Linux’s CVE Flood, Google’s $1B EU Fine, Samsung’s $22.8B Mistral Bet

Hey, everybody. It's Friday. Welcome to Techstrong gang.

And guys, Friday. In the summer days of July, used to be the dog days, that's a baseball term, where really nothing happens, but presumably almost everything happens now. This tech cycle is absolutely crazy, and same even applies to sports.

I didn't mention this in the green room, but LeBron James just signed with the Philadelphia 76ers. So he just dropped that. So it's a huge news cycle and a big tech cycle.

And who better to help us than the three amigos who join me. I'm John Swartz. I'm in for Alan, who is in a meeting today, and Mike, who is, I believe, on his way back from Quebec.

He was up in the Montreal area. So they'll be back on Monday. But first, I want to introduce you, and we're going to go west to east, and we're going to start in New Mexico.

Presumably, that's where Tracy is. Hi, Tracy. Hi.

Yes, I am in New Mexico, where it's hot and dry. But you had a pretty eventful day yesterday, I believe, as before you wing off to Ireland. Maybe you could fill us in on- I'm headed to Ireland and spent yesterday getting a root canal.

What a nice thing to do. At least it's not going to be a root canal in Ireland. God.

You know what? I'm jealous. How long are you going to be there?

Just about two weeks. Yeah, we're going to go and cruise around. Going to start in Dublin, go to the south, then go to the west.

Then going to go across the Midlands, and then spend a few more days in Dublin. So it should be an interesting experience. We aren't going to some of the top spots, though.

We're not going to go to Dingle or the Ring of Kerry because I like to hike where there's no people. Mm-hmm. So we're going to try to hit the national parks.

And I've done some digging on the trails that the locals like to do, so that's what we're going to try. That sounds wonderful. Well, have a great time, and we're happy you're here today.

And we're going to move to Texas. Jeff Reich, are you in Texas today? I am, yes, very much- And Jeff- ...

just a few hundred miles east ... And I'm wondering if it's Friday and Jeff is on the show, it must be National something day. It is National Tequila Day.

I won't play that whole thing. There we go. But yes, it is.

Well, yes. We're happy to have you here. I love your background.

I love the backdrop every time. It's always appropriate. Thanks.

Well, I like to celebrate, and what better way to celebrate than with tequila? Absolutely. And then, of course, ending up in North Carolina.

Are you there, Jack, in North Carolina? I am in North Carolina today, yes. Okay, good.

Fresh off his triumphant appearance yesterday, Jack Palmer is back with us. Yes. And how are you doing?

I'm doing pretty good. I actually have a question for Jeff, though, is did you feel the earthquake the other day? I did not.

I'm about 400 miles east of where it happened. Ah. So that's something not everyone gets about Texas.

In fact, I would offer Tracy had a better chance of feeling it than I did. I'm glad I didn't feel it. I was raised in California, and earthquakes freak me out because I've been in many big ones.

Yep. So they're not fun. Yep.

I remember being at game three of the 1989 World Series. Mm-hmm. That would've been horrible.

Yeah. That was a long night. Yeah, that was a long night.

Took hours to get home. But anyway, I digress. We have a pretty varied show today.

It's interesting. We're going to start with security, then we're going to go into regulatory issues, and then finally talk about chips and an investment that's really interesting. But we're going to start first with what Mike called, I think it's appropriate, the Linux security strain.

So I'll just do a little bit of an intro, and then I'm going to toss it to Jack. So systems administrators and enterprise security leaders are grappling with the overwhelming wave of security advisories following the release of 432 Linux kernel CVEs across Sunday and Monday. This surge in security disclosures reignited some industry debates over vulnerability prioritization, operational patch management, and the disruption, increasingly, of AI and automated bug discovery.

Jen Shallman, who is the CISO architect at Akamai, warned the sheer quantity of the advisories renders individual code review and traditional patch prioritization virtually impossible for enterprise environments. Linus Torvalds actually had referred to this potential problem and said the pace was almost entirely unmanageable back in May. So Jack, I'm wondering first what you gleaned from all this, and how seriously should we be treating this?

Well, it's no surprise that we're uncovering a lot of critical issues and security issues with AI, right? We've been talking about this for a while with Mythos, and now there's a number of different AI models, LLMs, that are targeted specifically at finding these types of security issues. And I look at this as sort of this is both a good thing and a bad thing.

The good thing is we're actually finding these things, and hopefully finding them before the bad guys find them. The bad thing is it makes it a lot harder for the developers to do feature functionality when they're spending all of their time working on fixing the security holes. Now I also sometimes think that that's a good thing.

I think we really do have a lot of security issues that have to be fixed. I think from, and I've talked about this before as well, that you have to look at the two different sides of the software chain. One is the producer side, and I know Tracy's deeply involved in being a software producer.

" Okay, that's easy to fix. " Nobody's paying me to fix these things, right? This is free open-source software.

There's not a lot of money behind it, and not a lot of resources available. " It's going to take me time, and it's going to take me time once I fix those to validate it and make sure that I get it right. And I don't want to rush a patch out the door by fixing something and breaking something else, or introducing another security hole.

So this is a really big challenge on the producer side. On the same token, on the consumer side, the entire world, every web server you ever looked at, all of the AI engines, all of the infrastructure, 95% of this stuff is all running on Linux. " How many of those affect me, and what do I have to do in response?

There's not a patch available for 399 of those 430 yet, because it hasn't been produced on the producer side. So as a consumer of this, what do I do, and how do I manage this? And this is a very big, scary thing for consumers of software, because with this many issues, they don't know how bad it is and what they really have to do.

So it's very scary and concerning. I was going to ask this of Tracy perhaps, given what she does. So given the sheer numbers, how can enterprise security teams modernize their triage pipelines, prevent alert fatigue, I guess you would call it, when AI-driven bug hunting tools generate hundreds of maybe low-level CVEs faster than humans can review them?

Well, there's already tools out there to make this easier. But we have a odd habit, we meaning the industry, developers, is we keep focused on the vulnerabilities that have been reported, and we don't know how to ask the right questions. So as Jack pointed out, it's the response that's important.

And a CVE tells us that a weakness exists somewhere, right? But the report doesn't tell us if the weakness can actually hurt our production system. So I feel like we keep addressing and getting freaked out about this number of CVEs that are coming our way, and we spend less time trying to sort out what is actually in our threat landscape.

" And we don't ask that question enough. So there are tools out there that the OpenSSO has been working on, other open-source tools. For example, VEX.

If you haven't used VEX before, it helps explain whether a CVE actually affects a specific product, but the decision must be supported by real evidence, so there's still work that needs to be done. Our open-source project, Ortelius, maps a CVE to an endpoint, so you can answer that question, "Is it in the threat landscape? " And it also brings up de-bloating.

I bring it up all the time. We need to learn to de-bloat some of the software. I would bet that the 400 vulnerabilities that were reported, some of it may not even be in code that is being actually distributed.

I don't know. But it may not, and it could never be reached. So modernization is the right term, I think, John.

Yeah. There's a show I like to watch on, I think it's on Prime. It's called "Clarkson's Farm," and this season, they're doing everything about making the farm tech, and how to make it produce more product.

And it makes me think about what we're doing in the security space. We are so focused on reporting vulnerabilities that we're not focused on making it easier to respond to them, and which ones should we respond to. So let's say out of that 400, and I didn't look at the report, let's say 50 of them are critical.

Then what we should be doing as consumers of that is looking to see if the version that we're running has less critical, and determine if we want to try to fix them ourselves and push that back into the community, because we're consumers of that code base. We just have to stop thinking about it in terms of the number, and we get so focused on the number, and it does freak us out, and it is scary. But when you really get right down to the brass tacks and you start looking at what is actually impacting you, those numbers decrease substantially, and we have to get in the habit of focusing on that as opposed to the total number, because the total number is overwhelming.

Yes. Jeff, I was going to toss this to you, and I'm wondering, should Linux kernel security teams reevaluate their policies of assigning formal CVEs to almost any stable fix, or does the responsibility lie on external security vendors and threat intelligence platforms to provide better risk-scoring frameworks? If you go back to what Jack mentioned, this is open source.

So it's not like a bit of proprietary software where you can go back to the vendor and said, "You gave me something that's bad. " So it's hard to pinpoint where that's going to be. I'd like to take us back naively to exactly one week ago, and the three of us were having a conversation about the National Clearinghouse, and we talked about, well, it's important to see it, but we need to be able to fix it.

Wasn't life wonderful last Friday? And now, one week later, we realize, holy cow, we have over 400 things we need to fix, and that's on top of everything else we were doing. So if a tree falls in the forest, you always need to patch it.

I don't really know. Obviously, prioritization has to be in here, but the bloat that Tracy was talking about, if you bought a boat, and the person from whom you bought the boat told you, "There's different types of woods used in the hull. " Your question is going to be, "Well, where are the oak planks now?

" And that goes to the, we don't know what's running in these, and how many things we're running that we don't actually need, but provide a backdoor. So, I don't think there's a regulatory perspective, John, that's going to say- Yeah ... " I think we need to completely change our perspective and, a year ago, we were talking about SBOMs.

When's the last time you heard the value of that? Boy, that would really go a long way to say with an SBOM, how many of these affect you? Exactly.

Well- Oh, go ahead, Tracy. There's a part of the de-bloating that SBOMs, and I haven't done SBOM, so I'm not sure when you look at the details what it covers, but I loved that you brought up the de-bloat issue because when you go and you do a default install of Linux, it dumps in a whole lot of stuff, and 95% of it you don't need to do what you want to do, but it fires it up and they run by default. And so one of the things I've got in the practice of doing with me and my teams is, I say build a minimal install with absolutely nothing in there, and then install the software package I need, and then only add what its minimum requirements are to run.

I don't need a GUI for this, and I don't need a GUI for that, and I don't need this. Hey, if there's no wireless interface on the box, we're not running Wi-Fi, I don't need that part of the package. And so much of what we do is we just accept what's default and just pile it on and pile it on because we're moving so quickly, and not- And- ...

taking security into consideration. It's a bad habit we've had, Jack, for a long time. I've been in the build management space for my entire career, and I hate to say it's sloppy and lazy, but that's what it is.

You don't take the time to understand what you're pulling into your distribution and understand what you're installing. You're going to get stuff you don't need, and now we need to start looking at changing those habits. That's what I'm saying.

Our habit is to count the vulnerabilities, get freaked out, instead of saying, "What are we doing? " But we have been, as a group, when it comes to builds of anything, we want it simple, we want it fast, so that makes us sloppy and lazy. Can I just- It's true.

Oh, yes. Just as a follow-up to what you just said, Tracy, you said earlier, so we see this sheer number. Maybe this is a naive question, but do we divide those into like here's a high risk versus a lower risk, or is this all lumped into one big number with no distinction between the level of security risk?

Well, there are so many discussions going on right now about how well they're being evaluated, and if they are critical or if they're high risk or if they're medium. But there is also an argument that even a low-risk vulnerability could create problems for you depending on where and how it's used. So if it is a low risk that's in every single application that you're running in your production environment, you probably want to fix it.

So again, it comes to how broad it's impacting you, and that is going to be unique for-- These are snowflakes, right? Everything we do is highly bespoke and customized. So it's the job of every single development team to start understanding better the packages that they're consuming, what vulnerabilities they have, and from an operational perspective, where are those packages running in our operational environments.

That is a defensive discussion. We're defending production from vulnerabilities that now have been reported when we already did our deployments. The offensive technique is find it before you ever roll it out.

But offensive alone is not going to win the game. You got to have both sides, and we do not talk about defense as much as we talk about offense. And offense says we're getting 400 shots on goal in two days.

Right? Yeah. And that's a lot of offensive work to do.

So we get distracted by we don't have a good defensive line, and that's- Sometimes- ... kind of how I see the problem. Yeah, and sometimes the snowflakes accumulate into a snowball, which kind of leads to panic.

And when people see numbers like that, they immediately overreact, and there is some nuance to it. But in any event, that was interesting to see what happens with Linux, and we're going to move on from there to something that's a little bit off the beaten track, and it involves Google and the EU, which is something we've all grown accustomed to. So the European Union has slapped Google with a $1 billion fine for anti-competitive behavior.

I think this is the first time Google has been penalized under European Bloc's Digital Markets Act, which was established a couple of years ago. So in this case, the regulators determined that Google violated digital market rules by favoring its own commercial properties, including shopping, flight, and hotel services at the top of search results, while burying their rival platforms. Additionally, the commission found Google unlawfully restricted app developers from directing users to cheap subscription offers outside the Google Play Store.

It is a significant amount of money, but just for perspective, the $1 billion penalty is just a fraction of Alphabet's massive financial footprint, which I believe is Reported $400 billion plus in 2025. And this isn't even the largest regulatory setback in Europe for Google. 5 billion EU antitrust fine over its Android mobile operating system.

Now I know there seems to be kind of an element of Groundhog's Day to all this, Tracy, with EU fining Google yet again, but I wonder if this one is different in light of Google's grand ambitions that we've talked about earlier this week to dominate the AI space and become the de facto destination for all content, among other things. I'm just wondering kind of what your take is on this. Well, I think it's a testament to how Google has dominated the marketplace, right?

Google is like, they own the shopping mall, and they operate the stores inside it, and they control the signs that tell customers where to shop. And that's what they got fined for, right? They got fined for using its marketplace to favor its own products, in essence.

I think that in that article that was sent out, they said it was divided into kind of two violations. One, Google favoring its own services and search, and shopping and travel results. And then the other was for restricting app developers from directing users to less expensive and outside Google Play Store.

To your point, Tracy, that's Google operating the mall, operating half the stores in the mall, operating all the advertising for the mall, and forcing every store to use Google cash registers and pay Google a tax for every transaction, even if they're not a Google store. Well, and they're the wholesalers that sell to the stores. Right, and the wholesalers that sell to the stores.

So they get a piece of the action no matter what happens. Exactly. So Google and Apple are kind of committing the same crime, right?

Right. So the bigger issue is not simply the size of the fine, it's whether companies that control a digital marketplace should be allowed to decide which products that we see and how developers reach their customers. Because that's a lot of control right there.

It is a lot, but is it really a crime? Because you use that term. Yeah, by the EU definition, within a jurisdiction, it qualifies.

But is it really a crime? Is Google a utility? I think that's the contention is a lot of people argue that because of Google's monopoly on the search business, it is de facto a utility, right?

Up until the last six months, maybe a year, the only thing you had in town was Google and search, right? Mm-hmm. I think AI is changing that game, and I hate to bring up the topic of AI this early in the conversation.

But AI is having a big impact on Google and the search world in general. You know what, Jack? I hardly ever do Google searches anymore.

I don't like the Play Store. I go right to ChatGPT and I start asking it, "Give me products. This is the kinds of things I'm looking for.

Give me things that are cheaper. " And it provides much better insights on what I should then go look at. So I've gotten away from using Google for buying things.

Right. And I've done the same. And Jack, you were 21 minutes in.

That's not too early. You checked the time. I wonder if we could ever do a show, just off topic, have we ever do a show where AI wasn't mentioned?

If we could try to figure out a way to do that. Is it achievable? I don't know.

100%. There are so many topics that we could cover that don't conclude AI. AI is not dominating our world right now.

It's just dominating our mind space. Right. That's an interesting point, actually.

I know it's the case in Silicon Valley, people just don't think straight anymore. Not that we ever did, but there's just kind of this preoccupation with inserting AI into every conversation or into every presumable product strategy or plan. And it's just interesting.

All of you, I guess, mentioned you don't use Google Search as much as you did before. Given this ruling, it's interesting to see where that's going to go, as well as this idea of what's going to happen to the App Store. Because I'm wondering what structural changes must major App Store operators make to comply with anti-steering rules, and how will allowing direct link outs impact the traditional platform fee revenue model used by the big tech firms like Apple and Google?

So I think there's two aspects to the App Store, or three, I guess, to separate. One is the walled garden aspect for security, right? So Apple originally, and now Google sort of was forced to.

You can't publish an app through the App Store, and in Apple's particular environment, you can't get something on an iOS device unless Apple looks at it and blesses it for safety, security, et cetera. So there is an argument to be made that there's value in that, and Apple actually charges developers to be listed in the App Store to cover that. The second aspect is the search.

How do you find the apps you want? And in which case, if you have this walled garden and you're participating inside the marketplace, then you can influence whether people see other's apps versus your app, and that's kind of sketchy from a free market perspective, right? You're warping the market artificially because you have that monopoly.

But the third aspect is also forcing people to go through you as a payment processor. And this was the big Spotify thing, and this is the half of the fine that Google had, which I can't remember the numbers, like 420 million euros, which was less than Apple's fine for the same thing. Apple was fined, I think, 500 million euros for the same thing, which is just because you're in the App Store, why should Apple get a cut of every dollar revenue that you make through it, right?

You're Amazon, you've got an entire- ... business built on charging people for books and digital content, why does Apple get a cut of that? " So there's that.

Obviously, there's a cut in that that goes back to the mall owner. And I'm willing to bet that none of the stores in that mall are there rent-free. They all had to pay rent to be there, to have a storefront, to have a presence.

I'm not advocating that we should just completely let monopolies take over everything, but I'm not certain that, and I'm surprised I'm saying this and not you, Jack. " I think that's a very valid point. It is a shopping mall.

It is a great analogy. Thank you, Tracy, for that. But yes, it's in the shopping mall analogy.

Yes, the stores, as I said, an app developer rents storefront by listing the app in the Play Store and whatever Apple calls it, right? Google Play Store, Apple. I'm not on the Apple ecosystem.

I don't know what they call it. But then again, like you said, but the thing is, they're not issuing store dollars and saying the store can only charge... Sorry.

They're not issuing shopping mall dollars and saying, "We're protecting consumers by not having cash and eliminating fraud, and so you have to use our store dollars," and everybody gets a cut of it. What they're basically saying is, "We don't care how somebody buys from you outside of the shopping mall. " And I think that's where people are.

That's sort of where the monopoly play comes in, is they just- That's the problem, that it feels like it... I guess we could say it is a monopoly. So if you only have one shopping mall to go to, should it be regulated to make it fairer for smaller companies, right?

Right. But the smaller company still has to make sure that the person walks through their door. Well, those- It's not going to solve everything.

Yes. How do regulators effectively balance antitrust enforcement under, say, the DMA with consumer convenience? I mean, Google's argument is that removing integrated search features like instant flight and hotel widgets degrades product usability for European users.

Well, and the argument is, that would be interesting if there were no alternatives, right? Mm-hmm. And the EU is basically saying, well, you have how many different...

I mean, there's what? 30 some odd different travel sites. Travelocity, Expedia, you know, I don't- Yeah ...

name them all, right? All of whom provide that exact same service, and all of whom are listed as apps in the App Store or are listed as websites in a Google search, right? And they pay Google for the top listing on the search, right?

You pay Google for ads, try to pay them enough, bid enough money to get top listing so that people select your site, right? " And what happens if Google says, or the shopping mall owner says the biggest signs are reserved for the shopping mall owner. And it doesn't matter how much you pay me, my sign is always going to be the first thing the customer sees when they walk in, and it's going to be the biggest thing that they see when they walk in.

That's not- Sounds like a trade show. Right? Sounds like a trade show.

That's exactly how we do it. Exactly. And that's another great...

You're wonderful with the analogies today, Tracy. You're on fire. Maybe we'll get in the business of suing all the trade show companies for the exact same thing.

Exactly. But, you know, I think that- Now I'm never going to get an invite ever again to any of these things. But the truth is that Google makes billions and billions of dollars.

I don't know if a billion-dollar fine is going to really impact them. It's probably just going to become a cost of business, and they'll continue doing what they do. And I'm guessing that American politics will start getting into this and complaining about it.

That was where I was going to go next. Yeah. I'm glad you brought that up.

I mean, I'm concerned- Yeah, they're going to say, you know, the EU is targeting American companies ... if you were going to say any foreign tariff threats or what? Yeah.

I mean, could this become geopolitical? Oh, yeah. I think it could.

But it's not that much money in the world of the high tech, right? It's not that much money in Google world. So- Yeah, but this could become geopolitical at the drop of a tariff.

Yes, it could. Yeah. Absolutely.

Yeah. And it's possible. I think from an international perspective, it can be geopolitical, but I think on the US perspective, I'm going to say that word again, AI is a much bigger and juicier target for politics than Google.

Oh, yeah. We should do a show on that. I mean, oh my God, all these lawmakers, they can completely disagree on everything, but if you mention AI, they see an opportunity to seize on it.

Yeah. Well, Anthropic just got hit with a billion-dollar fine today. Yeah.

Yeah, so there's- Was that the publishing, the publishers fine? Yes, the pub. Yeah.

We talked about that. Yes. 5 billion, over half a million books.

$3,000 a book. That's not a bad investment to make. No, it's not.

No, it moves to the common, it moves towards the ultimate goal of wanting to go public. And speaking- John, if we're going to do that show, we need one panelist to be a deepfake. I think we keep talking about this idea of having some AI-generated character or a deepfake appearing on the panel I think we should do it, maybe on April 1st or something, and see what happens, and play it straight.

You could do digital twins of all of us, and then just- Wow ... have the AIs talk to each other for a "Techstrong" show. Is the world ready for that?

Geez, that'd be crazy. Hey, we're going to wrap up. It's been very interesting.

We've got lots going on. And speaking of money, and lots of money, Samsung is in advanced talks to make a massive financial bet on Mistral AI in Europe. This is based in France.

This would be a major shift in how global semiconductor leaders compete for dominance in the AI hardware supply chain. So according to the "Financial Times," Samsung is negotiating an investment of up to a billion dollars. 8 billion.

It also, in a sense, highlights Samsung's aggressive push to secure influence over next generation hardware architecture as it seeks to outpace companies, rivals such as SK Hynix in the HBM market. And so rather than operating merely as an external memory vendor, an equity stake would presumably give Samsung direct influence over Mistral's long-term infrastructure decisions. Jeff, what do you make of this?

And I have a follow-up question after you tell me what you make of this. Okay. So first of all, we're staying in the EU, and we're staying with big amounts of money, right?

And I would offer, we are kind of talking about the same thing. Yes. Let's fast-forward four months, six months maybe.

Is the EU going to make the same sort of accusations with Samsung as they are with Google today, saying you are trying to corner the entire market by having a stronger influence on the chips that are going to be there, the software that's going to be there, the hardware infrastructure that's going to be there, the use of AI. There you go, Jack, we're deep into the show now. How is this different?

And by the way, how do you double the valuation of a startup to over $22 billion in six months? I've had some startups. Boy, did I miss that in how to make your startup work.

Yeah, I was blown away by that number. " That is a huge number. But continuing on my analogy day, this is- ...

it takes a village moment. That's what this is. Because it tells us that the model developers needs the chips, the memory, the data centers, and the distribution, and the semiconductor and device companies need the models that can create this demand.

And maybe at the end of the day, we're going to see smaller models running on smaller devices, and the AI market may not be another chat box, but it may be this intelligence built directly into the devices people already use. And I'm all about that. Right?

I've been talking about smaller models and more efficient ways to do this, as well as the agents for a while. Distributed AI has to be the future. I think this is the second week in a row that I've made that statement, only because take a look at all the other issues that haven't even come up yet today.

The big mega data centers that are being built or are being prevented from being built, and all the downstream effects of that, that is not a sustainable model. Whether you look at it from an environment or economics or chips, that's just not sustainable. It has to become distributed, and it's the only way it's going to work, it's the only way it's going to grow going forward.

And John, I know you had a follow-up question. Did we step on- Oh, yes. So yeah, you kind of answered it, but I'll rephrase it maybe.

So does a billion-euro investment from a South Korean tech conglomerate compromise Mistral's original positioning as a self-sustaining champion for European sovereign AI, or is this capital an inevitable necessity for non-US startups trying to build full-stack physical infrastructure? Part of the answer to that depends on what the investment is. And I didn't get a chance to dig into this, but if you look at a lot of the US AI investments, and we did talk about this a little bit yesterday on yesterday's show with regard to AMD and Anthropic and Microsoft.

We've seen a lot of what I call washing machine investments, where I invest a billion euros in you- Yes ... and then you pay me a billion dollars back, and then I give you chips for that, and my balance sheet goes up, and your balance sheet goes up, and everything looks hunky-dory. But there's a lot of funny money and a lot of hand washing going on here, and it's not clear what's going on from a pure financial investment perspective.

On the other hand, from an LLM perspective and a sovereignty in the EU building, or in particular the French building a French sovereign LLM, you can't build a large, and I think large is somewhat a misnomer. These are jugundously huge things these days. You can't do that without investing in hardware or having a partner who has invested in the hardware.

Mm-hmm. No matter what you're doing, if you're in the LLM business, if you're in the model training business, you're talking Scrooge McDuck swimming pools full of cash in order to do this. It is not a cheap endeavor.

This is not two guys in the back of a garage building an LLM anymore. That's not the way this world works. And if the EU wants that, they're going to have to come up with the money, and if the EU isn't going to come up with the money, Mistral's got to go somewhere else for it.

I think that's inevitable, and that growth is going to be there. Keep in mind, what we've been talking about, if you say, "Hey, I'm AI," you're going to attract investments. Some of them may be stupid, and I'm not saying this is or isn't, but AI attracts investments right now.

And Jack, what you just described in a different role that I've played before, sounded a whole lot like money laundering. I was going to say, this is like a money laundering kind of... I'm not accusing these companies of this, but you're right, Jack.

" Like, what's the actual dollar amount? But if there wasn't a contract behind this, if this was all done on a handshake and not with 1,000 lawyers involved, it would very easily be an interest to a RICO prosecutor. There is a whole lot of, I give you money, you give me money, our numbers go up.

I give you product, you give me stuff. It's very funny. Very racketeering, right.

But they still have to prove themselves. With a $22 billion valuation, they have to prove- Oh, yeah ... that they can grow beyond just the being a respected model developer company.

It has to show reoccurring revenue. It's going to have to defend that position, right? So- Yeah.

They're going to be- ... it's going to be a challenge for us to do that. Well, and the other question about that, I think is, does the market care that it's an LLM that is EU sovereign or data sovereign or trained on EU data only, or what is that?

Or does the market want the LLM to do something, right? $22 billion, I can't remember their total investment, but it's got to be in the order of tens of billions of dollars plus now the Samsung investment. For that, are you building something the market cares about?

Do you have product market fit, or is this just yet another LLM that's going to be somewhat in the game of second or third, or has been compared to some of the others? Yeah, and that's why, Jack, I was saying that they've got to show that the intelligence can be built directly into- Mm-hmm. Yes ...

the devices that we're using. And- And not just a large chat box, so to speak. Is the EU very concerned about where money goes but not caring about where it comes from?

No. The fact that these two articles are based in the EU are almost contradictory to each other. Yeah, they are.

Absolutely. That's a great point. It is.

These are two trains on the same track that eventually are going to collide. Absolutely. And does that mean it's a bubble?

Probably. It's going to be really interesting when these companies get close to their IPO dates, and I'm thinking of OpenAI and Anthropic, and just all the money that's going into this building out, and how much revenue's going to come in. And boy, is that going to be anxiety days for some of those companies.

In the back of my mind, I'm still hearing Alan Greenspan talking about irrational exuberance. Yeah. Mm-hmm.

Yes. Exactly. That's a good way to end it, actually, I think.

Well put. Jack Palmer, Tracy Reagan, Jeff Rich, thank you so much as always. The dream team on Friday.

You guys deliver, and you know how to take in conversation and go through all your opinions. I think it's admirable, and I appreciate it. Hey, thanks again for watching this week on the show.

There's lots of great content on Tektron TV that's coming up afterwards. We want to wish you a great weekend, and happy summer. And for us, we're out.