WOYM: META LLAMA 3.1 & Tech Earnings - Google, Tesla, TSMC, ASML w/ Michael Parekh (Ex-GS Partner)

All right, welcome to our next edition of Lumida Non Consensus Investing. You can find us on Spotify, YouTube, we're doing CliffsNow on Instagram, TikTok, Twitter, uh, anywhere you can, uh, consume content. We are there. I am really pleased to have back A repeat guest. Michael, you might be like our first repeat guest.

‍

Actually, Frank Rotman was our first guest from QED. Uh, so we're pleased to have you back. Uh, Michael is a, uh, world class analyst, you know, Goldman Sachs, uh, retired partner was studying the internet back there, uh, in the late nineties. And, uh, you know, you really saw this industry grow. You helped take Loudcow, uh, You know, public, for example, of course, that was founded by Mark Andreessen and Ben Horowitz.

‍

And now they started a VC firm called Andreessen Horowitz, which has become a real dominant presence. So you've done quite a bit. Thrilled to have you back. So I reached out to Michael. Thank you. I reached out to Michael last night. I was like, Hey, Michael, we've got Google earnings tomorrow, Tesla earnings.

‍

We've both separately done our prep and homework on these. So we're going to cycle through that and maybe we can get to other topics like CrowdStrike and crypto and who knows what else. We'll see where it goes. Let's dive right in. So let's start with Tesla and, and Elon. Michael, what are your initial reactions?

‍

Well, he's done an amazing job, obviously managing both the operational aspects that they've had to manage in terms of the productions, the issues with China and, and then managing expectations in the context of. Uh, full self driving, robo taxi, and the robots. And he did a good job on the call yesterday, continuing to paint the long term picture and prospects for those things.

‍

At the same time, the reality is the near term issues are the near term issues, uh, and that is some, some of the things that are reflected, I guess, in today's market action. But recognize, and I'm sure, you know, and you've pointed it out in your tweet stream The company's basically gained a Toyota in, um, in the last, uh, last few weeks, uh, in terms of this, uh, this mini rally that they've had.

‍

From 195, 196 to 250, so, so it's, to me, it's the, it's the, it's the, it's the biggest, you know, meme stock of them all. Uh, it puts the GMEs and the AMCs to shame, it's the super duper version, and a super sized version, uh, not because I don't, I don't agree with the fundamental directions that they're going to, going in terms of EVs and full self driving, it's just that it takes a long, long time, and, And the markets are, especially the professional investors, the, you know, my former peers, et cetera, when I look at what, you know, comes out of the Wall Street houses in terms of their, uh, assessments of the prospects for robo taxis and prospects for robotics and how they operate.

‍

Added on as a sum of the parts exercise to the valuations of Tesla, having been focused on internet and PC for 30 years, to me, all of those things have to be taken with a lump of salt, because these things take so much longer than people think, and, and that's sometimes not It does not come through in the stock volatility.

‍

Well, you're nicer than I am. I have a very critical take on the earnings call. Let me actually play a snippet of this call. I mean, the tone and the energy that Elon had, I thought, was very low. Of course, the day before, he had an interview with Jordan Peterson. But, like, take a listen to this. This is Elon talking about humanoids.

‍

And limited production starting early next year. Uh, this will be for Tesla consumption. It's just better for us to iron out the issues ourselves, um, but we, we expect to have several thousand Optimus robots produced and doing useful things by the end of next year, uh, in the Tesla factories, uh, in 2026, wrapping up, um, and at that point, we'll be providing Optimus to, uh, outside customers, That'll be production version two of Oculus.

‍

So,  so look, it was low energy. You go first. I'll follow. Yeah, low, low be very halting. So just to recap, what he said is, Hey, humanoids aren't production exciting, but Tesla will be the initial consumer, which means they're really not for commercial ready purposes. Uh, and wait until 2026. By the way, it looks like Robax are getting kicked out also.

‍

So it's, uh, yet another pattern of. Uh, selling the future seems to me to like pump the stock. And then there's like the, the rug pull where it gets, gets kicked out. Right. And, uh, that's happened again here. Uh, and you know, he said like humanoids have quote unquote, basic factory functions. If you recall this TFO call, I want to say two or three quarters ago, you know, he mentioned around humanoids, he said, well, we're still looking for a use case.

‍

Uh, and then Elon intervenes right there. I mean, it was like, you know, that, that, I think that's the issue is, uh, where's the delivery? Where's the commercial traction? Where's the earnings growth? You know, this stock is the highest valuation in mag seven. It is over twice as expensive as Nvidia on a Ford PE basis.

‍

And it's got the worst margins in the business and the margins have compressed six quarters in a row. And now they're facing the threat of competition from international players, uh, like China that can produce these efficiently, uh, and you've got players like Fabric in Texas. So I think he's got a hard hand to, uh, play here.

‍

See the humanoids, uh, uh, the, they sell well because people know this from science fiction. We know what humanoid robots are supposed to do. Mostly bad things we seldom have seen other than the Jetsons, we haven't seen robots actually do anything except in the cartoons where they, you know, make food and, and, and clean the house and that sort of thing.

‍

The issue here is that in an environment where after over a decade plus promising a 20, 000 30, 000 EV, the notion that we're going to have humanoids in our house, Or in our businesses that are going to be the 20, 10, 20, 30, 000, so the price points he's directionally pointed out are not realistic because you look at, I've spent a lot of time with, with robot, the robot market in the industrial market, right?

‍

Traditional robot companies that have been selling robots that are 100, Uh, from, uh, from a range of companies for the last 10 or 20 years. And there's an immense amount of complexity with them. The, the, these, it's literally like, if you again, uh, compare it to the car market, uh, what, what Google has done with Waymo, every one of those Waymos driving around in, in San Francisco or LA, or now, you know, in Arizona, et cetera.

‍

Those things have sensors that, you know, traditionally have cost hundreds of thousands of dollars. Those things are two, three hundred thousand dollar platforms. They're a long way before we ever The LIDAR, the LIDAR stuff, exactly. And, and so that's issue number one in terms of just the cost and getting it to a consumer point.

‍

When, when Elon says, Hey, where this is going to be a big market. And I've seen Wall Street reports from institutional firms that point to the human, the humanoid robot market as a bigger market than automobiles, which again, over a 10, 20, 30 year period, we may get there. But I think you touched on a key piece.

‍

What is the use case? What are you going to use these things for? We barely have figured out what to do with the Roombas running around. And, and, and Amazon tried the Astro and then they killed it because they were losing billions on Alexa. And they're re redoing that business. And Jeff Bezos is gone now.

‍

He's not focused on, on, on that Alexa and the robot piece, the home robot piece is gone. The last thing I'll point out on the robot side of the equation, humanoid robots, is there's a big transition going on, just as you'd imagine, from traditional robotic software, the Boston Dynamics of the world, their software stack is traditional machine learning, and when you go into the weeds of how this stuff works, it's just in the last year or so after the pandemic.

‍

chat GPD moment of open AI that the whole world of robots is moving into generative AI or large language model foundation robotics it's called. And Jensen Wang of NVIDIA has talked about it a lot and he's got, he's got some interesting platforms around robotics that he's, he's putting out, you know, and that those are tech stacks to look at and they're interesting and we can put up some details on that later.

‍

But the, the, where I'm going with this is that investors generally assume that, you know, because Teslas have this Hu hu huge Hoover vacuum of 2 million Teslas, uh, with their cameras basically providing all this data that, that data is, is not, it does not transport over to the world, world of the physical economy in, in, in your houses and your factories.

‍

We need. Whole new pools of data and you, there is no easy mechanism to capture that data initially inexpensively and then train the models around that. So Tesla, I'm sure with the humanoid optimist robot program, they're doing a lot of, uh, synthetic data, uh, creation, et cetera, but we are a long way from any real usable data on robots.

‍