Limitless Podcast - The OpenAI Phone is Coming... With an AI Operating System

Episode Date: May 6, 2026

OpenAI's phone, led by iPhone architect Jony Ive, is set to launch in 2027-2028. With a goal of shipping 40 to 50 million units, we explore the implications of an AI-first operating system that redefines user interaction.We compare OpenAI's consumer-focused strategy to Anthropic's enterprise approach, discuss the challenges of competing with Apple, and ponder the hurdles for user adoption.------🌌 LIMITLESS HQ ⬇️NEWSLETTER:    https://limitlessft.substack.com/FOLLOW ON X:   https://x.com/LimitlessFTSPOTIFY:             https://open.spotify.com/show/5oV29YUL8AzzwXkxEXlRMQAPPLE:                 https://podcasts.apple.com/us/podcast/limitless-podcast/id1813210890RSS FEED:           https://limitlessft.substack.com/------TIMESTAMPS0:00 OpenAI Phone1:28 Features and Design3:01 AI Operating System5:10 Supply Chain and Production Concerns6:09 Johnny Ive's Design Influence8:58 Market Competition and Consumer Interest11:53 OpenAI's Strategic Roadmap14:27 Anthropic’s Shift to Financial Services20:20 Comparing OpenAI and Anthropic24:39 Future of AI-First Hardware25:40 Final Thoughts and Speculations------RESOURCESJosh: https://x.com/JoshKaleEjaaz: https://x.com/cryptopunk7213------Not financial or tax advice. See our investment disclosures here:https://www.bankless.com/disclosures⁠

Okay, this is pretty insane. OpenAI is launching their own version of the iPhone and it's shipping in less than 12 months. Yesterday, an analyst which has the highest hit rate of any Apple product leak dropped a report that revealed that OpenAIA is building their own phone and it's been designed by the very guy that designed the original iPhone, Johnny Ive. They're targeting around 40 to 50 million units for year one, which is larger than the opening sales of the iPhone itself. Now, this is a very convincing rumor, but we wanted to investigate one whether this was real, and it let us down the craziest rabbit hole ever. We pulled up Anthropic and Open AI's job postings, and it revealed a hidden product roadmap that we're going to talk about on this show. There's over 750 open roles between them, and it revealed a 24-month, two-year product roadmap that no one else is talking about. It didn't only confirm that Open AI is building a phone, but it also tells us how the companies themselves are actually building in a new. a very different strategic path from each other. Yeah, I didn't believe this, and I still don't believe

it, but every single rumor is pointing towards the fact that Open AI is building an Open AI phone, and it is going to be designed by the same person who is building an iPhone. And we kind of have a brief of generally what we can expect. I think most noteworthy of this most recently is that they've expedited timelines. So there's a high probability that one year from now, we will be holding an Open AI smartphone in our hands. Now, what is it going to look like? like, what is it going to feel like? We have some loose ideas of that starting with the form factor. It seems like they're going for the traditional smartphone, not the screenless Johnny Ive device, this isn't a pendant, it's not glasses, they're going for like a proper phone replacer.

So I would expect it to look like the iPhone. The big idea is that they're going to replace the operating system with an AI first operating system. And we've always talked about how OpenAI is trying to build the AI first OS, but this is going to be the actual hardware in which it can be integrated on. Traditionally, iPhones, you have that grid system where all your apps live. Yeah, we have a great, great rendering on screen. This will be fully predictive. So this is pending on a few things. One is that Open AI is able to figure out the software stack in time, which I mean, 12 months seems like an eternity. I'm sure they'll be able to figure it out. And the other is that be able to actually convince people to go and use this device instead of an iPhone. Now, this is

separate from the product suite that we've mentioned in the past, where Johnny I was developing earbuds or a pendant or a little puck that sits on a desk. This is a full-blown, a display with a screen and a smartphone all integrated. It seems like it's going to be built on top of Open AS Cloud. So I would expect memory and a lot of the graphical processing power to not be as important because a lot of it is going to be offload to Open Ice Cloud. And I think this is really interesting news. This isn't what I was expecting. I thought they were going to be the third device, like your computer, your phone, your maybe earbuds, but they're going for that phone slot. And this was really surprised for me to hear. But this analyst who was reporting on it, he is the most

accurate Apple supply chain analyst in history, and he also seems to have a lot of other people backing him up as well. Here's the thing. I do think OpenAI is building the third device. They're pitching it as a phone, but it's going to act very and maybe even look very different to an actual phone. So a few differences I want to point out from the off. This phone is rumored to not have any apps on it at all. You will interact directly with an AI, presumably chat GPT, and you will feed it, your data, your thoughts, your intentions, and it'll either generate apps or user interfaces on the fly, or it'll interact with other AI agents

that will conduct the research, process, information retrieval, whatever you might need for you on your behalf. So I'm describing here two different things. One, a completely different operating system that is more engineered towards the AI future or AI-centric world. And the other thing is a completely different hardware component system if you want to start serving this. You can imagine a few different things off the,

off the back, which is normal phones are oriented around, you know, running video, pictures, making sure apps can run optimally. This one is going to be oriented around inference and making sure prompts are process and AI agents can run locally, that your data can remain private and a bunch of other things around that. So I've written a summary here, you know, it's meant to allow Open AI to win at the operating system level. And what I like about this is there's no other frontier AI lab, namely Anthropic, that is going

after the vertically integrated layer. You have Google, which plays on every single layer of the stack for AI. You have Apple that kind of fumbled the model building layer, but has all the distribution through their hardware components. And Open AI has realized that probably quite early, actually. I have to give Sam credit for this about a year and a half ago. He said in a very weird tweet that now, in hindsight, makes a lot of sense. I am not, or Open AI isn't competing with other AI labs. We're competing directly with Apple. And that is like, I need to find that. But it's making a lot more sense now.

Not only is he trying to produce an iPhone competitor or a killer, but he's trying to create a range of different devices that can ingest your data, either passively or unintentionally, feed it to chat GBT and make it the greatest product experience vector that he can do. Now, a few other things, because I had an immediate question around this, right? How on earth are they going to deliver this in the first half of 2027? OpenAI does not have the supply chain prowess that the likes of Google and Apple already have. Well, the answer is they're going to be outsourcing this to a bunch of third-party users.

I believe it's Lockshare, which also happens to be the packaging and assembler of Mehta's rebands. Now, if that's anything to go by, I'm not feeling quite confident at this point. But there's also MediaTech and Qualcomm. MediaTech does a lot of the high processing packaging for a lot of the GPU and Neum. cloud assemblers. And Qualcomm has had a lot of favor in the mobile phone market. They have like Snapdragon, which runs in a bunch of Chinese phones specifically, but they have the scale to help open air operate this at a larger extent or to match with iPhone. And the final thing is, what gives me hope is Johnny I being in the mix here. Josh, I know you probably have some strong

thoughts about this, but I think his design touch on top of this kind of like hardware scaling bet will give them the best shot at, you know, at least competing with the iPhone or whatever the next iPhone is going to be later this year. Yeah, it's so hard to imagine. I would want to put another phone in my pocket that's not an iPhone. But if anyone's going to design it, it's going to be the guy who designed the iPhone. And I think this post says 2008. We actually have an update today from Ming, the analyst, who is pushing these timelines up to 2027. In fact, it also gives us a few leaks of kind of the specs that we could expect inside of this phone. And this is something that's interesting too because we already know what's going to be in this thing. And it's amazing the supply

chain and the leaks that are able to get out and stuff. But basically what this is saying is that it's going to be running on all of the cutting edge technology that you'd expect a frontier iPhone 19 model this year to be running on. They have a deal with TSM to create those two nanometer class transistors. That is the current state of the art. There is nothing better. They're using a lot of low power components. So I think one of the more important things for this phone, because ambient capture is going to be so important, is energy efficiency. over raw power. So it seems like some of these design choices they're using are for extended battery life. So it can be on all the time, constantly listening, constantly taking information.

And you have to imagine this form factor, the reason they decide to go with it is because it's the only one that makes sense to be a true assistant throughout your day to day. It has the suite of cameras, the microphones, it has the accelerometers, it has the GPS baked into it. That's really difficult to include in this suite of devices like the earbuds with a small puck. we have a lot of information on that. We also have, like you were mentioning, the suppliers who they're going to be building with. I mean, these two nanometer chips are allegedly going in production in the second quarter of this year. That means these are going to be start being made in the next couple of months. And we have deals with MediaTech, Qualcomm, TSMC, and Luxhare. And like

you said with Luxhare, they do make meta's classes, but they are also Apple's number two assembler behind Foxcon. They've built AirPods, the Apple Watch, and an increasing share of the iPhone. they're using people who are basically the best in the world. They're using best in class components. And it's hard to imagine that this device isn't going to be excellent. It seems like Foxcon is also involved, which I didn't realize. And Foxcon currently, they build Open AI Stargate services. And for those who don't know, Foxconn has basically built every Apple device you've used. Like Foxconn makes all of the iPhones, a bunch of iPads, Macbooks, the whole thing. Foxconn has been building Stargate servers, but now they're also going to be building those Johnny I've designed devices while

Luxhare takes the smartphone build. So it's this crazy assembly manufacturing setup that they have where I think it's actually going to work. If any company is going to build a product like this, it's going to be open AI. They have the right people in place. They have the cutting edge products. I guess the question is, is like, are you actually going to want to put an open AI phone in your pocket instead of your iPhone? That's an easy answer for me. Like, absolutely yes. But I'm also like a sucker for like some of the new devices. Like I'm not going to replace my iPhone, but I'm going to have it next to it because I want to see what this new operating system is going to look like. And to be honest, I don't think current mobile phones are built for what the future of AI is going

to look like or function as. Like, if I'm going to be interactive, like, look at Siri. Do you use Siri? I sure as hell don't. And like, I want to be able to use a Siri enabled, I don't know, chat GPT that can do things for me and access all my apps so that I don't have to like scroll and tap into them, right? Anyway, to your question around like, this has to work. I'm probably not as confident that it can work.

or not as sure that it can work, because if you look at like this stack that they're building, there's a lot of new things here, right? So OpenAI, ChatjipD, is known for absolutely nailing the large language model architecture, right? That's what they've made their bread and butter on. What they haven't proven is that they can get it to function optimally on a small enough device.

Now, if there's one thing that Apple has all the experience on, is hardware optimization. So if anyone's going to know how to run locally run models either on device or through private cloud, because obviously there's a lot of data encrypt, stuff that needs to happen there. Apple has all the historical knowledge and expertise to be able to potentially pull that off. Open AI, I'm not entirely sure. Number two, Open AI really is trying to define like a new AI stack here, like an operating system. That's not only going to be

large language models. That's going to be images and image gen. Now, they have the world's leading image gen model, which is great news, right? So they're going to combine these two things. I think their dual architecture allows for like a vision language model to work with an alarm, which is going to be very cool. But I'm also curious, like, how are they going to function with the camera ingestion in lifetime? Like, we're going to use this camera very differently. Maybe it's not to take pictures, but maybe it's like to show the LLM what we're doing or what we're seeing, and then it can kind of address things. There's a lot of newness is what I'm, I guess I'm trying to say at. And I'm not entirely sure they can pull it off, but I'm hoping with Johnny I've at the helm,

and again, I'm a Johnny I fan that he'll be able to figure it out and pull it off. Yeah, and I think I like to, the next thing I'm looking for with these leaks is confirmation that Johnny's actually going to be designing this phone. It seems like he is. I don't understand why they wouldn't put him and the love from team, or I guess the I-O team in charge of designing this. But I want to make sure it's actually him or if this is actually a hedge against his screenless device is not working. So that's going to be an important thing to follow as we get more information. A lot of the next questions that I was asking myself is, why are they doing this and why are they doing it so quickly? And I think there's a few reasons for this, right? Because when you think

about the amount that they're going to be able to actually build, you said, 30 million units is what they're kind of aiming for. 30 to 40 billion. So assume these costs $1,000 a piece, which is kind of like a high end phone price. That's 30 to 40 billion dollars of revenue over two years. Anthropic just made $25 billion of revenue in like the last two months. So I'm not sure how much is going to move the needle. And I have to assume this is probably for the beginning of that harbor mode, but also the timing is funny because this is rumored to be when they want to IPO. And what better catalysts for getting people excited about your company than to build a new suite of hardware that can compete with Apple.

Whether or not it actually does a good job, that is TBD. But I assume this is probably a nice, really big catalyst for them as they go into IPO season as well. And I just can't help but think that that has to play some sort of a role in it. And then there's also the business side where you could bundle subscriptions with it and they get developer ecosystems for the hardware and you get that hardware remote. But man, like Apple's got the hardware remote. And the problem is just a skill issue with software. They just can't build the software.

opening eyes approaching it from the other direction. Can they do it? I don't know. We'll be buying it. We'll certainly be trying it. I promise you that. If it's around, we'll be buying it. But of course, I want to take the bull hat off for a second and talk about some of the challenges here. Number one, we need to make sure that it actually ships in 2027. Number two, we actually need to figure out whether people actually want this thing.

I know I do, but I'm a fan of AI in general. Will the averages with my mom and sister even be interested in this? I think the allure of chat GPT is good enough. You mentioned bundle subscriptions earlier. I do think that is a good enough. enough selling point versus like, hey, we're launching a phone and we don't really know what to do with it. And then I'm also thinking about Apple's response specifically because I know Apple is undergoing a very aggressive transition now to orient themselves around the new AI world.

They've launched their new M5 chips, which is probably the best consumer device components that can run AI on device. And they've got a new guy at the helm, John Turner, which is a hardware expert. He's spent the last 35 years building Apple's hardware. He should be the perfect person to build whatever that new device is. So they're going up, it's David versus Goliath, and I don't know whether David's going to win this particular battle,

but it's interesting. I think opening I will be able to pull something off. They have, heck, they have enough in financing. I think they're going to be raising how much in the IP?

I think it's like 70 to 80 billion, maybe even more, at a $1.3 trillion dollar valuation as a loss that I've heard on the rumor mill. So if anyone's going to pull it off, it might be done. I hope they do. I think if I'm giving my personal take on this,

I'm pretty bearish and pretty pessimistic on the phone. I can't imagine, like, the amount of love and addiction that everyone has for their iPhone is so strong. And for Open AI to come, especially when I'm flip-flopping, people are flip-flopping between Anthropic. We just filmed an episode on Codex when a month ago we were saying Codex kind of sinks, the amount that people are flip-flopping between these services, the amount of kind of buy-in that's required to not only commit to Open AI, but then transfer your phone that you've had for the less 15 years away from that with all of your eye messages and your social apps, that is going to be incredibly hard to do. And the auxiliary device pathway, the one that Johnny was making,

the screenless devices, the accessories to the phone, that made sense, compete directly with the phone. I feel confident they'll be able to bring it to market. Will people be able to buy it? Will people want it? I don't know. I don't think that's the case, but we'll see. We'll see. And that is one of many things that we are going to be kind of discussing as we uncover this roadmap as detectives. EJS. You's had your magnifying glass out recently, going through the job board listing, seeing what these companies have been up to? What have you found? What is in the roadmap for these companies as we progress, I guess, towards the Open AI iPhone early next year? Okay. So just to explain my logic, I was curious by, my curiosity was piqued by this new phone, right?

And I was like, how can I convince myself that this is actually going to be the thing? So I was like, well, let me look at the job postings. If they're hiring for like, you know, a mobile phone engineer or whatever you want to call it directly, then, you know, it's hint, hint, that's probably what they're working on. But I ended up discovering a whole other gold mine, which is their entire product roadmap, Open Air specifically, and I did the same with Anthropics over the next 24 months. And what surprised me the most is both of these companies, I thought would be very similarly aligned in their strategy. They're not. They're actually doing two very different things. And I want to kind of like tell that story over the next couple of minutes. So I looked at Open AI and I have it displayed up here.

I'm not going to walk through the specifics, but I'll give you the general idea of what I've found. So there's around 330 job postings currently live as of today. A hundred got posted in the last week and a half. So Open AI is really ramping this up. Now, as you'd probably expect, 25 to 30 percent of these roles are roughly sitting at the point where it's directly for AI forward engineers. So we're talking about researchers, engineers that can actually build and train the model and make this thing actually useful for them. 12 to 15% of the roles are specifically around infrastructure and engineering roles. And this is where you see a subset of hardware-specific engineers that are focused on devices

specifically. So one could assume that if Open AI is building a mobile phone or a suite of different devices, this is where the roles are being hired for. And it takes up a non-trivial percentage of the available roles. Now, when I look at a setback from this, right, this seems, I guess, kind of surprising, but pretty obvious. Open AI. underwent a code read over the last, I was going to see six months, but it's literally only been like three or four months, and it's crazy that they flipped script with Anthropics so quickly. But they've been able to undergo this process of change where they're hiring for very compute, intense, research-heavy things. They've trimmed a bunch of their other efforts

to focus on building the best model, specifically the coding model. And that is represented in what I'm seeing right now. But there's a strategic move that I want to call out here, which is Sam made a very big bet about a year ago, which is to go all in on compute, Stargate specifically. He also made a very big bet three months ago to go all in on coding AI specifically. He's treating this like a warlord empire where his entire ancestral lineage depends on this.

So he's not treating this like an enterprise company. He's not really treating this like he has an IPO coming up in a few months. He's like going all in. It kind of reminds me of Larry Page, where he was quoted saying, we're going to go all in and I'm willing to go bankrupt losing this race,

but I can't afford to lose the race. Sam is approaching this in a very similar manner. I don't know if you have any thoughts on Open Air specifically before I moved to Anthropic Josh, but this was like pretty revealing for me. Yeah, it seems like as they move forward, their focus is on Stargate and building like actual data centers, their focuses on building more efficient code,

and then they're focuses on these hardware devices. And I guess these Johnny I have devices, the I.O. devices that are shipping in Paris, seems like the timelines are pretty tight and we might even get these before the end of this year. So I think those are probably the three pillars, right? It's just like the core software, the hardware and then the infrastructure. And notably, missing from this is roles for alignment and safety, which seem to be in the minority here, if I'm correct. But that's not the case for Anthropic. Is that a correct assessment? That is correct. But before we move on to Anthrobic,

Josh, there is 12% of job postings are in another sector. Can you guess which sector that is in open air? Huh. Hopefully marketing. They got to work on that. Think about a recent alliance or agreement that they had to sign with a big government entity. Oh, is this government work? Yeah.

12%. Whoa. Interesting. Do you want to know where their second largest employee base is? Where? Is this the Pentagon? Washington, D.C., dude. Oh my God.

Yeah. Wait, that's kind of crazy. Interesting. They've aggressively hired on the DC side, which kind of makes sense because Sam has been a huge fan of kind of trying to lobby government entities or just trying to lobby AI policies in general to become more favorable or AI favorable. So it makes sense that he has the largest presence out there. Anthropic specifically doesn't. But yeah, we can move into Anthropic now. Okay, yeah, that's interesting.

I had no idea. What is Anthropic working on? Okay. So Anthropic has, thankfully, a much more well laid out. website which can kind of fall through. And a lot of open roles. Everyone's saying AIS killing jobs, man, these companies are hiring. Did you hear the news? So with the folks listening, Josh and I are based on the East Coast. And in New York specifically, Anthropics currently about to buy one of the largest office spaces that any tech company has ever bought in the most

recent couple of years. They're about to have a huge presence out here. And they're going to be focusing on three specific roles. Number one is go to market, obviously sales. Number two, enterprise-specific, and number three, the finance bros. They are going wholeheartedly into the financial services sector, and they want every analyst, MD, and fund manager to be using Claude to direct all their different workflows. I don't know if you saw the news yesterday, Josh, but they just signed a $1.5 billion joint venture with Goldman Sachs, Blackstone, and I think it's Hellman and Freed, to basically integrate Claude into every single one of their portfolio. Follure companies, we're talking about like 250 plus portfolio companies for Blackstone alone.

And the idea is Anthropics is going to shove a bunch of anthropic engineers. These are costly people, the most valuable people, into these enterprises to build custom workflows just so that they get hooked on Claude, just so they get addicted on the LLMs. And then the rest kind of like pays for itself, I guess. Okay, first, where is this building that they're buying? I'm so curious. I think it's near the Google Office in Chelsea. It's like a massive space.

Yeah, nice. And I definitely see, I noticed the shift to move to financial services. I've seen this a lot with Claude recently. In fact, as we were recording this, they just released an update to a finance for Cloud, basically, where they're releasing a series of different templates that can do a lot of different things as it relates to financial analysis. So it does models, it does spreadsheets, it does financial projections, it does all the things that you would hope a analyst would do. Claude is building these as kind of pre-templatized workflows that are built. directly into co-work and Claude code. So if you want to get evaluation, if you want to get

trajectories of pipelines or revenues, all of that is now done within Claude. And I like the shift that we're kind of seeing into the financial services business. It's mostly been open since now. Perplexi is really the only one that's moved into financial services. They've tried that Bloomberg alternative. It's pretty good. But Claude moving into this, yeah, here's the post right here of all the things that they're doing. KYC screener, statement and auditor, it's got all of the tools, one-click plugin. So it makes sense that they move to New York, get a little bit closer to Wall Street, and start moving into financials. This is, I mean, this is exciting.

This is really cool, particularly when the market is ripping the way it is. Yeah, it's so interesting that like both companies, probably at the same time, have gone into becoming kind of like a sales and BD-driven business. I thought the LLMs would just speak for themselves. But the truth is, the technology and power is here. It's just not well enough dispersed. And it's not going to diffuse until you handhold the people. people are the blocker right now are humans. The blocker is silo data sets in all of these top

enterprises and companies that don't really know how to integrate the thing with AI and the AI is not smart enough to know how to integrate itself into it. So it's kind of like a chicken and an AI problem and Anthropic is taking the first step here. Open AI as well, they've signed a similar joint venture with similar companies to basically put both of their products at the helm here. If we go back to Anthroping in general, the point I want to make around these roles is, listen, they have around the similar amount of roles that Open AI is hiring for, but the strategy that they're pursuing here is a bit different. They are treating their company as an enterprise services company. So they famously or infamously have not bought as much compute as open airs.

As I mentioned, Sam is like the conqueror. He's like, I need all the compute. I'll lever it all up. And if it pays off, it pays off. And it is paying off because now he can serve his Claude Methos competitor, GBT 5.5 to anyone and everyone called Methos is still restricted on the Anthropic side. Anthropic is being more strategic. Their bread and butter, the majority of the money that they have made so far and still, I think it's something crazy like 70%, sorry, 60% of their revenue recently comes purely from enterprise and they want to clearly see that reputation and wrap it up even more. They are focusing primarily on enterprises, which is why we don't see more of a consumer base. You know, we've seen all the crazy social media spread around like restrictions on rate

limits and stuff like that. I think that's because they're just focusing hardly on enterprise customers. They don't want to restrict them. I think they have like, um, As of two weeks ago, they have like 500 to 1,000 clients that pay them well over one to five million dollars per year. If they can ramp that up, that's a lot of ARR, basically. And they recently hit, what is it, 44 billion error? Yeah, pretty good business. And you could start to see the clear divide between these two companies as we go through these job boards and the rumor mills where Open AI, they currently have the best image model. They're working on consumer hardware. They're building a lot for consumer,

although they're working on their kind of commercial enterprise segment as well, but Anthropic is all in on enterprise, on financial service. on selling to businesses. And I think that's why we've seen their ARR go completely nuclear, because they're selling much larger ticket items to much more sticky customers that are going to stick around and endure 12, 24-month contracts with these companies. So it's interesting, it's fascinating. The hardware story to me is the most exciting. I cannot wait to get AI-first hardware. I hope it's good. I hope it's good. Like WWDC from Apple is happening next month, where hopefully Siri doesn't suck. But the first AI-first devices are going to be from

Open AI, and now the timeline might have moved up a little bit sooner. So that news at least is very exciting. This has been an investigative review of what is in store. I think the prompt for you guys is, one, will you use an opening iPhone? Like, will you swap your iPhone for this phone? And I know it's tough to tell without actually having it in your hands to see, but it seems like such a stretch. You just, what about you? Any final parting thoughts before we head out for the day? I want to see the device. I want to see some visuals. You know why I feel this way? Like, don't even give me the device. it's that dime advert with Alexander Scars God or whatever, which I don't know,

turned out to be a fake or a dud or a Chinese pair of earbuds. I need something from Open AI. So bad. So bad. Well, hopefully we get that before we get, will we get a better Syria before we get an Open AI Harvard device? That's the question. Who knows?

Probably not. Probably not. But that is the episode for today. EJOS, thank you for all of this investigative journalism into the job boards here for giving us some guidance on all things opening eye all things anthropic if you enjoyed this video don't forget to share it with a friend who might also enjoy this who might also be interested in the crazy cool new hardware leaks that are coming if you enjoyed leave a comment down below uh you subscribe on youtube leave us a five-star

review on wherever you listen to your podcast and as always thank you so much for watching and we'll see you guys in the next episode