42:53
This week, President Trump and some of the biggest names in AI are heralding the era of superlearning, while the companies building it are still trying to figure out what that means. We'll also look at the IPO market, from Anthropic's potential massive debut to Aura's decision to withdraw its offering. Additionally, we'll highlight a few deals this week, including a startup that raised $140 million to track events at sea. Welcome back to Equity, TechCrunch's flagship podcast about the business of startups. Today is Friday, October 2. I'm Kirsten Korosek, TechCrunch's transportation editor, and I'm joined, as always, by our weekend editor Anthony Ha and senior special projects reporter Sean O'Kane, and I'll see you in a few days.
This is already very close a little over a week we will be in San Francisco for Disrupt. We have a promo code for Equity listeners-Equity 25, you'll get 25 % off, and we'll be recording the episode live on Disrupt. I mean, there's a lot more going on there . I will be the host of the Startup Battlefield. What else is happening? I'll be interviewing RJ Scaringe, founder and CEO of Rivian, and we have a lot to talk about since he was on our stage four years ago. So I'm looking forward to it, and you, Sean, have a lot planned there too Yes, some really good panels. I think I'll also be on stage with the robot, so come see if the robot tries to fight me or who knows what. Or you try to fight a robot.
Listen, you know, if we sell out tickets to Disrupt, I will do whatever the audience wants, and that's a promise I will definitely keep. Yes, I was looking at the agenda right before the recording started, and we'll have a lot of speakers from Anthropic, Amazon, and OpenAI. Mark Wahlberg will be there too. I think it will be a very, very intense show. My last big pitch-well, I'll say what I really like about Disrupt. I have been visiting it for 8 years. I have been working at TechCrunch for 8 years. And the energy of all these early stage founders is a
dream for networking to find your next employee or your next investor. And that's actually what I enjoy even more than doing the interviews on stage-just getting to know all these founders. So, use this promo code again, it's equity25. And also come to our live recording. I'll say, I'll say, yes, come to our live recording. This is how we start the whole event in the morning. I have fond memories of when our, um, beloved former co-host Alex Wilhelm was on the show and there were superfans of his in the audience who were actually recording the show. So we need to bring that energy back. So please come and help us start Disrupt.
Yes, come pirate the recording on the Builder's Stage at 9:00 AM on Tuesday at Disrupt. Well, that's our call. Let's talk about what happened this week. And a lot has happened this week, including about superintelligence. Anthony, what happened? So, we are officially, by order of the President of the United States, no longer artificial intelligence. This is superintelligence. We talked about this, I think, last week, because it's something he hinted at, or I'm not He doesn't make hints. He's been saying this for a while now, but now it's signed into law as an executive order that everyone, if you're an official representative of the United States, you have to say " superintelligence" and not
"artificial intelligence." And it was also part of a broader announcement on AI security, which I think is interesting both because of the deal itself and because happened in the first place, because it all came together pretty quickly, because, again , just two weeks ago, a week ago, President Trump was saying, "We don't need any of this." Like, you know, we should be leaders in AI. We still can't stop him. We don't need any regulation. Anyone who says there is any danger from AI is essentially a Chinese or democratic agent. And yet all these CEOs of AI companies got together with him earlier this week and signed this deal. So, first of all, I'm wondering if any of you saw the seating chart at this dinner, which was pretty amazing. But before that, it's worth
saying that not only was there resistance to regulation, and we can talk about whether this executive order does any of that, which I would say it doesn't. There's a bit of theater there. But what is important and interesting is that just a few days ago, Anthropic was, let's say, out of favor, given what was going on behind the scenes with President Trump. And everything changed when Dario Amodei was invited to dinner at 10:00 p.m. He had this meeting. And just a few days later, we had this dinner. And this dinner was attended by almost every influential person from every major tech company. So, Amazon, Mark Zuckerberg was there, Jeff Bezos was there, Elon Musk was there, Dario, of course, was there too. But perhaps my main question
to both of you is: what does it all mean, and does it mean anything at all? How much time do you have? Uh, I mean, first of all, I'll say, you know, I got.Of course, they had dinner, Trump and Dario. If you read the Time magazine interview that Trump just gave, you know no more than I do about how much he was even aware that this dinner was going to happen, which is a strange reality, but that's the reality we live in. And I'll say one more thing that, you know, makes me skeptical- surprise, surprise- about how much the relationship has improved between the White House, or the administration, I should say, and Anthropic, because, despite the fact that there are people who are clearly trying to bring these two groups closer together,
the Department of Defense is still very adamant that Anthropic is horrible, a "woke" and a. And Neil Michael, you know, it seems like he's been bringing Anthropic to X almost every day. So I think there 's still a certain distance between them, despite Dario showing up at this event. As for the deal and all that, you know, let's be realistic. They even made a mistake in the name of the United States in the agreement that everyone signed. Moreover, it is absolutely not binding. That is, it is completely voluntary. I think what I liked most about this is that Trump described it as, quote, "morally binding," which is a level of abstraction and terminology I've never heard before.
Yes, I think to reinforce both of these points, in the context of Anthropic's role and their relationship with the Trump administration, there have been a lot of reports over the last year, especially during that high-profile scandal with the Department of Defense, that different parts of the White House had different relationships with Anthropic, and for example, the Treasury Secretary seems to have been a big supporter of them. While the Department of Defense was definitely not a big fan of them, or at least was unhappy with some of the conditions Anthropic tried to impose on the use of their technology. And in this sense, it seems that nothing has essentially changed, right? That there are people in the White House who want to use
Anthropic technologies, who want them to be involved in these decisions, and there are others who really, really, really do n't want it. And regarding this agreement, I think, yes, it doesn't look very meaningful. And what really matters is that they thought it was worth doing. What they, how would you say it, as evidenced by the fact that they made a mistake in the name " United States," seems like they were in a hurry to get all this ready. So the fact that such a gesture was needed seems significant, even if I don't think the deal itself will lead to any major changes. I thought Mark Zuckerberg said that this was a first step, which I think is pretty impressive, when even Mark Zuckerberg says, " Yeah, it's not really much, but at least we're all
saying here that we're going to try to do something good." Yes, let's all join hands. I really think there was one important thing that happened, and for me that's the main goal-it was the rebranding around AI. AI will kill us, that's scary, it will take jobs, but superintelligence-no. And for me, that was the main message. This is not my personal belief, for me it was a message. And you saw that the name is now officially from the White House- superintelligence, so SI instead of AI. You saw a bunch of executives, including Elon Musk, correct themselves, you know, along with Trump, saying, "It's SI." Um, and you see how in the last day or two, a lot of these tech leaders have been posting on X with the term " superintelligence," and the
hyped startup Safe Superintelligence now seems to be well- positioned to capitalize on this name change. But, in my opinion, the main purpose of this was rebranding. It's not scary. We are monitoring everything and operating as normal I think this makes a lot of sense, but I also want to point out that I think there are several levels of self-deception going on here. The first is that only rebranding and changing the name will change anything. People are demanding more transparency, more details about what the technology is, how it will be implemented, and how it will affect them. What does it really mean to have a
data center in your own backyard? Why does this need to be discussed in secret? These are the reasons why people resist this more than anything else. So, in my opinion, simply changing the name and shifting the goals, which are still designed for a hundred years ahead, will not change anything. Hmm, it's very similar to me too. I just published an article yesterday about Elon Musk and Tesla, who even a year later are still trying to define, quote, " amazing abundance." It's clear that all these people have communication problems. Other levels of illusion-for example, David Sachs' posts in X about this being a Woods, a conference convened after World War II to revise the world order, and oh, there's a lot to think about. Well, I do n't even know. That's all
I can say on this subject. I'm sure there were a lot of deals made there when so many people gathered. Let's just say, I think there are a lot of critically important things happening in the background that will likely impact the story, but not in the way it's being presented. Yes, but it's not as if these people weren't already close to each other and involved in each other's affairs, right? It wasn't like, "Oh, they're finally all in one room." They all have stakes in each other's businesses in different ways. Yes, I wouldn't attach that much importance to it. You know, it seems to me that this all looks a bit theatrical, like a PR stunt, like many other similar things. At the same time, announcements of consumer AI
products are appearing, and they make a completely different, let's say, very warm and pleasant impression. I'm referring, of course, to Muse and her cheerful avatar, as well as OpenAI's Dots, released last week, a little bubble-like cartoon character that can be shaped into different shapes. Of course, there were plenty of memes and rivalries from the Meta team, who used Jolly to, you know, make fun of Dots. But while this meeting and lunch at the White House were taking place at the highest level of federal policy, real products were being produced at the same time. And it seems that we live in two different worlds. Yes, although, again, they are related, because you already said that, right? The main result of this meeting at the
White House was, in fact, an attempt at rebranding. To some extent, we are talking about both the new products of Meta and OpenAI, and the attempt to make AI more attractive and friendly. As for Dots in particular, in both cases I found it a bit strange and even puzzling. And not in the sense that, of course, when creating a consumer product, you want to make it cute and attractive. However, in both cases it looked a bit out of place. For example, in the case of the Meta avatar, they actually say, "This is for adults." "This is for adults." So why does it look like it's for kids then? This is very strange. As for the avatar, or rather the OpenAI agent, the use cases described in the TechCrunch article were not consumer-oriented at all. It was
about programming and scientists. Therefore, it is strange to think: "Why would someone like that need this very friendly "wrapper" when it comes to scientific or corporate tasks?" This again seems to have been more of an exercise in branding than product development. We talked about this last week, and in the case of Meta, it seems logical. Regarding OpenAI, as you noted, there was some disagreement because they are promoting consumer AI, but the tool doesn't do what. So I'll make a prediction right now.
I know we usually do this at the end of the year. But I don't think this project will go far. This is not a criticism of OpenAI. I just think they're going to have to make a very drastic U -turn to achieve the desired result. Well, I mean, again we're dealing with communication issues, but also, I think in this case, it's something more. it's a more fundamental philosophical issue, which is that the way these things are advertised as useful often doesn't align with how most consumers use technology or what they need solutions for, right? So often we see these agents being used to. I think the example that everyone has been picking apart this week as a failure is making reservations at restaurants, or creating multiple
reservations, or making reservations at places that are hard to get into. You know, all these things, while most people are short of money because food is getting more expensive and fuel prices are just terrible. And people just don't understand, they're like, oh, this.this just won't change their lives significantly. And I still think these companies have a lot of work to do if they hope to make this work, to find use cases that make sense for most people, and there just aren't any. Our colleague Russell wrote a really good piece about the current landscape of consumer AI, about the obstacles that stand in the way, and, you know, he tries to outline an optimistic scenario, but then, I think, he goes into detail about why you shouldn't be
optimistic. And he was talking about this Andreessen Horowitz presentation, which they tried to present as a positive thing, but I found it quite strange, is that right now, I think, and I think this is likely in the United States, but actually about 2% of people are paying for AI right now, for some AI-based product, and you don't see, you know, sometimes, of course, you look at the growth curve of a startup, right? And the numbers are very low, but every month you see this huge exponential growth, and so if you just keep drawing that line, you know, for another year, it gets very high, but actually from a paid consumer growth perspective, it's not. It seems like it's just, you know, a small increase each month, and obviously with a lot of negative sentiment
around AI, it's unclear if there will be a dramatic change in that growth trajectory. So again, it makes perfect sense why a lot of these companies that are still getting ready to go public are instead focusing on the corporate sector, because, you know, corporations have money to spend. And it's an interesting dilemma, because I think it looks like the consumer market is not where the money is going to be, at least in the near term. But also, to justify those estimates, those promises of how much AI will change the world, I think they also need to have some degree of presence in the consumer market, otherwise, I think some of them just won't live up to the hopes that people have for them Yes, I'm glad you brought up the topic of
justifying grades, and that 2% is a pretty wild number. Because if you focused on, um, certain circles on social media or, you know, certain accounts on X, you would think that 90% of the population is using these AI tools, but that's not the case at all. But I will say that despite this, we are seeing IPO activity, which I want to talk about in more detail. Um, an interesting thing happening right now is that you've had some really positive IPOs. And Sean, you're probably not going to agree with me, but I'm going to put SpaceX in the positive IPO camp. I mean, I know it hasn't held up to some of the heights it had in the beginning, but this is a company that has been able to hold its own.
Next we have Anthropic, which appears to be moving towards its IPO. Their S1 form was merged. But there are also companies that have backed down, like Aura, which just canceled its IPO. So, it seems like a contradictory situation. And there are probably two companies that I think hold most of the valuation of recent IPOs, and that's SpaceX and SK Hynix. And besides, this is the lion's share of all the value that has appeared since the summer. So what does this mean for you both in the future, I think? Sean, I'm sure you have an opinion on whether the This is definitely the main question at the end of the year, isn't it? The fact that Aura canceled its IPO indicates a slowdown or greater caution in the world and the market. And I think it's worth noting that Sam
Altman has finally come out publicly about this, even though it was previously only reported by other media outlets, he said, "We're probably delaying our IPO, even though we've filed confidentially, until we feel like our technology is secure." Oh, and how do you determine that? The more we learn, the more dangerous it can be. But most importantly, I think they're not only holding back, but there are reports that they're going back to the private markets for another huge round of funding, which I think is probably the Of course, not every company is OpenAI, and not every company will be able to raise funds like them, but as we mentioned earlier in this episode and in previous weeks, there's a lot of money in the venture world right now
, especially after exits like SpaceX, and there are these huge funds. So the idea that resources are exhausted and these companies have to go public is not necessarily true. So I think, you know, the fact that OpenAI is going back to the private markets in the next 6 months or so makes me think that more people and companies are probably going to be hesitant to go public than they were 6 months ago. The last thing I'll say on this is that I'm curious, you know, there are a few other companies outside of the AI world that we know are moving toward an IPO, and the one that really comes to mind for me is Waymo. Waymo just closed what seems to be a $ 16 billion funding round earlier this year, in February I think. And many thought their next step would be an IPO and
some kind of spin-off business or separation from Alphabet. If they go back to the private markets for another round, I think, then yes, we're really in the conversation that the IPO market for tech companies is at least cooling down. Yes, I have to assume that a lot of it will also depend on how the Anthropic IPO goes. And if it's really as successful as I think everyone hopes, it's not.you know, I think there are not many other companies that fall into that category. But at least then it will be a sign that there is some positive feedback here, some hope. If it's, you know, pretty ambiguous, then I think we'll definitely see the window close, because it's going to look like this: If Anthropic can't succeed in the public
markets, who can? I also wanted to go into a little more detail on OpenAI for a second because, as you said, Sam Altman has now given several interviews and statements about this idea that we should focus on security rather than going public . And I don't think that's necessarily a completely misleading statement, but certainly a lot of reports suggest that's not the only reason they're delaying it. Um, and this, for example, I think there was, um, a sense, or, again, a reported sense , that they looked at the SpaceX IPO and were like, "Oh, this is not at all how we wanted ours to go; it started out really high, but then it couldn't sustain all that value, and if our, you know, financials were valued in the same way, we could have a similar
experience," and that's not what they want. So it will definitely be very interesting to see what happens there. Of course, the capital markets, the public markets, um, provide a lot of capital, and a lot of companies make that choice. But there are also many more eyes on your balance sheet, and you have to justify those assessments. You have, um, you know, shareholders who take risks differently than a venture capitalist who, you know, knows the founder and is willing to make a bet, and so I think a company that's still figuring out its business model, um, you know, I would be hesitant to go to the public markets as well. Not to mention that any of these companies have ever gone down this path, by the way, because we see plenty of examples of
it. But I really think it's important to remember that, so to speak, the siren call of the public markets, you know, can be hard to ignore. But once you get there, the people evaluating you, the investors and the shareholders, you know, they're not kidding, and they, you know, have very specific expectations. And so if your company is like OpenAI, which, you know, has users and is growing, but is burning a lot of capital, maybe you'll stall. I don't know. Yes, we have to say that Reuters did some really good reporting. They clearly got some or all of Anthropic's-1 form, and we're talking about $42 billion in losses last year alone, ahem, spending over $7 billion on computing power last year. Income, as many
have reported over the past few months, has been increasing. Obviously Anthropic wanted to spread this somehow, or someone at Anthropic or at the capitalization table wanted to spread this before the actual S-1 filing came out. But we also got, you know, some details of what we were waiting for, which was a warning in the risk factors section, which is 80 pages long, and I do n't even remember seeing one that long before. Usually, risk factor sections are, I don't know, 20 to 40 pages long, with 40 being the maximum. And in particular, the warning that AI could pose, quote, " catastrophic or existential risks to humanity" due to the behavior it has exhibited in the past, such as, you know, quote, "resistance to shutdown or hiding or manipulating
information," or other behavior "that resembles blackmail." So, we have, you know, a lot of reading ahead of us in the coming weeks. There are few such risk factors. Yes. I wanted to mention another financial aspect that seems to have come up in the reporting, which, again, you know, is not yet public. I definitely didn't have a chance to read it. I look forward to the opportunity to check it out when it happens. I think there were reports that they had profitable months recently, and that, you know, was a surprise.
This has probably been mentioned before, but it also depends on how you adjust certain things and what you factor into the costs. But the thing is, hey, Anthropic is probably a profitable company or will be profitable when it goes public. I am skeptical about this. I would say that, despite my love of talking about IPOs, we have two deals that I want to delve into at least a little bit. And one of them was a story, actually, I think, a couple of stories. Sean, you have everything ready here, but one of them is a maritime intelligence startup called Quartermaster, and it caught my attention not only because of the amount, which was $140 million, but also because of what they do, which is quite interesting.
Yeah, not only that, but the fact that they're already raising a B round after, you know, closing a round, which was also quite substantial-43 million in May. So, it was a very quick turnaround, and actually their round was pre-empted by one of the investors. They were planning, you know, to come back to the market and raise some more money relatively soon, but this happened even faster. And I spoke to the founder, you know, Neil, who is a really interesting person, and I think he's a perfect example of being in the right place at the right time with a good idea. Actually, they have this thing they call a "smart mast." You know, imagine a mast or any high-up place on a ship-it could be anything from a cargo ship to smaller boats where they just add
weatherproof sensors, sometimes even regular, off-the-shelf stuff. It's really nothing too complicated or technological. They don't put forty lidars on one mast or anything like that. They simply view the maritime industry as one that has been left out of many modern technologies, including the ability to use AI to interpret what is happening. The idea is that now, if you use their technology, you have a much better understanding of what's going on in the ocean. And not just from the perspective of the operator, responsibility, or general questions like: "What happened?" And also: " Can we be absolutely sure what exactly happened?" -but also from an insurance perspective, and for people on shore, such as local authorities whose ports receive these ships. And
so, they raised $140 million in a Series B round, and I think the other point I wrote about in the article is also pretty obvious when it comes to "the right place and the right time." With everything that's happening in the Middle East, in the Strait of Hormuz, at least that's how become even clearer to potential investors what huge opportunities are opening up for them in this area. While reading this story, I just thought it was really, really cool. And that there is quite a big opportunity here. I also like the name. I like the product name, which I think is becoming increasingly rare for modern technology companies. I was curious since you touched on this a little bit in your story. I also heard about the competitive environment.
This doesn't seem to be an area where there are many other companies trying to do something similar. Actually no, and not on such a scale. I think this is, you know, one of those cases where they really found something quite new that had hardly been explored before. And this is not only a good time and place for funding, but also a good moment for the company to do this right now, when the capabilities of these sensors can be maximized. Unlike if you were doing this 10 or 15 years ago: recording video locally, uploading it after the trip, and then spending days or weeks analyzing it. This is much closer to real time. They seem to generate tens of gigabytes of data per day from each such
vessel. They currently have 650 vessels equipped, with orders for 800 units scattered around the world. They are everywhere, even in the Arctic and subarctic zones. So they also prove that the technology is capable of withstanding the fairly harsh conditions of the marine environment Yes, for me the most interesting, or perhaps the most surprising, thing is that no one has done this before. Because this is one of those ideas where you think, "Oh, I was sure something like this already existed," but apparently it doesn't. Given the vast amounts of data, I'm very interested to see how the company will evolve using all this information. There's another company that you wrote about, and it's also taking on an industry where you think, "Oh, no
one's doing this yet?" And this applies, in fact, to the supply chain. This company seems to be Atomic, founded by former Tesla employees. What's so three such guys there. One of them is now joining the company. Last year there were only two of them, but that's three people who were essentially working on the supply chain for the Model 3 and Model Y. And the Model 3, in particular, is known for its so-called " production hell." They solved the problems of how to get and move parts in the most extreme conditions imaginable in the automotive world, and they did it faster than anyone else. So, their idea was to use this experience and build an entire company around it with more modern
AI-based technologies. That's exactly what they're doing: over the past year, they've moved to partner with big tech companies like DoorDash and HelloFresh. The basic idea here is this: as a company dealing with any kind of inventory, you want to find a balance - minimizing the amount of inventory at any given time, but at the same time, being constantly ready to provide your customers (whether other companies or ordinary consumers) with what they need, when they need it. And that can be very challenging, especially today, when modern supply chains are constantly facing unforeseen challenges. So they have achieved significant success in this matter.
Interestingly, they are one of those companies that were sort of "incubated" at DVX-the venture capital firm of John McNeill, former Tesla president. But it was done a little differently than they usually do. They usually come up with an idea and select a CEO for it. This case was a little more organic: Michael Rossiter and his co-founders had an idea, wanted to pursue it, started within DVX, and then spun off. One of the most impressive points in your piece was mentioning DoorDash as one of your clients, which is already quite impressive for an early-stage startup.
Especially for a company like DoorDash, where logistics are critical. But, apparently, about 90% of purchases are made through the Atomic system. And I tried to reconcile that a little bit with the fact that, in my opinion, the funding rounds are relatively small so far. So this is a seems to me that just a few months ago, an AI company working with such large clients would have raised 10 times more. Am I crazy? I don't think you're crazy. I think this is a valid question. I think the explanation here has two sides. First, this is just their Series A round. And as I spoke with John McNeill for this piece, he thinks things should go up for them. Well, and
especially now that they have significant revenue growth that they can demonstrate by acquiring new customers. And they really proved the effectiveness of this model. He believes that things can move quite quickly from here on out. I think it's also worth considering that this is a fairly small team. I don't remember the exact number of people, but this is one of the companies where software plays a key role in this area. They don't deal with large assets or anything like that. So without a large team and without this component, the amount they will work with will naturally be less. Whereas, for example, Quartermaster, they manufacture these smart masks themselves, in addition to handling sales, AI, and everything else. So I think that explains why the
funding round seems smaller. I want to briefly mention one more agreement, because it is very timely. So, another story you wrote, again, Shawn: Tarta Space raises 5 million to bring insurance to the stars. And we're not talking about Hollywood stars, but about space. And the reason I wanted to mention this is because they were finalists in Startup Battlefield last year. And this is one of the most exciting parts of Disrupt, and it's already coming up. So how timely is this. It seems they have achieved some success by attracting investors, passing the competition, and having a sufficiently sound business plan to raise funds. Yes, I think many will be surprised that, despite the increase in activity in space, the number of satellite launches
and everything else, many of them are not insured. is to find a way to create an insurance market for space objects, satellites or whatever. It all started with them developing this kind of engineering software because founder Yuchi Chan was frustrated with the state of space software and how fragmented it was. He wanted to create a centralized system and then realized there was value in connecting it to the insurance underwriting market. One reason why companions often don't carry insurance is because it's simply expensive; because typical insurance underwriters don't understand all the intricacies, or they see some element on a satellite, like a fuel tank, and they think that increases the risk by a certain amount because it could explode,
you know? So their whole idea was to start it, you know, essentially from first principles and redesign a new insurance market for space. And that was, you know, cool. They were one of the most energetic startups last year. They just believed in themselves so completely both on stage and behind the scenes. And you know, it was nice to see how in less than a year they really turned it into some kind of success. Yes, Sean, you made a great point. I definitely remember their presentation and I remember meeting them, they were, I mean, they were great. And they, as you said, were very confident in themselves.
They absolutely believed in their own strength. Glad to hear that this confidence is bearing fruit. Unfortunately, our time is up, but Equity will be back next week. And until then, you can follow us on Equity Pod at X and Threads. Equity is hosted by TechCrunch senior reporters, producer-Teresa Lo Concello, editor- Kell. Subscribe on YouTube or wherever you usually listen to podcasts, and stay tuned for updates at techcrunch.com/events. Thank you very much for listening, see you next time.
