so there'd be generators and there'd be, uh, UPS systems on site in order to continue operations when the grid goes away. Well, what about if we start with those as the primary sources? You know, we've gotta have them there anyway.

They've gotta be big enough to support the workload, the actual maximum load of the data center. Why don't we just use those? Uh, historically they've usually been big diesel generators, but, uh, you know, gas power supply in or some other source of electricity here in New Zealand, a lot of our electricity comes from hydropower.

So I can imagine building a big data center next to a hydro dam and just plugging straight into the generator floor. Uh, no transmission losses, real efficient use of the power that's being generated. Um, be a bit of a concern if something went wrong with that dam and your data center got flooded.

But, you know, maybe you put your data center a little, little bit, uh, upheld. But this idea of generation close to consumption is a really good thing. We talk about food miles of generating or growing our food close to the people that are consuming it.

Well, what about power miles? That's definitely loss. And it also helps us with this idea of having impact on the local, um, the local community.

We were talking about this in some of the pushback earlier of, uh, communities not wanting mega data centers next to 'em because it'll increase the cost of utilities, particularly power. Hey, but there's one thing missing from the story. Water cooling.

All of these massive data centers need cooling. How are you gonna likely generate your cooling without using some massive supply? Tom, I think the, the real story that's sort of been buried in this is how do you keep these damn things cold over time?

Oh, we just launched it into space, right? No, we don't, we don't do that. Uh, by the way, uh, uh, check me out on Blue Sky.

I actually linked to a great video from a guy named Kyle Hill who explains why that's this really stupid idea. Another reason that Kyle Hill is a great person to follow, especially on YouTube, is because he really is a nuclear physicist. So he understands the, the challenges of using nuclear power.

Um, these things are hard to do, right? Like we've known that for a long time. That's why we've never built a power plant next to a data center.

Um, here's the problem that I have with it though, because more than any other thing, when you build a facility that's gonna consume a lot of power from the grid, you pay for that power. We know that because there's a utility bill. I don't even want to know what the electricity bill for your average data center looks like, but I'm sure it's astronomical and has lots of commas.

But when you pay that bill to the people who generate the power, do you know what they do with the money? Okay, do you know what they're supposed to do with the money they're supposed to use to upgrade their facilities? They're supposed to be able to increase power generation.

They're supposed to make it more reliable. Uh, they're supposed to explore ways to create power generation facilities that have a lower cost per watt so that they can make money off of it. And we've seen that through the use of things like solar and wind and hydro.

Do you know what happens when the companies who are consuming massive amounts of power don't pay for it from the grid? Yeah. Yeah, you guessed it.

Kids, uh, there won't be any grid upgrades because when it's just, uh, ma and paw baker's house that's running off of the grid, we don't pay enough in electrical, uh, consumption costs to, uh, to be able to do that. Worse yet, you think these plants are gonna be running at a hundred percent capacity at all times. I bet you they build a slightly bigger plant than they need with the projection in three to five years that they'll consume more resources.

But what do you do with the power generation? Because you can't just shut these things down when they're, you know, producing a certain amount of power. Well, how do you handle it when you have a solar array on your house that's generating power all the time and it fills the battery on your power wall?

Oh, you sell it back to the utility company? Oh, well now I have a, a facility that's generating power and it's dumping it back into the grid. And that's better for everybody except now the utility company is paying for the power that they're not maintaining, which means your rates are gonna go up because they have to pay for that power.

And more than likely what's gonna happen is either a scheduled expansion is going to be delayed because they don't need it anymore, or they may just cancel it all together because, oh, well, now that we have these extra natural gas plants that are throwing an extra a hundred megawatts onto the grid, we don't care. This is the typical, uh, late stage capitalism playbook of, um, you know, publicize the losses, privatize the profits. And so we're privatizing a power plant to get this thing online.

And then maybe if we can cut a deal with the utility company to make the power cheaper to consume than it is to run this, then we'll consider buying from them. But that means that the price per wat is gonna be so low that they can't really profit from that to increase the grid. And then everybody's utility cost is gonna go up because there's more power being generated and dumped into the grid.

And ultimately everybody is gonna lose out because what do you think is gonna happen if they built this small plant to get things running until they could get a grid connection? And then when the price of electricity goes past a certain point, they're gonna disconnect from the grid and go back to using that and expand that plant too. Look, I get this, this is Harvard Business 1 0 1, this is how you operate something, but you're dealing with an electrical utility and utilities should work differently than everything else because they're for the greater good.

And I know that those two words are a very foreign concept to people in the industry today. I don't care about effective altruism. Sometimes you have to lose a little bit of money on something to make everybody's lives better.

Now I know that AI companies are no stranger to losing massive amounts of money, but rather than investing it in hardware that needs to be built up, you know, year after year and buying out the entire flash storage and RAM market for the year of 2026, why don't you spend a little bit more money upgrading power plants so that everybody can consume that and pay for it and we're all better off. And oh, by the way, you might get to run your GPUs this, this year. Just a thought.

Well, you know what else is a thought that Tech Field day thing that we do when we're not ranting about the news every week and the next event that we have coming up is one that involves data centers and all kinds of fun stuff. And al that one's right up here, alley, isn't it? It is.

I'll be returning to, uh, Silicon Valley for March the 11th and 12th of Cloud Field Day. This is the 25th cloud field day that we're running. Uh, it's definitely getting to be, uh, uh, mature, uh, experience.

I had a great panel of delegates coming out to join us. A couple of new faces. I'm looking forward to meeting some people I've never met before.

Also at least one brand new delegate on their very first Tick Field Day event. That's always fun. Uh, looking forward to Hammer space and to VM whereby broad as our primary presenters.

And I'll take a a little look at some more of the things that future and research does. Maybe we'll have somebody else also as, uh, as presenters coming up for that. com, as well as across our LinkedIn and uh, YouTube channels.

Of course, I don't have an exclusive on those channels as well because Tom, you're gonna be returning to the area as well. I am, I'm gonna be at RSAC this year and I'm gonna be bringing some Field Day friends with me. We're gonna be having Tech Field Day Extra at RSAC March 23rd and 24th Monday and Tuesday of the show.

Uh, we got great presentations from Veeam Object First and Commvault. Uh, stay tuned. We might be adding some more names to that list very soon.

Uh, we're also bringing in some friends from the security space. I'm gonna be meeting up with some folks, probably even talking to some companies. com.

Um, I'll be there, uh, you know, we'll have some fun. And then I will be back April 8th, ninth and 10th for Networking Field Day. Uh, that is a jam packed event.

We have a ton of sponsors. We have a lot of delegates, some new faces to the crowd. Uh, in fact, uh, I was telling somebody the other day, uh, I finally got to invite somebody that I met at, uh, in the tech space about, uh, 20 years ago.

So it should be a fun time for everybody involved. But you know what else is a fun time is the rundown. Every Wednesday, new episodes, new news, you don't wanna miss it.

Uh, check out YouTube for all the latest. Uh, you can also subscribe in your favorite podcast application of choice if you'd rather get the audio version. Uh, we stream this on Techstrong TV and on all of our other Techstrong future and group programs everywhere, whether it's the Security Boulevard podcast, the Tech Field Day podcast, all the things that we do.

We're gonna be back next Wednesday to talk about all of the IT news in the week. That was as soon as I can get that generation facility built next to my house. Until then, for me, for Al, for everybody here at Tech Field Day, we hope that you have a great day and we'll see you next week.