Amazon & Apple Quarterly Results, Crypto Boosting Private Market, AI Customer Service | Oct 31, 2025

0:13Welcome, everyone, to the Informations TI TV. My name is Akash Pasritcha. It is Friday, October 31st. I do not have a costume, but I am wearing blue because my hometown, the Toronto Blue Jays, are playing for the World Series tonight. And let me tell you, I'm feeling pretty good. I'm not going to say anything else, but we're going to be watching. I think it's going to be a great weekend. And we have got a great show planned for you today. Amazon and Apple reported results last night, and we're going to break it all down for you. We've also got two exciting founders coming on the show. One is building at the nexus of AI, data centers, and energy, and the other has gotten a lot of funding to take its shot at the AI customer support market.

0:52We've also got two great stories out of our newsroom today. Our crypto reporter is going to come on to talk about how the rebound in crypto startup valuations is going. And we're going to get to our weekend big read about Benchmark, the storied venture capital firm. It's going to be a great show. And so let's get right on into our first guest. Amazon and Apple reported results last night. As usual, Amazon Web Services and the iPhone were big focuses for tech investors. To bring us up to speed, I want to bring on Austin Lyons, a senior analyst at Creative Strategies. Austin, welcome back to the show.

1:25It's great to have you. Hey, Akash. Thanks for having me. Got my blue on as well. Go Blue Jays. Go Blue Jays. I like to hear it. Okay. All right. Let's talk about the results. So Amazon revenue, it was up 12 % overall, AWS revenue up 20%, which was the fastest in 11 quarters. What were the headlines for you coming out of last night's earnings? Yeah, what everyone's watching with Amazon is AWS. Amazon, of course, is a big retail business. Retail, it's tough. Margins are slim. It grows slowly. But right now, during the AI craze, of course, Amazon also has a cloud business, and there's huge opportunity to grow that.

2:04historically, you know, people have been pretty bearish. It felt like AWS was falling behind. Azure was having good growth. Google Cloud was having good growth. It felt like AWS wasn't keeping up. They'd invested in Tranium. It kind of felt like they didn't have enough GPUs and actually what customers wanted was GPUs. That all flipped on its head yesterday in the earnings call. It was all about Tranium. Andy Jassy and team said Tranium like 36 times. And the name of the game was basically, they're fully subscribed. All their Traniums are being rented. They just turned on 500 ,000 of them for the Project Rainier.

2:39And they're growing faster quarter over quarter. So now all of a sudden, it looks like AWS, they were right on the Tranium bet and everyone's happy again. And we should say, of course, for people who are just getting up to speed here, Tranium is, of course, the chip that Amazon is building for itself, hoping to be in competition with NVIDIA's GPUs. Although Andy Jassy was very complimentary of NVIDIA last night, which is sort of a pattern in these calls. You know, the thing about Tranium here is it's really touting its end-to-end AI platform. The same way Google has its TPUs, Amazon was coming back and saying, look, we've got it too.

3:18And the question in my mind really was, okay, he's talking about price performance as a way of evaluating how good Tranium is. Is Tranium actually a better chip? How does it actually stack up against NVIDIA's GPUs from what you're seeing? Yes, so you're right. So Terranium is an internally designed chip. You can think of it similar to Google TPUs, like you said. You kind of think of Anthropic as Terranium's first and best customer. And if you're comparing the GPU to a Terranium, let's call it an XPU. That's kind of the industry vernacular. um xpus generally like you kind of they're they're usually not quite as performant in terms of like raw flops or amount of memory or networking because you gotta think at the systems level and of course you have to write software that's a little bit more custom it's not cuda software you kind of have to port your nvidia software over um but how it becomes more price performance is that you still they still have good enough performance but uh amazon because they've kind of cut out the middleman they'll rent you this chip for cheaper so if it might have been four dollars an hour for an h100 let's say maybe you could rent a tranium for say two dollars an hour and and that concept of good enough i think is something that that will be interesting to see how it plays out you know there was also talk about headcount and and we heard the news about amazon's layoffs i found it interesting that he basically came out and said look it was not a financial decision.

4:55It was a cultural decision, which is a company of Amazon size trying to say, we're trying to be a little bit more nimble, a little bit more like a startup. And it almost seems like they're going in the direction a little bit like Mark Zuckerberg saying, we're trying to build a very agile team here as we have to pivot quickly. And so that was certainly something interesting to see play out. I want to move to Apple very quickly. Apple, it was another big quarter for the iPhone. The company said that they are expecting double-digit revenue growth in the current quarter for the iPhone. What stood out to you about the Apple results last night?

5:30Yeah, you know, I think this was classic Apple getting back to Apple Basics. They sell great devices. They sell services because of those devices. And they make good money on the devices, and they make great margins on services. And they were back to, hey, we launched iPhone 17, Great camera, great user experience, great battery life. People are buying those and we're selling record services on top of it. And so we kind of got away. Historically, we've had a little bit of that distraction of Apple intelligence and did it land or not. And this felt like it was back to classic Apple, just disciplined and selling lots of iPhones and services to go with it.

6:11How are you expecting the AI story to come back with Apple? Just as you're watching the company and its products, Are you expecting it to come more into the conversation? Because it feels like it's kind of drifted out of focus for the company. Yeah, it totally has. And, you know, I think early on, there's just a lot of questions of should Apple train its own model or should they, you know, bring ChatGPT in the fold or buy perplexity? Kind of a lot of that conversation. But really, I think for Apple, it's just how can they make the best devices in a generative AI age, which means on-device experiences.

6:44They're working on improving Apple intelligence, but they're also integrating other players. And I think just, and by the way, they're also starting to expose with Apple's foundational models on-device generative AI for app builders to build local AI into their apps. It's just the normal apps you use every day that are third-party apps. So I think over time, we'll just slowly start to see more features coming, you know, being invented by third-party app developers, first-party apps getting better. And we'll just start to see over time more local generative AI. And what that means for Apple too, is that that just means they have to put more memory on device and continue to use Apple Silicon to have enough compute power to do that locally.

7:23So I just think we'll see a long tail of it getting better over time. Right. Well, Austin, there's a lot to take away. We could talk a lot longer, but I appreciate you coming on and helping us make sense of it. That is Austin Lyons, a senior analyst at Creative Strategies. One of the discussions that has come up in big tech earnings is the bottleneck that power is for building and operating AI data centers. Emerald AI is one startup that says it could have a solution to that problem. We reported exclusively that the company has raised$18 million this week from a group of investors that included NVIDIA.

7:55I want to bring on Varun Shivaram, founder and CEO of the company, to tell us more about what he's building. Varun, welcome to the show. It's great to have you. Oh, gosh, thanks so much for having me. So you had some big news this week. Congrats on the fundraise. Tell us a little bit about what Emerald AI is. Absolutely. You mentioned those CapEx numbers, Akash, and they're staggering, right? Hundreds of billions of dollars that are seeking to invest in building AI factories, this next generation of AI. But energy is the single greatest bottleneck that could prevent that future. There's 50 to 100 gigawatts of AI data centers that want to connect that may not be able to connect to American power grids and grids around the world.

8:34emerald ai we founded it just under a year ago 11 months ago to solve this exact problem and i think we've just come today to the point where the community in the industry is realizing oh my goodness this really is a problem and here's how we solve it just last week energy secretary chris wright announced that it was national policy to direct the federal energy regulatory commission to try and fast track the connections of data centers that can be flexible in how they use energy. And Emerald was founded 11 months ago to make data centers flexible. So we've been preparing for exactly this moment.

9:09And you guys are kind of interested because you're taking a software approach to this problem. That's exactly right. We take a software approach. So in Virginia today, it'll take you seven or 10 years to connect that next data center. But with our Emerald software that we've been developing, and now we're going to be deploying with NVIDIA, with Oracle, and with others, with our software, we can get data centers connected faster today instead of waiting years for brand new power plants or upgrades to the grid. Because here's what's happening. As we panic and we try and get those hundreds of billions of dollars of CapEx connected to the grid, we end up far over building energy infrastructure that we probably don't need right this second and raising costs for everybody.

9:52Emerald AI software, as you said, Akash, we're the software layer that helps AI data centers be flexible in their power consumption. And so what does the software actually do? So it basically helps balance the supply and the demand of the power. What does it actually optimize for customers? Yeah, it's a great question. Look, most of the year, 99 % of the time, power grids have plenty of power to supply data centers. There's actually 100 gigawatts of stranded power just sitting on our grids today that could supply data centers. the next several years of all that massive CapEx spending. But in very limited situations, the grid will reach a peak moment.

10:31We demonstrated this in Arizona in May in our first commercial demo with Oracle and NVIDIA. We demonstrated that you could take a real data center and take the AI computational workloads and our software would reduce the power consumption by 25 % on a hot day when all the air conditioners are running. And so when that grid is really stressed out in Phoenix, the data center's power falls for those three hours. and every other time that data center is able to take advantage of that abundant power on that grid. We did that in Phoenix. That's been now accepted for publication in one of the world's top scientific journals.

11:05We did another demonstration. Then with National Grid, which is the United Kingdom's national grid operator, we've announced a partnership to do a major demonstration of the technology at the end of this year in London, in the United Kingdom, to showcase that AI can be flexible anywhere as our software throttles the power either by slowing or pausing particular AI workloads that are a little flexible or by moving AI workloads from one location, say Phoenix, Arizona, to another location where there's abundant power at that moment. So you mentioned Oracle. So Oracle is a customer of yours, sounds like.

11:41Which of the other big tech companies are customers for you right now? Look, we're delighted that at our Arizona demonstration, we worked with Oracle very closely and we announced this publicly along with the Electric Utility Association, EPRI, that we're all working together to make AI flexible. In addition, we announced just this week, on the same day that the information exclusively broke our funders, we also announced that NVIDIA, along with Digital Realty, Emerald AI, EPRI, and PJM, the countries of the biggest grid in the East, announced the world's first power-flexible AI factory at commercial scale.

12:16And these are subscription deal? I mean, it's a software. So how is it structured? Is it an annual subscription? How does it work? That's a great question. So just very quickly to finish the announcement, 96 megawatts commercial scale. So that means that this data center will be the world's first facility at this scale to be power flexible and to meet the needs of the grid. And that's going to be deployed in Manassas, Virginia. And what Governor Youngkin, the governor of Virginia said is, look, this is a great deal, not just to get far more AI factories connected to grids, but also to help protect the rates of communities because rates are going up.

12:56But with this technology, you actually make power more affordable. You make rates go down as you have more AI factories come that can help the grid in moments of stress without having to overbuild expensive energy infrastructure. Right. All the software monetized, look, we create tremendous value. Every time you get a data center that's connected faster or a data center that's expanded, say 200 megawatt data center that's expanded to 250 megawatts, you create several billion dollars of net present value because you can sell on very lucrative cloud services earlier than you thought you could. You can upgrade to the latest NVIDIA B300 or Vera Rubin chip.

13:33And so Emerald AI is in the business of just making sure that our partners, whether it's the cloud provider, whether it's the data center developer like digital reality, or whether it's the utility, everybody makes money and sees values and serves customers. And yes, some portion of that value enables us to provide this software. But the critical thing is there's many orders of magnitude more value that we're generating. Four trillion dollars of investment off the sidelines into the AI revolution without raising everybody's power bill. Right. Well, great. Right. It's a fascinating approach to the problem that very much is in vogue right now.

14:08Like you said, CapEx and power was a discussion that was very much in the spotlight with big tech earnings this week. Thank you, Varun, for coming on. That is Varun Shivaram, the founder and CEO of the new energy company, Emerald AI, here on TITV. Okay. This past year has been a boon for crypto, propelled by the Trump administration's Lucent financial regulations and the appointment of crypto-friendly officials. To that end, we took the pulse of subscribers to the information, and a new survey that we conducted found that half of readers traded crypto in the past three months. 19 % say they did not trade crypto in that time frame, but have in the past, and 32 % have never traded crypto at all.

14:52Bitcoin, Ether, and Solana are among the most bought cryptocurrencies in the past three months. They are, of course, three of the biggest cryptocurrencies. And you'd think with the number of crypto IPOs we've seen this year, those would be popular buys, but readers mostly stayed away from IPOs. In the last six months, just 6 % bought Circle, 4 % bought Bullish, and 3 % bought Gemini. That is all compared with the 15 % of readers who bought stock in crypto's most mainstream company, Coinbase. For more subscriber insights, head to theinformation.com. We will link the story in the show notes. And speaking of crypto, my colleague Yueqi Yang published a great story today on how the crypto mania in the public markets is causing much excitement in the private markets.

15:37I want to bring on Yueqi to tell us more about that. Yueqi, welcome back to the show. It's great to have you. Hey, Akash. So you had this great story here today. And I want to start with Kraken because you had some great reporting on the company and its status with fundraising. What did learn about kraken and then we'll talk about the private markets at large yes so kraken is a very popular name right now among investors in the private market largely because they are seen as the next major crypto company to go public potentially next year and uh and it's being reflected in the valuations that they're getting uh this summer they were able to raise funding at 15 billion valuation but really quickly after that now they're raising a 20 billion valuation And this is a trend that we're seeing across multiple crypto companies in the private market, where they're all seeing this big jump in their valuations just over the past three months.

16:34And one of the things I like that you pointed out there is you took a look at some of the valuation multiples for Coinbase and where it's trading right now. And Kraken is actually trading based on that new valuation somewhere in the same ballpark, right? Yes. Based on the full-year revenue from last year, Kraken's valuation, if they do get to be able to close the$20 billion valuation round, it will be similar to the valuation multiple that Coinbase is getting. And this is very interesting because usually when investors invest in private companies, they're taking on more risk, especially in the sense that their shares are not as liquid as public company shares.

17:14and they're also taking on the risk that the IPO may not happen or may happen at a later point than they expected. You also mentioned some of the other IPO candidates people are watching. It's not just Kraken. It's BitGo, Polymarket Chainalysis, Anchorage Digital, Fireblocks. There's a whole host of private companies in crypto that people are watching. What's driving the excitement right now in crypto? We talked a little bit about the regulatory regime right now and how that's changed. Is it all that? Is there anything more fundamentally about blockchain being integrated more into the traditional financial system?

17:49What's driving all the excitement? I think the regulation factor is still a big piece. And that's because given how friendly the U.S. regulatory environment has become for crypto companies, the prospects for these companies have become a lot brighter. They feel like they're no longer under scrutiny. They're no longer facing enforcement actions from the SEC. And that gave investors more confidence that these companies will be able to do well in the longer term, and they will be able to gain more mainstream adoption. But then at the same time, the valuation is very high right now. And that's something that even crypto investors are acknowledging.

18:31And I think for these hot pre-IPO names, investors are trying to kind of find the right balance. of getting into interesting assets, but also not trying to overpay too much for the stake. And we should say one of the things you point out in the story is it's not a sure thing once you go public. You look at a company like Gemini, another crypto exchange that went public. You point out in the story that it's actually trading below its IPO price right now. Do we have any sense on why Gemini is lagging its peers? Is Kraken more of a Coinbase? Is it more of a Gemini? Where do we think it falls in that spectrum?

19:09Yes. So I think initially most of these crypto IPOs are doing very well, at least on the first day. And that's just reflecting the enthusiasm investors have for crypto names. Because for a long time, Coinbase was kind of the only game in town in the public market if you want to get crypto exposure. and now there's all this pent up demand from investors where they will just buy up almost any crypto name that's going public. But now what we're seeing and through the example of Gemini is that as more crypto companies are going public, investors will have more choices and they will be more discriminated and they will be able to scrutinize the business performance of each individual name and assign valuations based on that.

19:56Because they actually have comparable companies now to stack them up against. I think that's a great point. Speaking of which, very quickly, Coinbase reported earnings this week. What were the headlines coming out of that print? Coinbase is doing very well. Their revenue increased 55 % from a year ago, largely riding on this trading boom that happened during the quarter. Another interesting thing that contributed to their growth is that they're aggressively expanding the number of tokens that are available on the Coinbase platform. And as we know, during the prior administration, Coinbase was on the more conservative side.

20:32They couldn't list as many tokens as they wanted because a lot of these tokens have unclear legal status. They might be seen as unregistered securities by the SEC. And now that's no longer a concern for crypto companies. So Coinbase is moving very aggressively to take advantage of that. And it looks like they're diversifying their revenue pretty effectively away from trading revenue as well here, right? Yes, I think their revenue is still a large part of it coming from trading, but they do talk more about wanting to become a stablecoin payments service provider as well. And that's an area that I'm watching closely to see to what extent they can actually generate meaningful revenue from stablecoin payments business.

21:17And that's something obviously that's very different from the DNA of Coinbase, which is essentially a trading platform. Well, like you said, as more of these companies go public, it's going to be great to be able to compare and contrast the trading performances of all of these exchanges. Yueqi, I want to thank you for coming on. That is Yueqi Yang, our crypto reporter here at The Information. Okay, AI customer support has become a crowded arena, but it hasn't stopped investors from pouring money into it. Another company in that category is Maven AGI. The company most recently raised$50 million in June, and its investors include Dell, Cisco, Lux, and M13.

21:55I want to bring on Jonathan Corbin, the founder and CEO, to talk about where he's taking his business. Jonathan, welcome to the show. It's great to have you. Thanks for having me here. Good to see you again. So tell us about what Maven AGI is and what its unique angle is to the customer support AI problem. Yeah, Akash, have you ever had a bad customer experience? All the time. I had one last night, my cable decided to stop working. And what we're in the business of doing is helping people to prevent those from happening. We've worked across the entire customer journey from end to end. Whether you're having a bad customer experience because your Comcast cable is not working, or you want one single point of interaction across that customer journey, whether it's sales, support, making sure you're getting value of your product, that's what Maven's helping companies to do.

22:43Okay. And so, I mean, break it down for us. It's like an automated chatbot type thing that companies can use to handle customer service requests. I think it's really interesting, Akash, because in the old world, we kind of forced, like, you know, the last, the ML models and the legacy providers, we forced our customers into the hours that our team would work. Hey, sorry, we're only going to provide you with support between nine or five on Eastern Standard Time because that's when we work. What we've realized is customers don't want to be beholden to the hours that your teams work. They're using your products 24 hours a day, 365 days a week.

23:21You shouldn't have to work on the schedule that the support team works. And so what we found is in order for you to be customer centric, you don't want to just offer chat support. You need to be where your customers are when they are looking for support. You want to do it around the clock. That's right. Yeah. And whether it's voice, whether it's text, WhatsApp, voice messages, the search, what we understand is that most companies have three different types of customers they're trying to support. You have the digital natives. You have people who are looking for an interaction so that they can validate the answers they're getting.

23:54And then they want the human touch. And so we've built out a product that allows us to be able to support the customers that each organization has, regardless of which category they fall into. So why are you better than companies like Decagon, like Sierra, like Ada? This is a very crowded space. It is a crowded space. And I think, you know, it's really interesting. When you have all the different players in that space, each of the CEOs are going to tell you something different. Gosh, the ultimate measuring stick is what customers are saying. And what our customers are saying is, I went out there, I compared you to every other player in the marketplace.

24:30Companies like Clio and Thumbtack went out and evaluated 50 different folks in this space. And they said, the reason why it shows you is because you produce the best results for our customers across that entire customer journey. So I think it's easy for us to get distracted by headlines, by investor claims. At the end of the day, customers are the ones that matter. So let me ask you this. You've named the company Maven AGI. Surely you must have some sort of a concrete stance here on the debate that is ongoing with AGI. where do you fall on that? Is it two years? Is it 20 years? Is it even a thing that we are aiming for here?

25:08What's your take on it? Yeah, I think it's something that we're going to see in the next couple of years. There is a recent study put out. Like closer to two or 20? I think it's within the next three to five years. Three to five years. AGI, three to five years. Okay, and what exactly, how do you define it? For us, we're very focused on the business problem. And so how do you help departments be able to take action? at a higher level place. And so before starting Maven, I spent a number of years at a company called HubSpot. And what we tried to do is we tried to get our customers to get value out of the product by doing things like setting up a marketing campaign.

25:44And that kind of makes sense, right? Like you're purchasing marketing product, you do it in order to obtain leads. For us, when we think about business AGI, it's being able to tell the system, Maven, hey, produce 100 leads for me next month and it takes care of everything else. That's how we think about business AGI and where we're going as a company. Let me ask you this. Are you guys getting any interest from buyers? Any acquisition interest? I think there's a lot of players in this space. We're very focused on building a durable, sustainable company. Because that's how I see this shaking. I mean, there's so many different players.

26:20It's very fragmented in some ways, but the big tech companies are also building in it. I mean, don't you think this is just going to all get consolidated in the next two years or so? I think we've gone through trends from an industry perspective. You know, when you look at, you know, the cloud computing and you look at mobile, you saw, you know, kind of periods of consolidation happening. And I think what we saw is that companies with the best products are the ones that end up winning because your customers vote with their dollars. If you don't have the best product, you're not investing in that.