NVIDIA Investigations, Anthropic IPO and Microsoft AI Stack
The NVIDIA SEC probe, Anthropic’s Nasdaq debut and Microsoft’s push for a transparent AI stack lead the November 18 discussion on Techstrong Gang. Mike Vizard and Alan Shimel host Jon Swartz, Andi Mann and Elizabeth Safran. The panel looks at how investors and enterprise buyers now judge the biggest names in AI.
NVIDIA Investigations
NVIDIA beat third-quarter expectations with $115.8 billion in revenue. However, the results arrived alongside reports of an SEC civil investigation into its investor disclosures. A leadership shake-up following Jensen Huang’s exit adds more uncertainty. Coverage details how NVIDIA delivered solid earnings amid an SEC investigation and leadership turmoil.
The panel separates the reported financials from allegations that remain unresolved. It also considers what the NVIDIA SEC probe could mean for customers who depend on its chips.
Anthropic Anticipation
Anthropic is preparing for a Nasdaq listing. The offering could value the company at up to about $1.4 trillion after revenue grew more than tenfold. Reporting explains how Wall Street weighs Anthropic’s explosive growth against a $900 billion compute bet.
That growth comes with roughly $895 billion in multi-year compute commitments and steep losses. The discussion weighs whether public investors will reward durable demand or treat the listing as a bet on the AI cycle itself.
Microsoft AI Muscle
At Ignite 2026, Microsoft announced a Transparency Initiative aligned with NIST AI standards. The company also pitched Copilot as an operating system for work. Coverage outlines how Microsoft is pressing its case for a transparent AI stack.
The panel examines whether a single, more transparent stack gives organizations more control. It also asks whether that approach deepens dependence on one platform.
Together, these stories show that growth alone no longer settles the AI debate. Governance, disclosure and transparency now shape how buyers and investors respond. Watch the full conversation and catch new episodes every weekday at noon ET on Techstrong TV.
Transcript
Hey everyone, happy Wednesday and welcome to Techstrong gang. I know what you're saying. " It is.
I'm not in the studio today. I'm actually at my home office kind of playing hooky. This is the only video I'm doing all day, so I stayed out of the studio office and then from home.
Now I look like everyone else on this thing. I don't have my fancy background or anything. I feel like I'm now dressed as Clark Kent instead of Superman.
But we've got some supermen and women on the gang with us today. Let me introduce you to, well, he's over in Spain still, our one and only Caballero, Andy Mann. I almost said Andy, like with a Spanish accent.
Andy, good to see you. She was here a moment ago. Don't go blaming him.
No, never. She was here a moment ago, and I hope she'll be joining us back. She's coming back.
Yeah, she'll be back. From Brooklyn. Just- Liz Saffron ...
a quick- Of course, our Mr. West Coast and Mr. East Coast.
Mm-hmm. John Schwartz and Mike Vizard. I feel like you guys should be batting third and fourth in the Yankees lineup with that kind of gab, right?
Yeah. Speaking of the Yankees lineup, we got trouble next year. But anyway, guys, welcome.
Liz, I'm glad you're back. Yeah. So guys, it's a busy day.
We've got a huge item. Yeah. Yogen and Price, every network we know seems to be going down at one point or another, but not to worry, it's not cyber attacks.
It's just a crazy time. Mike, there's so much to pick from. What do you lead off with?
Well, I would like to welcome you to the latest episode of As the Nvidia World Turns. Mm-hmm. It seems like the CEO wants back in.
Sales to China are out, and the SEC is investigating perhaps, and investors are not happy. But John, walk us through what's going on here. Okay, so yesterday, it was Tuesday.
It's hard to keep track of the days with Nvidia. It all seems to be a blur. Yesterday, hours before they announced their results, which were actually fairly decent, they were fairly good actually, but hours before they announced the results, the Wall Street Journal reported that there's going to be an SEC investigation, a civil investigation into the company, and whether it misled its investors.
Around the same time this happened, Jensen Huang, the disgrace, and I'm going to say the word disgraced, former CEO of Nvidia, literally groveled to get his job back in a note statement to the board. Now, during the results, the board basically said he's gone. The chairman of the board, Harvey Jones, said he's not coming back.
We don't want him back. The interim CEO, who was also the CFO, reported the results, and during the results, I will not go down the item list of the results, but they beat expectations for the third quarter, and more importantly, they raised their implied calendar 2027 revenue forecast to $740 billion, up from $700 billion. Whether we can believe them is another matter, but they did nonetheless, and that did get thumbs up from Futurum's Ralph Polk, who I have a great deal of respect for as an analyst, among others.
What is interesting is that the interim CEO, Ms. Kress, avoided mentioning Jensen Huang by name, and she acknowledged the diverse opinions around bringing him back, but again, that's a non-starter. Now, to complicate matters or maybe to make things even more interesting in this soap opera, Kress mentioned that Nvidia's going to cease selling the H200 chips in China, which were being sold through a trade loophole just days after the departure of Jensen Huang.
There are going to be congressional hearings. God knows what else. I also saw another report that Ellis Huang, the paramour of Jensen Huang involved in this alleged spy scandal, had been questioned during the summer by US investigators and authorities.
To me, it's kind of like a combination of a different version of Enron, Watergate, Teapot Dome, throw in whatever cliche you have or whatever controversy you have, and we're just at the beginning of this. It's going to get worse. We're going to find out, unfortunately, and I'll leave it up to you all because I know you all have opinions about this.
This thing is going to become a monumental mess. So you're telling me I should go get popcorn and settle in, right? Totally.
It's going to be- I don't think it's Teapot Dome or anything. " Or maybe nighttime soap opera. " Yeah, totally.
Something like that. Little bit of a Bond. Little bit of a spy thriller.
Who shot Jensen? Yeah. Oh, God.
But John, I got to tell you the truth. I got to give you real bonus points. What did you call Ellis again?
His paramour? Mm-hmm. Oh, Ellis, yeah.
That's a fancy word. Alleged paramour. That's a fancy word.
Oh, no, he admitted to it. That's true, he did. He did admit it.
Yeah, he did admit it. No, he admitted to it. The alleged spy.
He said it was a piece of something. But paramour- He said she was the only one, and then that- ... is a fancy word, John.
That's why you're an author. You can make up for fancy words. Oh, yeah, so are you.
Amazing. Yeah, so are you. " And I'm wondering, is that a function of demand is so high and supply is so limited, or will there be pressure to reduce the pricing of GPUs?
Are you saying- No, there'll be no pressure. The only pressure to reduce pricing would be if there was a competitor who made a comparable GPU, and there's not right now. Not for the foreseeable- And I have to say, can I add something really quick?
I'm sorry. One thing I should've also mentioned among all this mess is the Singapore element. That a couple years ago, Nvidia stopped reporting their Singapore-related sales.
Mm-hmm. Yes, they did. Set off a lot of alarms, by the way.
That's it. That's in the report I did about China GPUs. But let me just clear the air on one other thing.
Let's not, for a second, give people reason to doubt that the results that Nvidia reported were not true. I don't think they fudged their books over there. You could talk about Jensen doing a lot of things.
No. Their revenue is their revenue and their profits and so forth, and they're amazing. I'm not referring to what they did do with this, but their forward-looking statements are...
I don't doubt- Well, let me ask you a question, John. Had Jensen made those forward-looking, make believe the last two weeks didn't happen. Mm-hmm.
Jensen reported those results and gave you his outlook. Would you doubt him? I wouldn't, no.
No. There you go. So here's the part that I'm not saying there's an alternative, a GPU for GPU alternative, but if you look over at what Amazon, Google, and Microsoft are up to, they are all positioning some alternative class of processors as an alternative to GPUs.
And they've only started that motion aggressively in the last, say, yeah, three to six months. So, I'm going to toss this to Andy, but am I going to see more pressure on GPU pricing? Because there are options, and it seems like other folks are going out of their way to create distribution networks for those options.
Yeah, look, eventually you are. It's going to take a while for this R&D work to show up in the marketplace. But it's very clear that GPU prices have grossly overinflated.
You look at the price versus just a year ago, and it's like 100% up, which incidentally was the same revenue growth that Nvidia enjoyed in the last year on year. But you look at what happened with Apple and the M1, 2, 3, 4 chip. They were able to build their own over time.
They moved away from that lock with Intel that they had, and that actually gave them separation and differentiation in the marketplace as well. Those are stunningly fast chips. And so look, there's a big incentive ...
for these competitors to come out with better chips, faster chips, take new techniques, new models. They'll be able to separate. A lot of AI is like, "Well, do I use Anthropic?
" Whatever. They're both going to hallucinate at me. There's opportunity, though, for them to have very different models and different ways of computing.
And so there's more than just the price incentive. I'm a little bit with Alan. As long as there's one supplier, we're not going to get cheap chips.
I'm a gamer, and I've just recently moved to Spain. Do you think I can afford to buy a gaming PC here? You've got to be kidding.
Sam's hoarding all the chips. Mm-hmm. All the gaming boards.
So there's obviously demand. And as long as there's demand, there's high prices. But if we can increase supply, and especially supply of different types of chip for different use cases, different purposes, that starts to rewrite the register.
Right. I think a lot of people can't afford to use GPUs even for AI, so they're going to be looking at alternatives. It's just the cost is too high.
Yeah, absolutely. I think it depends when you say using it for AI. If you're talking about inference, yes.
You've got Google, you've got Broadcom, you've got Amazon. There are plenty of inference chips out there that are viable. If you're talking about training models, however, they're the only game in town and continue to be such.
Amazon will tell you that they are using their other classes of processors for training as well. So, we'll- Yeah. Good luck with that ...
we'll see. I saw some of the Amazon models up on the top 10 leaderboard there in OpenRouter. No.
Not. Not. I don't think that they're saying that they need- There's a reason why all these Chinese companies were going through the machinations they were going through.
There's a word like Paramount, John, going through the machinations- Machinations is a good word ... they were going through to get a hold of these GPUs so they could train their models. Good enough beats best every time.
That's right. Part of the reason I have doubts about some of the numbers associated with Nvidia, maybe it has to do with the circular financing conundrum that was going on, and something that we talked about before Gate and got a lot of coverage, right? So I- But they're on the other side of the circular financing.
People are using their dollars as the backstop, almost as the currency of this whole AI market. Mm-hmm. But, again, if you go back to what Jensen said, I think at the last quarterly reporting, basically, they've already sold all the chips they can make next year.
Mm-hmm. So maybe it'll be propped up for a while, but one thing that I wanted to ask John about is that when you were talking about the creative accounting with Singapore and how they didn't book that revenue by region the way they were. Yeah.
So if Singapore, the Biden administration put export controls, that rationed back some revenue to China, then all of a sudden Singapore exploded. Exploded as a customer. Their revenues in Singapore, I don't have the numbers at my fingertips, but it was like 20% of their revenues in one quarter came from Singapore.
So- That's correct. Clearly, the whole Megaspeed, are they renting it to Chinese companies is probably true. And then what happens if that revenue was cut off with their growth because maybe Jensen is counting, he was counting on his relationship with Trump to loosen export controls, and he was counting on the revenue that he was making in Singapore to keep those numbers as high as they were.
That's a pretty- I find that's painful. That's an interesting theory. I think you're onto something there.
Okay. Can I say something in plain English so we can all make sure that we understand what we're talking about here? Nvidia is taking money that it has and handing it over to, quote unquote, customers who are willing to pay a premium to pay the highest price possible for GPUs out there that are supplied by Nvidia, which keeps the price of GPUs artificially high.
Yes, no? Yeah. It's a scarcity game.
This is a scarcity game. But it's an artificially created scarcity game, maybe. No, it's not artificial because let's say Nvidia said to, what's the Neocloud that- Megaspeed.
No, that's the second one. CoreWeave. CoreWeave.
" CoreWeave does not have the wherewithal to build that center. Yes. That's why it's important to follow what happens to CoreWeave in the future.
Right. CoreWeave does not have another market they could buy AMD GPUs. Right.
Maybe not as good. But that means another Neocloud didn't have access to those GPUs. No, they're backstopping all of these GPUs.
Well, not everybody's on the Nvidia- See, I just think there's so many factors that could screw up this merry-go-round, right? Like- Let me make another distinction, though. The thing with Megaspeed.
Look, Asia is rife. There's another good word for you, John. Asia is rife with data centers that are renting compute to Chinese companies.
That in and of itself, under the laws of the US, as of time they were doing it is perfectly legal. Yes. No need for an SEC or congressional investigation.
The allegation on Alice Wang and Megaspeed and her husband, the Griffiths here, is that they were buying chips as Megaspeed and perhaps taking them back into China. Diverting them, right. And that is illegal.
Yes. And that will get someone in trouble. Though, from what I understand, Alice is now living- In China She went to Beijing, I believe.
Alice is her stage name, like Tiffany and Brandy. Yeah. It's just four names, isn't it?
It's just like aliases, just a couple aliases. Yeah. Whatever her real name is, Wang Li or something.
Wang Li. Mm-hmm. Are you saying that that's not usually unusual in China, in Asian circles anyway?
Oh, no, they have a Western name. An Anglo name is pretty normal. Yeah.
But she's back in China, so she's immune from prosecution or at least extradition there. That's true. You said there's a legal loophole about renting computers or something.
Well, now supposedly the Trump- They're going to close that, right? That was the loophole ... They're closing it.
But we discussed this all week. I'm sorry, Mike. Are you saying that GPUs are falling off a truck in China somewhere?
Haven't that happened three months ago? I think Alice and her husband have a thriving GPU store somewhere in the backstreets of Beijing. Right next to the bats in the market where the honey comes from.
Andy, if you're going to get that gaming PC, I think you need to make a trip. There you go. The road to Beijing.
Yeah. But again, I almost felt bad for the interim CEO at Nvidia. She reported stellar numbers with amazing projections, and everyone's crapping on it.
Yeah. What is super ironic to me is that there were these accounts of Jensen Huang belittling her and calling her a deer in the headlights. Yeah, I heard about that.
Right? And it's just completely abusive behavior because she said this idea of circular financing may be a bit of a problem. She was trying to be the adult in the room, and he castigated her.
So, as we unpeel this onion, it's just an ugly business the more you look into it. The culture, the way they strong-arm their suppliers. It's the people, John.
We've said this before. The tech bro billionaires that are running the AI industry, and it's not just Nvidia, it goes across the board. Agreed.
It goes to what you're seeing in some of the VCs, and that whole tech bro culture out of Silicon Valley are the most dislikable. " I've mentioned the show before. I've told you to go check it out.
You really should watch "Cupertino" on Paramount or CBS. I think it's every Tuesday. They've got amazing storylines in there.
Mm. They depict these tech bro billionaires just These are some of the most venal people in business. I'm going to be honest about it.
Here's my prediction. Sometime in 2027, Nvidia's going to write a really large check to the United States government without admitting any guilt, but they're going to settle something. To the government or to individuals?
Or a certain person. Oh, they've already done that. Probably a donation to a ballroom.
Yeah. They've already done that, Andy. They already did.
5 million there ... the thing too, though, is this relationship between Trump and Jensen Huang. Trump didn't even know what Nvidia was a year ago, or maybe even nine months ago.
He'd never heard of it. He never heard of this guy, and then- Well, they do have a lot in common ... " We are over time.
Sorry. Too bad. I just want to ask one more quick question.
John, how was the CFO? How was she on the call? Did she present well?
She's well respected. I don't know why she's been given such harsh reviews. She's trying to be the responsible one and trying to keep this stuff together.
And I'm sure there's going to be a hemorrhaging of talent because the cult and the cult leader's gone, so some of his loyal subjects are going to jump ship. All right. Well, I think we're going to leave this here because I don't think anybody's shouting, "Lock them up," just yet, but- No, I think the real question is what's Calci in the prediction market saying about Jensen coming back?
I don't know. We'll see. Man, that would be something to behold and maybe not good for a president.
Hey, Sam did it. Sam did it. Yeah, I don't think Sam- But this was a more responsible board.
Sam, this is different though. There was no national security implications with that. Yeah, there was no national, right.
Exactly. Right. I don't believe Sam necessarily could be accused of sleeping with subordinates, lying to the board.
Sam's just an a*****e. That's it. Yeah.
Deceitful, untrustworthy, manipulative, and Machiavellian. That's the quote of the year on this show. There you go.
It's not accurate, so. Hey, it wasn't me, it was him. It was him.
It wasn't me, Mrs. Goldberg. It was you.
All right. All right. What else?
Mike, what have we got next before we go get ourselves locked up or something here? There you go. All right.
Well, so coming up to our next topic, which is Anthropic, which is I guess, IPO is going to go live tomorrow morning? Is that the general expectation? Tomorrow.
And everybody's kind of watching this maybe on pins and needles to see which way the market goes after this IPO gets launched. And John, I don't know, what is the number here again? What are they looking at?
It's changing. 6 trillion. So Liz, here's what I want to ask you.
Is this launch and this IPO going to set the standard for the rest of all these companies out there that are AI driven, and is this going to be the thing that we all watch to see where this lands? Is this the bellwether? I think it's a short-term bellwether.
So I think everybody is looking at this, especially in light of what's going on with Nvidia because the two companies are so intertwined financially. So I think it's going to be fine, right? Who knows what the actual number will be, but I think long term, they've got a lot of debt and a lot can change with this administration.
So I think it's going to go as well as they want it to go. I think that nothing's going to change. But I think it's very short-term thinking.
And they also have a lot of debt, and they burn through a lot of cash. So it's the whole, when John was talking about the adult in the room, this whole boom is predicated on future revenues and future growth and as much money as they are making, and they're raking in the cash. I am well aware of that.
What is going out is so much more. It's like when the music stops, what happens? It's like this game of AI musical chairs and I'm worried that when the music stops, the whole thing's going to fall apart.
All the chairs are going to fall over, not one person out, but everybody. So to your point about future revenue, just to be real brief, so Futurum Group, again, Ralph vaulted a projection into some estimates, and he thinks that they could scale their annual revenue. And I'm not an apologist for Anthropic, but I'm just going to give them their due.
He thinks that Anthropic will scale its annual revenue six to nine-fold to between 450 billion and 750 billion by 2030, which is kind of in line with some of the higher end projections. So it could happen. I think about Amazon, I think about this.
But in the short term, a lot of pain probably. So here's the part that I would worry about if I was them, and Andy, I'm going to toss this to you. But in the last 45 days, everybody I have talked to about AI is squarely focused on one thing, reducing their costs.
And much of that reduction cost is like, we got to figure out how to limit how much we're using the frontier models for and find some other ways to do this stuff. And that takes a little while for them to get there, but when I look at this stuff, I'm going, there's too much noise in the system where people are going, I got to find another way to do this. Yeah, look, that's a real problem because the cost savings aren't necessarily all there.
The spending absolutely is. To your point, they are making money hand over fist, Liz. But the enterprise CIOs, CTOs, just the token cost, the spend, the problem with the productivity is absolutely there.
I'm seeing it in my team absolutely. And we're getting incredible multiples in terms of our FTE equivalents by using AI to help and assist our developers in doing their job. We're actually getting pretty good results in quality as well.
We're actually using AI to help us make sure our quality's going up. There are definitely positive results, but the spend for a lot of these organizations is massively out of whack for what they're getting back. And so yeah, we're going to see not just the potential for the AI companies to not make that huge revenue, but we're also seeing the potential for the use of AI to start scaling back within the enterprises that are using it.
We saw that just recently. I can't remember, one of the big gaming companies is actually putting an upper limit on the advanced models that their people can use and how long they can use them for. And so they're doing that purely because of cost control.
So this is yet to play out. I don't know if we're going to make money or lose money with this. We're certainly doing a lot more stuff, but wow, is it costing us a lot of money.
There's a line among others, speaking about commitments. I think Anthropic has $900 billion in multi-year compute and infrastructure commitments among just Google, Amazon, Microsoft, and SpaceX. It's kind of staggering.
Which I totally understand because they need to make use of a frontier model for the things that they do. But 80% of the world doesn't need that. No, Mike, you got this backwards.
Anthropic is committed to paying those companies. Yeah. Regardless of anything.
So they're on the hook for that $900 billion. Well, they may be customers. That 900 billion number is for data center usage basically.
Right. But I would argue full circle that it's the big tech companies that are using those frontier models. So this may be another example of circular spending and financing.
But when I come back to the average enterprise, I'm going, you know what? 20% of what they may need to do requires these more advanced frontier models. Eventually.
Right. So I got two words for you. Not plastics.
Indispensability trap. Oh. That's also a great title for a book.
There it is. It's a whole book. Yeah, exactly.
And here we are. There it is. Well, let me do my Wheel of Fortune imitation.
Absolutely. I say check. But no, but seriously, this is exactly what the book is about.
Mm-hmm. There's not a doubt in my mind that this is going to be one of the most, if not the most successful IPO of all time. I haven't seen this much buzz.
There's a bigger buzz than SpaceX. This takes me back to Netscape and GeoCities. I was going to say, that's a great comparison.
Yes. Right? John, you were around.
We were around. I wrote about that. And you know what happened with that IPO?
And Liz, I'm not sure if you're old enough to be around when that was happening or you're too young. You're fine. You're very kind, John.
Yeah. We were all so young back then, but- Yeah ... I think what you said about what you think is going to happen to Anthropic happened to Netscape.
There was that- Exactly ... initial incredible jump, and then things settled down very quickly. Poor Marc Andreessen never made a dollar, huh?
Yeah, or Jim Barksdale or whoever it was. Yeah. But okay, maybe he did, but what happened with the company, right?
And that's the whole point of the indispensability trap. They don't become the enduring wealth. But here's the point, and I'm just about to put an article up, or submitted to put up on Techstrong AI.
Talk to me in six to nine months Right. Now, to be fair though, if you look at Anthropic and you look at OpenAI, and as Mike said, one might be run by an a*****e, the other one not, or they might both be, who knows? No, they're both kooky.
Yeah. Wait, I'm in the all of the above box. Okay, check them off.
All of the above. But here's the thing, Anthropic, I think more than OpenAI, has displayed an inclination and a direction of moving up stack. They know they're not going to justify these numbers just being a frontier model.
They have got to specialize into coding partners, testing partners, lawyer, law partners- Right, they're moving to Lascaux ... right? They are trying to move up stack into applications.
Oh, it's worse than that. Be very afraid. They're moving into verticals.
They're using the metadata and the workflows that they have been observing, not your data because we don't train models with your data. Yeah. But we watch everything else, and we are now launching applications into all kinds of interesting new verticals.
Yeah, they do this almost every few weeks. Every few weeks they do a vertical announcement. Mm.
But would people trust them? If they're going everywhere, first of all, they're going in all these different verticals at the same time, how well are they going to be able to penetrate? But would you trust Anthropic for your cybersecurity and not use your other vendor?
I just don't understand how- I think Mythos is the standard for AI-based scanners. No, I- Yeah, scanners ... I wish there was an alternative on the trust metric that mattered, but there isn't.
They're all trustworthy. And let me just tell you, a couple trillion dollars goes a long way. Jensen Huang, he did okay.
Yeah, I guess he could buy it, yeah. All right, well, Liz brought up the trust question, and we're at the bottom end of the 15-minute segment, so maybe it's time to move on because it's a lovely segue. Sure.
All right. So Microsoft had its Ignite conference, and one of the things they were talking about is that they can be trusted, and they are embracing this trust transparency framework along with, I guess, 39 other companies. It's not quite ready, but it's all about providing transparency into your AI workflows.
And I've got to be honest, I watched the keynote and I saw all the architectures and the platforms, and this stack that they are building, it's elegant, man. Everything is tightly integrated, and everything is designed to drive down costs, including a pair of AI models for coding that do 98% of what Anthropic does for 40% of the cost. Hmm, I wonder what that's about.
But here's the thing, and it's going to be the same play out over and over again. There is almost net zero in that Microsoft stack that is their idea. They basically have adopted and co-opted every other idea they ever saw in the AI space, reverse engineered it, and integrated into this lovely, seamless stack.
Well, so Mike, you're missing one thing. Wait, let me finish. Go ahead.
This thing is a work of art, but here's where it gets interesting. In my mind, the ability to reverse engineer somebody else's stuff is dropping to near zero. So the days when you could outrun Microsoft are probably getting lower and lower because their ability to reverse engineer stuff and put it into a more elegant stack is almost instantaneous now.
So this world is changing. Yep. Well, you missed the big picture, and it is really a Microsoft...
It's not reverse engineered. They've played this game before. Anybody remember trustworthy computing?
Oh my God. Okay? 0.
Bill Gates isn't here to write the memo. No. But that- When you said they played this game before, I thought you were talking about something entirely different.
This is Microsoft's MO. They've been doing it- Yeah, no, I was thinking the same thing, yeah ... forever.
They didn't create word processors. They didn't create spreadsheets. What I'm driving at is this is their MO, and the board is tilting further in their direction because they can pretty much, Alan doesn't like the phrase reverse engineering.
Sorry, they can capture your, quote-unquote, intent and create a duplicate of your intent in a couple of hours. Look at this stack. It is so well advanced.
" Why? AWS doesn't do that. They do it.
They replicate for 20%. Yeah. Google doesn't do that.
Ah, now you come to the point. So they all do that. What happens to the little guys in this industry and the innovative startups?
This isn't a little guy's game. Yeah, they do other stuff. There aren't little guys, yeah.
The little guys do the innovation and then Microsoft- They get bought. Microsoft does, yeah. They get acquired ...
and these others make a bill versus buy. Why would I buy them when I can just reverse... Excuse me.
Because reverse engineer may get you 80% of the way there and not 100%. So then I go, what is that phrase, where I just hire the people but I don't- Acquire ... acquire.
I'm not here to bury Microsoft, nor am I here to praise them. Yeah. I'm on neither side of it.
I'm just marveling at what they- But let me just say this, you've got to give them credit. Of all the companies we just mentioned, AWS, Google, Meta, the Mag 7, Oracle, they're the only ones that in spite of these big investments in copying everyone else's stuff, and building out data centers, and doing everything they're doing, they're still wildly cash flow positive. Hmm.
When Google is cash flow negative, when Amazon is cash flow negative, Microsoft's printing money, or at least profits. And if you look back in time, every time we go through one of these upheavals, everybody says, "Oh, Microsoft's doomed," and at the end of the upheaval, the only one standing up is Microsoft. Yeah.
I don't see why it's bad, though, that they're copying other people. Van Halen did a great cover of the Kinks song, kind of made it theirs, right? So what's the difference if they're using somebody else's playbook and they're still executing?
So- There's nothing illegal about it. It's perfectly legal. Unless you were one of those small companies that got copied.
Unless you're one of the companies impacted by that, right. Yeah. Well, think about these startups or smaller companies, they're pretty much screwed unless they're acquired, right?
They can't go public. They're not going to be acquired. They're just going to be steamrolled by these other guys, and that's a trend that's going to happen, not with Microsoft, but with Amazon and the others.
I'm really feeling bad for those venture capitalists on Sand Hill Road out by you. Oh, yeah. Terrible.
Speaking of a******s... Oh, sorry, go ahead. Did someone say Marc Andreessen?
He's number one on the hit list, yeah. Now, that said, though, it is cause for concern on this level, right? So this industry is built on the backs of all these little startup companies that kind of, if it wasn't for them, we wouldn't be anywhere near where we are today because that's where the innovation engine has been driven.
And what happens over time is when they disappear, the big companies don't innovate as fast because they don't feel the threat. Right. But then there's always a new crop of new companies coming up that do innovate.
And look, this is not an AI specific thing. This is the way tech works. It's the way it's worked for, I think, our entire careers.
Mm-hmm. One could say the same thing about Elon. Right.
He didn't found Tesla. So maybe we go back to the days of the two guys in a garage with Hewlett and Packard, and they're changing the world, and we don't need $50 billion from a VC just to launch a company. So that is an issue, though, Mike, and you brought it up, right?
It could just be Woz and Steve, right? Mm-hmm. But I do think that's where things have changed.
This is a deep pocket game. To play the AI game right now, it's a deep pocket game. What I think is going to happen, and what I think in the indispensability trap I write about happening, is once this $8 trillion or whatever we're going to bury into this infrastructure is in the ground, and we can start building on its bones, you will have fresh innovation.
You'll have new companies that don't need a trillion-dollar IPO to turn a trick. Right? You'll have innovation.
But until you have that fabric laid down- It's a big boys game. It's a trillion-dollar ante, basically, to play this game. How long before you think we get to that?
Because is that a year out, two years out? Where is it? Look- Utility ...
I can only go by my own life experience. And so I grew up, business-wise, during the dot com fiber optic. And again, it's a whole section of the book.
How long did it take for Global Crossing, MCI WorldCom, Level 3 to lay their fiber, go out of business, and then how many years was it till AWS and Google and SaaS and everything built on top of that fiber, the cloud itself, take over? The bust was late 2000 going into 2001. By 9/11, the boom was over.
But cloud, if I remember correctly, is around 2006. So I will say- Five years ... I will give Microsoft props for one other thing that they were talking about, which was essentially AI fatigue.
And what they were pointing at is people don't want to build and maintain their agents, and they'd just as soon buy it and have it as a capability, and it does something, and that's great. But they were basically saying it's pretty clear from the level of experimentation that's been going on for the last couple of years that maybe the average organization isn't going to be able to harness all these different AI agents and manage them and keep them from squabbling. So best to leave this to somebody who has good hands, like Microsoft.
I don't disagree with that either. I think one of the interesting things that we'll see if it holds water or not is the do it yourself agent crowd, so you can conjure up your own agents. I think most people, it's going to be too easy to buy an agent or download an agent, install an agent.
The real issue is how I manage them all. And from an enterprise perspective, never underestimate Microsoft to start with, but never underestimate the value of having a Microsoft in your environment that will do these things for you. We saw, for example, with cloud computing, Google's always struggled a little bit against, everyone struggled against Amazon, but Google's always struggled a little bit because it's not very enterprise friendly.
One thing that Microsoft's always had in its favor, it knows how to do business. It knows how to support enterprises. It knows about departmental feuds and budget cycles and support areas, things like this.
It knows how to do business with large enterprises. And so for an enterprise buyer who wants to use an agent, have a new agent, who wants it all in one, this is absolutely going to have legs because this is the same Microsoft they've always worked with. It's not necessarily the best vendor in any given category.
Word is a piece of bloatware for 100 years now. Certainly is. Especially if their stuff is good enough that they can undercut on cost.
Got it. I'd just be damned we'll get over that. Right.
No, but let me- But you know what? It's true. Hey, speaking of reverse engineering, I read an article this morning.
Anthropic is coming out with their own Office suite. This is true. And you can find that- Talk about moving up stack ...
ai as well. But here's my prediction for tomorrow. Tomorrow, you're going to see Microsoft extend Anthropic a very nice backhanded compliment.
" Right. You know what? We'll take the baton and run away.
So I'll leave you with this, my final thought on it. In every generation of new innovation, there are a handful of companies, as Brad Feld used to tell me, If you're not in the top three, get the hell out. There are a handful of companies that reach the promised land, that reach escape velocity, that become the next companies to take on.
There was a Microsoft before there was a Google. There was a Google before there was a Meta, Facebook. There was a Meta, Facebook before there was- TikTok ...
TikToks. There will be an Anthropic, and is probably better suited than anyone else, companies that reach that pantheon of the gods, of the tech gods. Will it be Anthropic and OpenAI from this generation?
Will it be Perplexity? Will it be someone we don't even know yet? " Well, who knew?
Who knew it was Google? Wait. So you mean that sound I'm hearing is the sputtering of all the engines of these AI companies that are about to flame out?
Is that what I'm hearing? They might. Well, or they'll get...
You know what? They don't flame out. They get absorbed.
They're like old soldiers in Douglas MacArthur. They just fade away. Yeah, I was thinking the same thing.
And there's already so many tentacles across the fabric of these companies that getting absorbed might not be that hard. Make no mistake, though. Look across the big guys.
Their stacks are already built. There's no more room for the startups unless they got something really cool and new. But a lot of the existing guys, the ship already sailed.
But that's what makes startups so exciting, because it's like a giant Petri dish, and yet they will come up with something exciting. And I guarantee you- But I hear it's harder than ever, though, now. I think that's what Mike is getting at.
Well, there's a higher bar of entry. You need more work, you need more resources than- It's a high bar ... the bigger guys can mimic what you do and maybe- But I think the big guys, rather, do they want to put in the investment resources to solve a problem that a startup might be doing?
They'll just have a startup be like their R&D lab when they get to the traction point. Well, they'll just reverse engineer what that company's doing and get it out the door within a couple of months. Right now, there are some people working in their garages who have some new ideas on using AI in a way we haven't thought of before who are going to be the next crop of Microsofts, Metas, and Anthropics.
I guarantee you that because that's how this world works. No, they're not. They're working on quantum computing, by God.
Well, maybe it'll be quantum computing and AI or cold fusion or whatever it is. But I'm telling you, this isn't the end. It's the beginning of the next chapter.
I agree. All right. Great show today.
I was thinking that chapter's called quantum, but what do I know? Hey, Alan, what was that- We'll be talking about it in a different dimension. Yeah ...
show with Groucho Marx where they had the magic word? You Bet Your Life. You Bet Your Life, right?
Yeah. That felt like that a little bit today with the magic words. We definitely did have some words in there.
And look, it's going to be interesting. I actually put up a good article on Techstrong IT today about robotics. Yeah.
What a great invention the human hand is. And look for robotics combined with AI. Yep.
Integration. 100%. It's all about the development of the human hand and the fingers right now.
That's their- Anyway, we got to pull the plug on it. It's 12:45. Guys, what a great conversation.
What a great gang today. I hope you guys all enjoyed the talk as much as we did ourselves. I think sometimes we like to hear ourselves more than, I don't know, other people.
But hey, we're live every Monday to Friday at noon Eastern Time here on the Techstrong network. tv, Techstrong TV YouTube channel, or our OTT app for Android, iPhone, Roku, Apple TV, I don't know, Amazon. Wherever you got a screen, you can watch the Techstrong gang.
Until tomorrow, though, on behalf of Mike, John, Andy, Liz, and the rest of the gang, this is Alan Schimmel. We're out.