Uber CEO Dara Khosrowshahi’s Autonomous Bet, Apple's AI Threat & Overhyped AI Startups | Aug 6, 2025

0:13Welcome everyone to the information's TITB. My name is Akash Pasricha. It is Wednesday, August 6th. It is the busiest earnings day of the week. Uber, Shopify, Shopify, and Disney all reported earnings this morning. We have also got Airbnb, DoorDash, and Lyft after the bell. Speaking of Uber, don't go anywhere because we have got the CEO of Uber, Dara Khosrowshahi, coming on the stream. In just a few minutes, we're going to ask him about the quarter and about what he's got cooking for the company. We have also got Snap earnings that we're going to get into with our friends over at Creator Ventures.

0:47We're going to talk with Adrian Aoun about OpenAI's new open weight model. And finally, we're going to hear from our financial analysis columnist, Anita Ramaswamy, about her thoughts on AI coding startup valuations. But let's start with the headlines out of Disney and Shopify this morning. Disney reported earnings, and the big news was that the company is launching a new streaming app for ESPN on August 21st. And ESPN is also acquiring the NFL network in exchange for a 10 % stake that the NFL will take in ESPN. The other big news with Disney is that the company said it is expanding Hulu into international markets, and the full integration of Hulu into Disney Plus is expected to be done by next year.

1:31On to Shopify, the company's shares were surging Wednesday morning as revenue grew 31%. That was much faster than it did last quarter. And the most important thing here is that the company said tariffs did not impact the business as much as they would have originally thought. Okay, let's talk about Snap. Snap reported earnings last night, and it was quite the contrast with other social media platforms that have reported. Total revenue was up 9%, but ad revenue was up only 4%. Just for context, Meta's ad business grew 21 % over the period. Joining me now, I want to bring on Sasha Kolecki, Managing Partner at Creator Ventures, to talk about the results.

2:10Sasha, welcome to the show. It's great to have you. It's great to be here. Thank you very much, Kash. So what stood out to you about Snap's earnings last night? Well, so the first thing to say about Snap's earnings is it's always a rollercoaster, whatever happens. If you look at the last 10 quarters of Snap reporting earnings, there, I think, has only been one where it's moved up or down less than 10%. So a 15 % move off the market, 19 % this morning is pretty much par for the course for Evan, Spiegel, and the team. But in this case, it was a disappointing quarter from the market's perspective.

2:42Like you said, revenue met expectations. applications, ad revenue a bit disappointing, earnings per share just about missed. But I think what the market was really responding to was although global DAUs was up 9%, you know, not all DAUs are created equal and US DAUs were actually down 2 % year on year, which I think a lot of analysts consider the kind of leading indicator for long-term revenue growth. And finally, oh yeah, go ahead. Oh no, I was just, I mean, let's dig into the DAUs because that's what jumped out at me too i mean north america users declining why well so you know the kind of stereotype of snap is that it's kind of for kids and you have this long-term issue they've been trying to get over of new users and cohorts of users joining when they kind of join high school and college and then gradually churning out over time uh but you know that that's been snaps long-term issue they are trying to retain those older cohorts, but they have been struggling of late.

3:39And, you know, you must use all these platforms. I mean, I'm curious about what you make about the direction that Snap's platform has been going, Snapchat. I mean, I don't use it as much. So, what have been your thoughts about the platform itself? Well, actually, I think although these results are not great, there is still a lot to be excited about in the snap platform the snapchat platform it's they've they've invested incredibly aggressively in spotlight through their creator program which is now spending 40 of usage watching content is now in spotlight which is basically a tiktok like product they launched in 2020 uh snapchat plus which is uh effectively a paid subscription a way of getting more features out of snapchat the most relevant of which is to kind of see who likes you and all those kinds of social features um are are also surging.

4:28So there's actually a lot to like about what they've been doing on the product. It's just the business model and the usage is not quite catching up yet. So if you're bullish on the platform, then what's the solution here, do you think? Yeah, I think they're doing a lot of things they need to do. Like I said, their Spotlight product really is doing well and it is working. I think probably the biggest upside case for Snap is unfortunately really out of the hands of the team. And it's always been, you know, TikTok seeing an issue with the US regulators and those trade talks not going through. Obviously, if TikTok were banned in the US, this would be incredible for Snap and Snapchat, which is not necessarily something they can control, but it's definitely an upside.

5:15One of the most interesting parts of the call last night, I thought, was the technical glitch that they talked about affected revenue? Can you share any of your thoughts? What happened? Yeah, so the way these ad platforms work is it's effectively based on bidding, as you know, as Google kind of pioneered a long time ago. And if there's a slight error with the algorithms, the way the bidding works, you can find some companies get extremely discounted ad placements. Effectively, the market's not working correctly. And that's what happened in the last quarter, the CFO said that actually it had a multi-percentage point impact on their year-on-year growth, which is staggeringly huge, if that's true.

6:01Right. And the company's outlook, I mean, sometimes the company doesn't actually issue any outlook. This time, they did give us revenue guidance. What did you make of the forecast that they had for next quarter? Yeah, they said that they were actually tracking towards it already. They were obviously getting back to growth. And they implied that they were actually the way Q3 was tracking so far, they were already on track to hit it. Maybe they needed a little bit of a boost to get there. But yeah, they seem relatively confident on making it. If your Snapchat and your earnings are always a bit of a kind of can go one of two ways, I think you're only probably going to give earnings or guidance that you're kind of likely to hit.

6:41So I do feel somewhat confident that they're more likely to hit it than previously. Great. Well, Sasha, thank you so much for coming on the show. Every time Snap reports, it is always a roller coaster, as you said. So it is great to have someone help us make sense of it. That is Sasha Kolecki from Creator Ventures. This morning, Uber reported strong earnings thanks to growth in ride hailing and delivery. Revenue grew 18 % and the company announced a new$20 billion share repurchase program. We are very excited to have Uber CEO Dara Khosrowshahi here to discuss those results and the future of the company, including their bet into autonomous vehicles.

7:18He'll be talking with our editor-in-chief, Jessica Lesson. Jessica, over to you. Thank you, Akash. Dara, thanks for joining us on TITV. It's great to see you. Happy to be here. So strong results today, nearly 20 % growth. And as you've shared, you're really seeing that combination of ride hailing and delivery build a flywheel for the business. Can you talk about that? Yeah, definitely. We're super excited. I mean, ride hailing is what Uber has been known for. And the ride hailing business grew really nicely. Transactions were up 19%. Bars bookings were up 18%. Profitability was up 22%. But then what we bring, what's special about Uber is that we are both in ride hailing and in delivery, and we're able to bring consumers on to one side of the business and then promote the other side of the business, which really is an advantage that we have over the other players.

8:16So about 30 % of, for example, Uber Eats' new customers come from a rides business. We've got the Uber One membership program that allows you to save money on riding or on delivery as well. So what you saw with our delivery business is actually gross bookings growth accelerated. You know, we're doing over$20 billion in gross bookings, so usually you don't see businesses of this size accelerate. But the gross bookings actually on our delivery business accelerated, really great profitability. So right now we're very happy about this combination of both working together. So you also gave an update on Autonomous.

8:53And I think this is one of the most interesting parts of your business, because you're striking a number of partnerships with companies like Waymo, but so many more. Why the eggs in multiple basket approach? Does it have to do with your questioning of whether any one product will take the market? Give us the update on Autonomous as you see it today. Yeah, definitely, Jessica. So I think the one thing that's a little bit counter instinctive about Uber and I was actually surprised when I joined is that we're actually a supply led business, which is the more drivers we have on the platform, the lower surge we get on the platform, the better reliability in our average ETA across the world is now four minutes.

9:37And as prices come down and reliability continues to improve, more people come onto the platform because it's a really easy way of getting around. So just like we are looking to add more human drivers onto the platform, we want to add more robot drivers onto the platform. And really what we want to do is be a catalyst for the entire autonomous ecosystem, so to speak, by sometimes partnering with, sometimes investing in a number of partners. We obviously have the Waymo partnership. We expanded in Austin. We entered into Atlanta, and that's doing really well. But we also have terrific partnerships with WeRide and Abu Dhabi.

10:22Again, a nice expansion there with Wave in the UK, Avride, Maymobility, Baidu, Neuron Lucid together, many others, because as we get more autonomous robot drivers onto our platform, our platform becomes more reliable. And as a result, we succeed. So we want the entire autonomous ecosystem to succeed. And I think partnered with Uber, we can drive a lot of growth there. Do you have any forecasts of sort of where a tipping point might be and when the robot drivers outnumber the drivers or just how do you think about where you are? I mean, forecasts in this space are notoriously off by decades, but now you're in it.

11:05You have partnerships up and running. What's your outlook? All right. So you just told me how all forecasts are wrong and you want me to make a forecast. Absolutely. So I can get you back on. So I think actually that the most challenging part that we're finding in Autonomous isn't on the software. And a lot of kind of these larger models and the advent that you see in large language models has an echo in Autonomous. And we're seeing the software, the newer software players really accelerate in terms of their time to market. The biggest issue is actually the hardware side and the OEMs. How quickly can you get cars that are affordable?

11:48And I would say less than$100 ,000, for example. Right now, these autonomous cars are really expensive. They've got sensor suites, multiple cameras, LiDAR. They've got big compute in the back, often with NVIDIA chips. These are really expensive pieces of hardware. and in order to make autonomous commercialized at scale, you need to bring the cost of that hardware down, which is, for example, one of the reasons why we partnered with Lucid. It's gonna take, I would say, two more generations of vehicles to get to the prices of well under$100 ,000, at least available in the US. China may be different or the rest of the world may be different.

12:30For example, Apollo, who is a part of Baidu, is already producing cars for less than 50 ,000. But in the US, it'll take two generations of cars to get there. That's probably five to six years to really get started with these lower cost cars and probably another five to 70 years to really scale to be close to the numbers of human drivers out there. So there you go. Another prediction that'll be wrong. 33, 2033, if I didn't know. Yeah, yeah. And how do you think about the competition from some of these partners? You know, I live in the Bay Area, so I've got my Uber app, I have my Lyft app, and I've got my Waymo app.

13:08I mean, how do you consider that in entering into these partnerships and that these partners are also looking at building their own networks? It's a way of life. Competition is a way of life for Uber. We've always competed with other networks. But I think that, you know, if you look at our Uber Eats business, I think you're going to see a roadmap to the future. in that, you know, Starbucks has their own app and they attract consumers to that app and they want to sell direct. But at the same time, in order to drive the utilization in their stores, they also work with Uber Eats. They work with DoorDash as well.

13:45I wish they didn't, but they do work with DoorDash as well. So essentially, I think you're going to see the same thing with autonomous. These are very, very expensive cars. You want to get as many rides as possible to as many riders as possible into these cars. And that will be possible with Uber as a partner, but it doesn't mean that some of these players like a Waymo or like a Zoox aren't also going to build a direct channel. So there's kind of the drama, which is, hey, either you compete or you work together. The fact is that in our ecosystem already in food, we compete and work together. And I think that'll be the same.

14:23The same will be true in Autonomous. Got it. I want to ask you about M &A. Obviously, we've been in a period for a couple of years where everyone's put the brakes on big deals. You know, you're in a position of strength. What about going out and gobbling up an Instacart? How do you look at the opportunities? Well, we just announced a really big acquisition, and that's a buyback of$20 billion of a really cool company. No regulatory approval required. Yeah, exactly. And I don't have to pay a premium. So it's pretty cool. So I think what you're going to see is our biggest capital spend is going to be with a buyback because we think, frankly, it's a really good deal right now.

15:03We want to reduce our share count, all the goodness that comes with it. We are going to look at M &A. We bought, for example, Trendy Algo in Turkey because we think that's going to be a huge growth market. We're already in there with mobility. Now we have mobility and delivery together, which is really strategic. But I think, Jessica, the best way to make sure that you do great deals is not to have to do deals at all. Organic growth for us is really great. Sure, we're going to look around. But at this point, I think the organic path is the most likely path that we're going to take. What about, I mean, you're so strong in many markets.

15:39And also some of your competitors are, you know, really taking steps to compete with you, I'd say, especially in Europe where we've seen more. What's your view on the European landscape right now? Well, we've been able to grow organically in Europe. We built our UK business. We're now the number one player there. We're number one in France. We launched in Germany about five to six years ago. We're seeing really good product market fit there and terrific growth there. We're number two in Spain, but gaining category position. So we've taken the organic path there. And frankly, it's hard to take the organic path.

16:13You've got to build in the market. Every single market is different. You've got to train your AI algorithms to work in the streets of Manchester and the streets of Lisbon as well. So it's not a surprise to see some of our competitors buy their way into the European market. The fact that it's high GDP, high average purchase rates, and highly profitable for us. You know, M &A sounds good, but the integration is actually pretty painful. So I actually think for the next 12 to 18 months, we're going to lean into Europe while some of these other players are distracted by all of the stuff that comes with integration.

16:53So I see this as an opportunity. But listen, we take competition seriously. Process who's bought in, DoorDash who's bought in. These are big companies. They're worthwhile competitors. And I'm just glad we've been able to build organically. Well, I just before we let you go, a couple more questions. But I actually want to zoom out. You know, you, today's results show you really restored, you know, Uber to really impressive growth and you've been in the role for a while. I'm kind of wondering what you've learned. I mean, what are some of your big lessons that, you know, going through the pandemic and getting to this point sort of stand out to you?

17:31I think for us, it's about product and it's about speed. A lot of the stuff that we build looks easy, but actually the work that our engineers put in, that product people put in, the algorithms that we build, the tuning of these algorithms, we're constantly reinventing ourselves. Every two years we build a product and then we replace it and replace it. In the end, it's about the tech and the product and kind of the data science that we have. That is what wins. Obviously, there's a lot that comes around at marketing operations. You know, we have to be responsible financially. But it's really the focus on products and engineering that has got us where we are.

18:12It's sometimes underappreciated, but it's by far the most important part of the business. Move fast. And Dara, I do just want to ask, because there is a story today in The Times that is alleging that Uber hasn't been aggressive enough in combating sexual assault through the platform. Do you have anything you'd like to say about the piece? I think, you know, first I tell you that sexual assault is a problem, you know, across society, obviously. And with Uber, we're doing 36 million rides a day. That same issue is going to show up on Uber. The fact is, I think as a company, there's no company that has done more to combat this, to be transparent about our safety record.

18:53We have now released a number of safety reports laying out exactly what are the issues that we've had. And I think we've innovated on safety like no one else. You know, you can track your ride now. You can reach 911 in an emergency. if we see the Uber go off route, we're gonna try to contact you. We have audio recording now for riders and drivers. We now have video recording available. We're absolutely leading in terms of safety, which has led to, if you look at our safety record and our last safety report, we've improved by 44 % versus our first one. That's not enough. As long as there's any sexual violence happening on the Uber platform.

19:40We're not going to be satisfied, but every single day we're working to get better. And, you know, I think that we're going to keep innovating and hopefully others can come in behind us. And we're going to be transparent in terms of what we're working on and also what our problems are. Well, it's a busy day for you, Dar. So we appreciate you being here and giving us the update and we'll be watching. And especially when it comes to autonomous. I I want autonomous everywhere, personally. So do I. Thanks for joining us. Thanks a lot. Back to you, Akash. That was Jessica Lesson with Dara Khosrowshahi.

20:17Okay, before we get to our next guest, I want to bring your attention to a story that we just published late last night about AI video company Runway. The information was first to report that the company is raising around$500 million at a valuation of at least$5 billion before the investment. That is up from the$3 billion valuation that the company fetched last year. And we also learned that Runway is doing about$90 million in annualized revenue. And you might recall that we had Runway CEO Chris Valenzuela on the show just a few weeks ago. One of the things that he talked about was that the company is focusing on the enterprise consumer.

20:54And they're also innovating this new offering where they actually put consultants on the ground or Runway employees on the ground with these enterprises to help teach them how they can use the software. I should say, Runway isn't the only AI video company that is busy fundraising. Luma AI is also raising money. They are raising at least$1.1 billion at a$3.2 billion valuation. That was also in our story. And a bit of juicy intel here. Meta at one point talked to Luma about potentially acquiring the business, I will link that story in our show notes. AI video is clearly very hot. Okay, yesterday, OpenAI released a new open weight model, GPT-OSS.

21:38It's the latest move in this ongoing AI war between the new titans of tech. And here for an insider's reaction is Adrian Aoun, a serial entrepreneur who spent a number of years at Google after selling them an early machine learning startup. I want to bring Jessica back to run this interview. Jessica, we'll turn it over to you. Thank you, Akash. And Adrian, welcome to TI TV. It's great to have you. Thanks for having me. Excited to be here. OK, so it's a hot AI summer, which just means headline after headline after headline. And for some reason, someone got the memo that this week is the model upgrade week.

22:13And so we've seen something new from OpenAI and are expecting more later in the week. Same from Anthropic. as Akash said, we've got an open weight model here from OpenAI, maybe taking aim at Llama a little bit. But what did you make of OpenAI's move and kind of doing their, I guess they had an open weight model earlier, but this is really one of their first big moves into the space. Yeah, I mean, the first thing I think we all have to say is congratulations, it's a damn good model, right? And so - Already played with it? It's been like 12 hours. You know, I don't sleep much. So look, take a step back.

22:53This shift all happened when Zuck and Meta started a few years ago and said, we should just be open sourcing these models because it's inherently deflationary to all the model providers. Now, Zuck can do this competitively because he can make money off the ads, whereas everybody else at the time was making money off the models. And so he said, I'm going to make OpenAI dance. And the problem is they danced incredibly well. What they did is they shifted their revenue model from we're just monetizing the models to now we've got applications and APIs that are built on top that they're monetizing incredibly well.

23:29And so here, if you're Sam, all of a sudden you go, wait a minute, that was actually a really good strategy. I didn't like it when it came out, but now I should make the models just as deflationary as anybody else. So if you're training these models and they cost billions of dollars, but they're a rapidly depreciating asset, then you're moving from a world where anybody can build models because they just got some funding in a research lab to actually the only people who can be in the model game are people who have real revenue streams. So from my perspective, if I'm Sam, this is fantastic. Now, doesn't this affect the ChatGPT subscription business ultimately too, though?

24:07I mean, if you're giving away some models, doesn't it hurt that part of the business, which for them is the cash cow or at least one. It turns out not. It turns out that users are trained not to download a model and use it, but instead to go to the website the same way users aren't flocking away from google.com right now because we're so ingrained and so trained to it. We've all been trained to chatgpt.com and we're using it at scale. But what you mentioned is interesting because who is this really good for? Well, the reality is this is very good for Apple. Apple doesn't have a whole bunch of on-device models that are best in class at this level.

24:44So they can start employing these. Now, who might this not be good for? Well, in some ways, it might not be good for society. It's a little too early to know. But we're putting more and more powerful models in the hands of anyone. Now, we were doing this with China, right? Launching DeepSeek. And now at least we've restored this. And the standard is now going to be, let's build back on the US, which matters both for public perception, but also because the US is worried about bias being built into these models. If everybody's building on DeepSeq out of China, maybe it starts to affect our culture.

25:20Or some of us are even more worried that there's hidden attacks in these models. What happens if we're all building on DeepSeq, but DeepSeq is very quietly inserting malware attacks or backdoors into the software that it's writing, writing nefarious code. So I think that generally this is going to be a good move, but it only increases the speed of the race of the big model providers. You mentioned this could be good for Apple. Say more. Obviously, they've felt a couple clicks behind in earnings last week. Tim Cote came out, went out of his way to say, no, we're in this race. We're going to stay aggressive.

25:55They have a partnership with OpenAI. Is OpenAI one of those sources behind Siri. How is this good for Apple and how should Apple play this moment? Yeah, well, let's first take a step back and say that Apple is just clearly behind on this one, right? And you can see it in their products. We're all using chat TPT and talking to it and having profound conversations. And we go to Siri and it can barely call my mother. So obviously we're not playing at the same level. Now, if you take a step back, the biggest risk to Apple, Apple says, well, we're not seeing it in our revenue, right? We're still selling phones.

26:30Apple's franchise is the iPhone franchise. And what keeps you using the iPhone? Well, at the end of the day, it's the apps on the phone. And that moat of a - What the camera? Anyway, yeah. Sorry? I think it's the camera, but yes, no, it's the apps. It's the camera, okay. Yes. But the camera is easy for others to replicate, whereas the apps are very hard for others to replicate. Or a better way to say it is they used to be hard. Right now, thanks to AI being able to write any app in a few minutes, thanks to this model context protocol, this MCP technology that everyone's standardizing upon, we're getting towards a world.