AI Espionage Risk, Financial Pressure and Infrastructure Reality

Hello, everybody. Welcome to the Techstrong gang. I'd say happy Monday, but, well, that might be a little premature, shall we say, but we'll see how the day's going to evolve here because while there's scandal.

It's hard to believe, and it's amazing to me. You just can't make this stuff up. But let me introduce our members today.

Jon Swartz, good to see you, my friend. You are not in your usual location. No, I'm in Nashville at Commvault Shift.

And as you alluded to, Mike, yesterday was just another slow weekend news-wise. Nothing happens. Too silly.

Amanda, good to see you again. How are you? Doing well.

And Tom Hollingsworth, of course. Good to see you, my friend. Coming to us, once again, from Oklahoma.

Yes, where absolutely nothing exciting has been going on at all. I promise. All right.

Somebody's probably trying to build a data center somewhere in Oklahoma, but we don't know for sure. But anyway, the issue of the day is Jensen Huang, the CEO of NVIDIA, has admitted to an affair, which in this day and age is not in itself maybe the worst thing that could possibly happen, but it apparently involves some sort of spy. And this is where things get a little more interesting.

But John, there's been an update this morning on this story involving a Commerce Department investigation. But bring us up to speed here. Oh, yeah.

So the latest just happened. New York Times, which is really killing it on this story, is reporting that there was a US probe of Megaspeed. That's the company affiliated with this paramour of Jensen Huang, who is alleged to have some sort of spy connections to China.

Anyway, the US was going to probe this company because it's known about the situation for quite a while. But that probe was stopped after NVIDIA intervenes, which kind of brings to mind the government's involvement in this. And to be really quick, Donald Trump knew about this evidently while it was going on.

Marco Rubio is warning about what the implications might be. And as you said, Mike, Jensen Huang admitted to an affair with this woman named Alice Huang, no relation. The board of directors of NVIDIA basically announced Sunday that they're going to launch an independent investigation looking into whether Jensen Huang was targeted by a Chinese espionage effort.

And it just gets seamier and seamier the more you look at it. It all kind of ties to Jensen's many travels to China and his lobbying on behalf of NVIDIA and others to sell more high-end chips to China. And it's beyond a soap opera.

It's kind of reached a point now where there's open speculation about what his future, Jensen Huang's future, is going to be with the company, and based on this relationship, based on the government's interaction. In a sense, there is going to be a hearing behind closed doors November 18th. Allegedly, he's been subpoenaed by Congress.

The company announces its earnings November 17th. And it's going to be interesting to see how this plays out. It doesn't look good.

It's getting worse and worse by the day, especially when NVIDIA puts out these press announcements that refute almost everything involved in the story, only to have to backtrack and acknowledge. But the fact that the board said they're going to do an independent investigation, to me, is kind of a nudge to him, and it doesn't look good, and it's going to get worse. So whether he survives, do you think he will survive as CEO?

But it also seems at this point that we aren't talking about a crime. And at that point, is there going to be an investigation and then that will force an issue? And then we'll get into the espionage aspects of this thing.

But what's your read on this thing? Does he survive this or no? So he's one of the co-founders.

6% of the stock, so he possesses some substantial institutional influence that could help him withstand the inquiry, yet he doesn't have super voting shares, so he remains vulnerable to the board pressure and potential removal. There is so much at stake here. He is the center of the company that is at the center of the AI boom.

But his timing, and the fact the timing of all this is quite curious. I'd be interested to see what you all, what Amanda and Tom and you, Mike, think about this, but with the Democrats coming in to control the House and the Senate, and they're already talking about bringing people up before them, this is really dire for him. And I don't know.

The Justice Department, the government as it is right now, has been so up to now looking the other way or sympathetic towards him. But given the American sentiment about tech bros and the oligarchy, this is a devastating thing. I'm looking at the NVIDIA stock, it's down 6%, which is not as much as I thought it would be, but this is going to have a cumulative effect on the company.

And it also brings into question if they need a successor. If it ever gets to that point, there's already talk about his daughter, who's really deeply embedded and really high up as an executive. They're already using her as a perhaps proxy to help as part of the new management team.

But this guy's so identified with this company and driving everything it does, this is a devastating blow to this AI boom, I think. Yeah, and to your note, I do think that it feeds into the concern that's already there about Chinese tech. So it certainly doesn't help in that area.

So to Amanda's point, Tom, you think we're going to discover that the plans for the next generation of processors from NVIDIA are now, or should we just assume they already are in the hands of Chinese rivals, and that they will be coming out with chips sooner than we expect? I don't think Jensen did that, and I'm not trying to apologize for Jensen. I know that Jensen's a capitalist, and he wouldn't give away something that he could sell.

And I think that that was the motivation here, was not to give the whole game away to the market. It was to find a way to open that market for him to sell into. The key question that I've heard discussed over the last 24 hours is exactly when Alice became an intelligence asset because I have heard that their relationship began before- Yes ...

she became an intelligence asset. Asset. And that seems to be what we're seeing, is that she and Jensen had whatever they had, and then she was turned or was whatever, and that's where everything started from that point forward.

That makes it slightly better, but not a whole lot better. But what it means is that he had enough control in the situation to prevent this from becoming a honeypot. Because this is what we typically see in these kinds of situations, is the asset, it gets involved with the person, and then it becomes a blackmail situation.

Mm-hmm. I think that because it went the other way, where the relationship started before the intelligence asset situation happened, is that he was leveraging this to be able to make even more money in this market. It's still not going to save his job because, look, any situation that you've looked at over the last, I don't know, 50 years where a CEO's been caught doing something and then publicly said that he wasn't doing that or she wasn't doing that, and then the board was like, "Oh, yeah?

" They've all been fired, allowed to resign, whatever. Not because they disagreed with the behavior, but because they can't have the executive lying to the board. That's the thing.

We don't care what you're doing. You just can't lie to us about it. And especially now that we know for a fact, from what we're hearing, that people inside of the government were briefed on this up to a year ago- Mm-hmm ...

" It was like that relationship was an open secret in Asia, evidently. So if you read, I'm monitoring of the Asian press. It was pretty well acknowledged.

So in a sense, you're right, Tom. And it was a romance, whatever you want to deem it, and then perhaps that she was approached by these operatives about leveraging him. But I think about Mark Hurd, and I think about HP.

I think about executives who are not exactly upfront with their board, and it gets them in hot water. And I also think that the way this has been handled by Nvidia is just absolutist denial, makes them look really irresponsible, kind of clunky. But behind the scenes, if this new New York Times story is correct, Alex, this has even taken another escalation and a level of really cringiness.

I look at it past other companies too, as a warning that when you're a leader of any big company like this, you've really got to watch your Ps and Qs. So I agree with Tom in the sense that I don't think that Jensen gave anything of value deliberately to China, but we don't know what she did, and we don't know what she had access to and what she handed up. Correct?

Right. There also was the other story where-- It's just ridiculous. Nvidia's shipping product to a company that doesn't exist, that has military ties, has ties to the Chinese military.

A vacant office space that the New York Times wrote about. We're going to hear more of these types of disclosures, I think. And whether Jensen-- I mean, Jensen didn't knowingly share probably more than he-- Who knows, actually?

The level of confidence or level of what he would talk to her about. He may have gotten excited. He loves talking about his company and what it's doing, and maybe he talked to her about future plans or in general terms.

I don't know. Everything's open at this point. But John, the biggest problem that you talk about is when you look back at shipping these GPUs to a company in Southeast Asia that then rented them out to the Chinese government.

Shipping equipment to an address for a company that doesn't exist. These are the hallmarks of getting around things like embargoes and stuff like that. But more importantly, it's also the fingerprint of an executive who thinks they're above the law.

Mm-hmm. Who thinks that the rules don't apply to them, and I can do whatever I want. And I hate to say it, but you have to make an example out of this.

Yeah. No, I agree. You have to go in and say, "You broke the rules.

You created a national security nightmare. " Right. A couple of weeks ago, we were talking, it was on a Friday show, and Tracy Reagan brought this up, and I always think about this.

We were talking about how Jensen's everywhere, right? He shows up at every event. He's constantly traveling.

He's almost as bad as Musk. " And I started-- Now we might know or think that the reason why he spent a lot of time in Asia is to be around Alice Wong. I don't know.

But it just brings into effect this kind of governance and who's in charge. And the one thing that's going to be really interesting to see if any of his CEO pals who have used him and leveraged him so much at their announcements rally behind him, or we start seeing less of him. Maybe he'll spend more time stateside, but I never could grasp or understand how this worked.

Now, I don't know if this is true, John, but I'll toss this to you, but he runs that company in a unique way. He has more direct reports- 60. Yes ...

which is unheard of. And so if you do take him out of that picture, you're talking about restructuring the entire company. Yeah, he's got 60 direct reports.

So ironically, I feel terrible about this, but his daughter would travel with him quite a bit. She actually is the one who took a number of the photos of him with Alice Wong. But she's not up to speed.

Obviously, he is the architect. He knows this company better than anyone. If you pull him out, I've read some reports from Wall Street analysts who are speculating what the damage could be to the stock.

And some of them are hard to believe, up to 40%. I don't believe that, but we'll see. This whole era, it's like a, what was the name of that game?

Where you put the little piece, Jenga, whatever? They're at the center of the Jenga pyramid. That was- You're talking about pulling out the central piece right now.

" From now until the first of the year, we might hear some kind of discussion around that. I don't think it'll come to fruition. About three days into the new year, I think the conversation is going to change significantly, and it's going to be less about how can we salvage this company and more how can we make sure that this never happens again.

Because we know for a fact that letting all of the economic output of the United States be based around a cult of personality is probably not the best way to run a government or an economy. You know what? It's funny you mention that, Tom.

You think about what's just happened in the last week. So Elon Musk spent half a billion dollars on the midterms and that blew up in his face. They killed the Cybertrucks.

Anthropic's valuation has been nearly halved, and there's talk they might delay the IPO. Now you've got this with Nvidia. A lot changes, and it just shows the dangers and the fragile nature of depending on such a few individuals and companies to drive the economy as it has.

John, you forgot Sam Altman telling me that these are- I was going to mention, I have that. Please share, yes. But the thing is, you mentioned that.

Is every one of these people, Jensen, Sam, Dario, Mark, do we even care about him anymore? All thought that they were too big to fail. Not their companies.

They were too big to fail. And all it took was a stumble. And I know that for a fact because when I went on my first trail run on Saturday, I kicked a tree root at about a mile two and a half.

And despite the fact that I think I'm a good runner, I was not too big to fall. And boy, it was a fun crash. But that's the thing is it was the first stumble and then trying to right yourself with every step going forward, knowing that you can't do it.

And I think that's where they're at right now. They know that they can't stop this fall, and all they're trying to do is prevent it from landing on the concrete and having a hard crash versus landing in the sand and having a slightly less destructive crash. And I wonder what company or what companies are frontrunners in case Nvidia sees this big crash.

Who's lined up to take advantage of this opportunity, too? Oh, sorry, Mike. Go ahead.

I apologize. So the OpenAI movie, the anti-OpenAI movie comes out December 23rd, I think. But now there's the infamous photo, and I keep seeing it.

We got Trump sitting in the middle of all the bros, and he's flanked on each side by guess who? Jensen and Elon Musk. And Zuckerberg is kind of looking plaintively on the outside.

He's not next to Trump, but he's kind of looking like he feels like he's on the outside. Maybe that's a little bit of solace, but in a sense, his name is as dirty as any of them as well. Yeah.

I'm going to leave this here, but I would agree with you on this one point. This just becomes too much of a distraction, and it takes away from the company, and it will never go away. So at that point, they need to find another way.

Now, does Jensen, if he survives all this, stay on the board, and he's still influencing the company? Perhaps. But to Tom's point, this isn't just somebody having an affair.

Criminal activity is being alleged, and if that gets investigated to the point where it has shown to either I lied to the board or I inadvertently disclosed stuff and then covered it up, it just all ends badly. All right. So let's go on to what is happening here with all of these AI investments.

And I think we have previously talked about where they need like $8 trillion in revenue ultimately to justify the AI investments that they're making by 2030. And we've seen over the weekend, I think Microsoft was talking about how it's shifting workloads away from Oracle because, well, it doesn't need a secondary player. And we're also seeing OpenAI raised another, was it 11?

11 billion, I think. 11 billion. And I guess they were looking for more, but is there enough money out there to fund all this stuff?

Now, the market today is up being somewhat forgiving for AI stocks, and it's bouncing all over the place. But Amanda, what's your read on all of this, and how does it play out? " Yeah, it seems like more companies.

So there was this big move to the neo clouds and all that, and then it seems like a lot of companies are taking back their data and their compute, and they're looking at data centers. And we have this huge data center boom. But the problem with this is the fact that most of these are stagnant or held due to regulatory issues or people in the communities causing recall elections, for example, in my community over this issue, and how the electric grid in general is going to support all this new infrastructure and data centers.

So the question is, if you are not using other companies and you're going this data center route, right now there's a big stall in the data centers. And that's also from a price standpoint too. So there are a lot of issues, water, electric price- Mm-hmm ...

the land. So how are we going to move forward when they're looking at data centers, but those are kind of on hold? So let me tell you a story because I'm here at KubeCon in Salt Lake, and Red Hat is here, and they're talking about how they went and used open weight models to shift all kinds of prompts and workloads away from frontier AI model companies, cut their costs on AI by well over close to half.

And they're probably not the only company that's going to go do that. So Tom, as I look at all the math that goes into this with these frontier AI companies, I think what we're going to see is more people realizing that 80% of what they need to get done doesn't need a frontier model. So will that make the math problem even worse?

You should ask McDonald's about that. And the reason why I say that is because there was this huge boom in McDonald's. They were getting more and more money, and then here recently, it's trailed off significantly.

As prices have gone up, what have people said? Well, we're not going to go to McDonald's, we have McDonald's at home. And that's what you're seeing, is that companies are looking at what it's costing to run these models.

They're looking at what the-- Because the costs never go down, right? This is like a baseball player contract. It never goes down.

It always goes up. So eventually what's going to happen is they're going to go, well, if I can get 85 or 90% of what I get out of this frontier model with something that I can run in my own organization, then why won't I do that? Because this isn't like a light switch, right?

It's not a binary where it has to be either all of the AI performance or none of it. We can get very close to it. And that's the same thing we've seen repeated over and over again in tech.

Cloud is another good example of that. We're going to take the ideas behind what we're doing, and we're going to run it on-prem so we can save money. And that's where the companies are going to make their money back, right?

They're not having to spend millions upon millions upon billions of dollars every year to rent this capacity. And then that changes the economics of how all of this is supposed to work. And how is McDonald's fixing their problem?

Oh, well, we're trying to find ways to bring people back in, and we're trying to do this. And of course, if you took a survey of people and you said, what's a way that McDonald's could bring people back in? What is the number one thing they would say?

Lower costs. Lower cost. Mm-hmm.

Lower prices. Get it so that it's now convenient enough for me to make that choice as opposed to going, no, I have hamburger meat at home. They will never do that because in capitalism you can't.

You cannot lower the price because what you've proven is that people will pay, what is it? 012 cents per token. 0011 cents per token because then that's admitting that the pricing doesn't work.

They won't do that. And they'll ride this thing all the way like Dr. Strangelove.

They'll ride the bomb all the way to the ground- ... because they would rather explode in a blaze of glory than bail out at the last second and admit they couldn't make it. Mm-hmm.

John, what's your read on this? Because everybody that I talk to who favors the continued investment levels in AI cites this insatiable demand for AI. But I think they're not accounting for the fact that there's other places to get the AI.

So maybe are they kind of got some rose-colored glasses on and so there's some wishful thinking going on here? There probably is. That was the whole part of the mindset of the AI data center expansion, and maybe they've got to do something to steal from Apple, have to think different.

Because there are two things that are going on right now. Rising electricity demands at the same time as local opposition is escalating against these new data centers. And I think about the epicenter of data centers is Texas.

You've got the incoming governor-elect who ran on against data centers in large part or some sort of regulatory stance or some sort of scaling back. So they're going to have to look at other alternatives. I've started to hear, remember, was it about a year or two ago, about they're looking at nuclear as an option, nuclear power.

I'm starting to hear that murmurs again because right now, as you said, Mike, the math is it doesn't work. Mm-hmm. The amount of money they got to spend versus the amount that's being generated, amount of revenue that's been generating is an incredible shortfall.

And maybe that's why there's some hesitancy or second-guessing about the Anthropic IPO. I don't know. There's all these factors, and it's economic, geopolitical, personality driven, that have all kind of conspired to make us rethink the way we've been approaching AI and this kind of rabid boom.

And I think maybe it's part of a kind of a little bit of a pullback and people are going to look at alternatives. Mm-hmm. I think.

They must be talking about this within the C-suite. Yeah, the infrastructure just isn't adding up. But to your point, just down the road in Abilene, which is an hour and a half from us, they are looking at it.

They have a nuclear option they're looking at. Nuclear option, that's a good phrase. Speaking of Dr.

Strangelove now. Flynn-Pickens riding the bomb all the way to the ground. Mm-hmm.

All right. Well, walk this through with me a little bit. Is there another way to get at this solution for these frontier model companies?

Or am I looking at some sort of bubble that forces, I don't know, OpenAI and Anthropic have to merge because there isn't enough support for the two of them? How far does this go? That is an interesting idea or concept because I don't think this is going to change.

I think it's actually we're going to deepen our way into this quagmire that these companies find themselves in. And it has material effect on Anthropic. Allen brought it up a couple of days ago, maybe last week, that they could delay the IPO.

And now, with OpenAI, they've pushed theirs back to next year. It just seemed too good to be true for all these folks. " And that extends to Nvidia and some of the others.

And they're not as much in control as we thought they were. " And Microsoft, if they don't have to share revenue, why would they? It's almost like they're distilling everybody's models and turning them into something that they can work.

Oh, wait, that's bad, I forgot. We don't do that. I think that the market will correct itself eventually.

And yes, John, we're going to go a lot deeper down this rabbit hole before we finally realize that, yes, there is a bottom to it. And the bottom is going to be when we have run to the end of the LLM branch of this and realize there's no path to AGI and there's no path to recursive self-improvement. Mm-hmm.

Whoever comes along after that, and I am not betting on OpenAI or Anthropic to come along after that because they've spent way too much time and money in the LLM branch to suddenly swing to whatever comes next. Whoever eventually does and create that RSI model that will basically improve itself going forward, they're the ones that ultimately win out and capture the market. Because I heard this great conversation late last week.

OpenAI and Anthropic are essentially betting on the future, right? Burn $100 billion now for the possibility of developing a system that can make you $100 trillion later, and you'll never have to work another day in your life. That's a bet you have to take under the capitalist system right now.

But if that bet doesn't pay off- Which a lot of skeptics are now saying that it probably won't. Then, yeah, we'll probably be reading a news story in four or five or six years where whatever remains of one of those two companies gets bought out by the other one in a fire sale because it's just too expensive to keep running down a track that has no payoff when other people have jumped to a completely different branch of the problem and are spending billions of dollars of capital trying to get there. Right.

Amanda, many of my friends have been watching their funds increase over the last couple of years around AI, and they've all been quite pleasantly pleased by it. But are you hearing the average folk out there starting to wonder about how much money they've got wrapped up in these AI stocks and maybe they're rethinking their positions? Well, we did see kind of a crash this week in a few areas, but I think with anything...

How do I put this? With anything AI related, I think we're still in an upward trend. AI isn't going anywhere, it's just a matter of which companies are going to come out the most successful and I guess keep growing.

But in general, I think if you're investing in the right AI stocks, it's an upward trend that we're not going to see go away anytime soon. But I'm not a financial advisor or anything like that, so. That's a good caveat to put in there.

We probably should put that in on every show. But yeah, I'm going to leave this here, but I think Amanda's right. AI's not going away.

There's going to be increased demand for it, no doubt. But that doesn't necessarily mean that the leading players associated with that are the ones that are ultimately the winners here. So be careful and figure out what the smart money's doing and watch them closely, because if you start seeing them heading for the doors, you might want to be not too far behind.

All right. I'm going to switch gear slightly, but we're going to talk about AI infrastructure, because Tech Field Day had another event on this topic, and Tom was there, and was the host for this thing. I think you were the host, right?

It was Alistair Cook, but I tuned in because there were some really great conversations going on. And I feel like this relates to what we were just talking about because, at least from what I can see, the folks who build the infrastructure for AI, man, they are just making that more efficient day by day. So, what were the highlights of this thing?

Alistair brought up some really good points when I talked to him about this because really what they're trying to do is they're trying to make AI and agentic solutions more of a production capability. Because when you think about all the stuff that's going on, look at the amount of money that you have to pour into building systems to do training, right? And they're kind of their own little islands that do that.

But when you get to inferencing, that has to integrate with everything. It has to be part of the broader business strategy. It has to integrate with AppSec and things like that.

So you have to work at making it more efficient, more cost effective. And so that's one of the things that a lot of the companies highlighted while they were there, is how they can integrate into an existing enterprise IT platform. Because you don't want to have to go out and spend millions of dollars on something that can really only ever be used for one purpose that has a finite end date.

And that also speaks a lot to the way that you design those systems and how you engineer everything together. Look, we all talk about the magic of AI and how special and important it is, and it's still running on servers that have network cards that need to be cooled in a rack as in a row in a data center. The only thing that has changed, other than maybe the size of these things, is the efficiency of the workloads.

We're still building them the same way we did. And in an enterprise, the fact that you can switch those things out and change their workloads is super important. I'd say we probably didn't get here without having some cloud functionality built into the enterprise data center, because if we were still trying to build them like we did, I don't know, 10, 15 years ago, I don't think it would work out quite the way that it did.

But ultimately, I think what's really critical for enterprise IT departments is you have to get to a point where you are the ones that are offering AI as a service to the rest of the organization. Think back to the heady days of cloud, when all it took was your corporate Visa card to do some shadow IT work, and how out of control that got, and how once we brought everything back into the control of the IT team, things became more efficient. We had standardization.

It caused security concerns to disappear. And then think about what's happened in the last five years. " And unlike the Cylons, none of these people have planned because they're jumping at the first thing that tells them that they should be doing stuff.

" Mm-hmm. So is it your sense that IT teams are going to take more control over these AI deployments to achieve this level of efficiency because they understand the infrastructure and the cost associated with it, and they also understand that I can maybe do this with less costly models? Yes.

And the reason why I can tell you that'll happen is I want you to flash back to the heady days of March when token maxing was a thing, right? Remember there were the tote boards inside of Meta that were, "If you're not maxing your tokens, you're not being efficient," and everything like that. And where are we now?

Oh, yeah, it turns out that doing busy work really doesn't help you there. What brought that under control? Well, when you have one central area where you can see the cost, when you can see what it takes to get to that point, then you can start matching that, and you can start understanding that.

We talked about this a few months ago where, was it Accenture, I think, figured out that something like 40% of their workloads were people that were literally converting PDFs and PowerPoints. That was a huge chunk of their token spend. It wasn't deep research, app development, reverse engineering, Photoshop.

" It's like, that's not the best use of what we're doing. But only when you have central control of that through an IT team can you see what's going on. Because if this is just a bunch of shadow IT, shadow AI used on credit cards, it's not going to help anything.

Mm-hmm. " Mm-hmm. Especially with token usage getting so expensive.

Mm-hmm. Yeah, it's funny. The actual cost of the token, I think, is coming down, but it's just that we're using it more and more, so the total cost of the spending goes up.

Or is that the case, Tom? Will the cost of the token get any cheaper? It's the idea that initially they wanted to throw everything in, right?

Just use it as much as you can. We need you to get used to the idea of it. And then when they realized that it wasn't a sustainable business model or there were ways to do things, then they changed it, right?

Oh, well, the per token cost is coming down, but the amount of tokens that you need to accomplish the same output is going up. It's every PE firm play that you've ever seen, right? It's like, "Oh, no, we're not going to give you the free fries with the entrée.

" I don't know. John, is it your sense that maybe the infrastructure players at the end of the day are the big winners in all of this, and maybe the investment should be going over there versus the model? It should be.

Yeah, it should be. You should be advising these companies. Yeah, no, that's where really the rubber hits the road.

I just think about these enterprise leaders under all this pressure to implement AI, given all their options and the costs and the learning abilities, the upskilling of their employees, and then the whole idea at the same time we're going through all the restructuring at these companies. It's like trying to thread a needle in a hailstorm. It's quite an undertaking, and they're going to have to lean more on the infrastructure folks to help guide them through because it's quite dizzying the way things are moving and the way things are up in the air.

Just what we've been talking about this show is the unpredictability is just over-the-top. It's crazy. Amanda, will there be like a 12-step program for AI for a lot of these companies where they're going to have to go in there and kind of get everybody detoxed a little bit to make sense that what to use this for and when to use it?

I think there definitely needs to be some sort of educational awareness program, just better communication between departments. Thanks. Tom, was there anything among the people who came to visit that you said, "Wow, this is the most amazing thing you saw," or anything that leapt out?

You know how it is, right? " Yeah, since what, exactly? Companies that we've seen do a lot of great things over the years.

HPE had a great co-presentation with Equinix. The delegates were really excited. They had a roundtable on Wednesday afternoon where they kind of got to chit-chat about some of this and bring up some topics that people maybe weren't really thinking about.

There's a lot of experts there, right, that understand this stuff. And I will tell you that sometimes the experts can be a little snarky, right? I'm not only a member of the Snarky IT Experts Club, I'm also the president.

But where the value in the snark is, is that a lot of times what they're trying to call out is the disconnect that they're seeing from the messaging to the reality. " And then what you need to do is you need to step out of your cubicle or your office, and you need to walk down the hallway, and you need to find your average IT worker and say, "Okay, show me how this works. " And watch them, because I promise you, they are a better bellwether for what you're doing than any analyst, than any board member, than any talking head on pick your favorite financial show on TV right now.

They're the ones that ultimately have to decide whether or not your company is going to sink or swim because either it makes their life better or it's so complicated and useless that they just throw it into the drawer along with their mini USB adapters and that keyboard that they promise they're going to throw away this time, and they'll forget about it. I will say I have faith in engineers in particular, and I know they abhor a vacuum. " And I think we're going to see all kinds of interesting innovations in the hardware space in the next 6 to 12 months that will get us to the point where AI becomes more cost-effective and more efficient.

But I don't know. John, am I- No, you're right. That's what's happening.

That's what I hear from these companies is that it's just trial and error. And to Tom's point, thank God the humans are actually considered in all of this. Like they finally started going, "Well, how does it affect the employees?

" But I do think it's a multi-step process and it's a work in progress. This too much too fast culture has whipsawed so many companies to the point where they're trying to figure out how do we slash costs before we actually implement some of this stuff. So I hope.

I don't know. Amanda, you used to work on Digital CxO with us and you're doing the podcast now on there as well, but do you think that these folks are going to play a larger role in maybe helping to drive some of this because they have some hard-won experience on transformations gone wrong? Yeah, absolutely.

It goes back to what Tom and John and I think we're all saying is, just sitting down and communicating with the actual users and the subject matter experts what works, what doesn't work. And then also having good documentation processes throughout. And also talking with other business leaders, coming together to collaborate and see what's working, what's not working outside of your own company could be useful.

So go listen to the Digital CxO podcast or other podcasts for tips and advice. Don't just stay in your bubble. Tom, do you think FOMO is dead, over, done with, or as it goes for AI, or are we going to be moving to the point now where maybe more rational thought is applied?

Rational thought in IT? Surely you jest. I think what's going to happen is that the budgets are going to dictate more about where the FOMO goes and that the speculators who want to make money off of the next big jump are going to be moving on to something pretty soon as this becomes a pedestrian daily thing.

Look, I survived the SDN wars 10 years ago and how that was going to be a radical shift in the way that IT works, and now it's a feature on a checklist. Although I'm putting my early money on quantum. Don't miss out on quantum.

Got to get on top of this. Get in front of this. Come on.

Yeah, got to get in front of it. Do you want another IT thing that takes tons and tons of investment for very little payoff? Then I got the thing for you.

How do you feel about liquid nitrogen and fuzzy computing? Do you know what a qubit is? As we say here in New York, we have a bridge we can sell you.

Yes, actually I've got another bridge I can sell you cheap. It's in Arizona. All right, cool.

All right. John, you were in Nashville at this conference, so I guess we're going to hear from you later this week. Yeah, I'm going to go to a briefing and talk to some of the executives.

I have the materials so I'm going to write a story for tomorrow. Yeah. All right.

We are going to be talking about all things related to Kubernetes for the next couple of days from here at KubeCon, and of course, that will have a big AI spin because one of the things we're talking about is all those AI workloads run on Kubernetes, and maybe we can make Kubernetes more efficient so that we can increase the utilization rates of GPUs. What an exciting topic, I know, but that's where we are at this point. So here we go.

Hey, I want to thank everybody for sharing their knowledge and their insights as always. Thank you all for spending some time with us. Please stay tuned with the rest of the lineup for the Techstrong TV replay.

And with that, here's a word from our sponsor because, well, it's our favorite because it's us.