2026-05-01 17:06:32
In 2021, Apple spent $11.1B on CapEx for the year. In 2022, that number fell slightly to $10.7B. In 2023, it was back to $11B. By 2024, it was down a bit to $9.4B. 2025 saw a jump to $12.7B. 2026? Well, with half the fiscal year in the books as Apple just reported their Q2 numbers, Apple is on track to spend... $9B to $10B in CapEx.
One could imagine a banker on Wall Street conjuring Matthew McConaughey in Dazed & Confused: "That's what I love about Apple, man. Their Big Tech peers go crazy on CapEx, Apple keeps the spend the same."
The dichotomy is so wild that it now gets written about every single quarter. But the dichotomy also keeps growing every single quarter as Big Tech keeps ramping CapEx and yes, Apple stays the same! I mean the chart above says it all by showing it all.
In 2021, Apple brought up CapEx quite a bit from $7.3B in 2020 to that $11.1B number – a huge jump, relatively speaking. It looked like they might be closing the gap with their peer group as Meta was at $19.2B that year, Microsoft at $20B, and Google at $24.6B. (Amazon was an outlier back then – $61B! – due to their unique warehouse and data center needs given their core businesses.) But over the next few years, the story remained largely the same, with everyone holding fairly steady (aside from Meta, which ramped CapEx in 2022 – but pulled back the following year). Then 2024 hit.
By then, it was obviously clear to everyone that ChatGPT was not a flash in the pan – well everyone except Apple, perhaps. Acting like the Eye of Sauron realizing the Ring of Power was nearby, the rest of Big Tech firmly fixed their gaze on CapEx. Last year, the numbers exploded. This year, they'll explode even further. Amazon will hit $200B spent for the year. Google will be close behind at $190B. Microsoft should be around the same at $190B. Meta at $145B. Apple? Again, they hit $4.344B in CapEx in the first half of 2026 – which was down a bit year on year – so they should end in that $9B to $10B range, assuming some level of ramp. Yes, that would be below what they spent in 2021.
It's just such a wild break from their peer group. And it keeps getting more wild. It seems like the most binary bet imaginable. Either Apple is right and the rest of Big Tech will have lit hundreds of billions – perhaps trillions when all is said and done – of dollars on fire, or Apple is going to be in big trouble.
Obviously, there will be some nuance there. Apple's bet relies on a few factors, including that someone else is spending this CapEx on which Apple can rely. This is the "hybrid" approach to such spend that Tim Cook and others at Apple keep referring to. They're basically saying that while they'll do some of their AI work in-house, on their own servers, for the most part, they'll partner with the others to rely upon their capacity. Most famously, Apple is now doing this with Google.
To that end, they may view it as similar to their longstanding – even after legal challenges – deal for Google Search. Apple could have spend tens or hundreds of billions building their own search engine. But they opted to partner with Google and instead got paid tens or hundreds of billions! And how well did that spend versus partner strategy work for Microsoft? They're making some money from Bing now, but it took years and years – and billions upon billions spent – to get to this point. And they're hardly making Google "dance" despite Satya Nadella's promises.
So yeah, that would be a best-case-scenario for Apple. The problem is that no one else thinks AI is going to play out the same way – clearly, or they wouldn't be spending these billions! At the very least, they view right now as a greenfield opportunity to own AI in a similar manner to how Google took control of Search. But more likely, they think AI is going to be a core bit of technology that they each need to control, lest they be beholden to someone else. This is exactly why Microsoft shoved OpenAI aside despite their early extremely prescient bet on the startup that remains the leader in the field. Even though Microsoft owns 25%+ of that company, with IP rights and access, they're now spending the billions to build out AI on their own.
What's especially wild there is that Apple is most famously the company that doesn't want to be beholden to anyone else if at all possible. It's the "Tim Cook Doctrine" for chrissakes! Either they've forgotten that fear, which stems from the times Apple nearly died in their history when others refused to play ball with them, and have been lulled to complacency by years of iPhone dominance, or again, they just think this will be like web search. Not something they need to fully own.
Or at least not yet. The other possibility here is that they view this AI arms race as nutty right now because the technology continues to evolve so fast and it's possible that their peers spending their billions just ends with them all more or less on equal footing as the technology stabilizes and perhaps commoditizes. That has not been the case thus far, which is why everyone makes fun of Apple's AI stumbles. But it doesn't mean it won't be the case eventually.
At the same time, Apple can be like the tortoise racing the hares. Slow and steady with their AI work to try to win the race in the end, years down the line, when the technology is finally mature enough to matter for everything. The problem there might be the fact that they won't have the muscle memory from years spend in said AI race. Maybe they can hire/acquire their way to the top then, but it's a massive risk, to say the least.
But it's the one Apple is clearly willing to take! Certainly it's the one Tim Cook has been taking over these past many years. We'll see if John Ternus thinks differently here, but all indications are that he will not deviate from this strategy any time soon. We could very well be in a situation where in 2027, Google and Amazon are spending $300B on CapEx and Apple is spending... around $10B.
One more thing: Apple is ramping up their R&D spend, which you have to believe is also related to AI work, at least tangentially. In fact, those numbers are at all time highs for Apple. Of course, their peers are also ramping such spend too...
M.G. Siegler
M.G. Siegler
M.G. Siegler
2026-04-30 18:18:56
Forgive me for getting pedantic, but I can't be the only one who feels like they're taking crazy pills here. Yesterday, OpenAI published a blog post declaring mission: accomplished in what they said was a key goal for their 'Project Stargate' initiative: 10GW of AI infrastructure. It's a big number, to be sure. And they note they've achieved it well ahead of their 2029 goal. Impressive.
The only problem? As best I can tell, they backed into this goal by not just shifting the goalposts, but by completely changing the field on which they play. And perhaps the sport too...
2026-04-30 02:49:45
Happy Earningspocalypse to those who celebrate. Amazon, Google, Meta, and Microsoft all report today after the market closes. Obviously, all eyes will be on AI spend – and any signal of actual returns from spend to date.
How high will the CapEx numbers go this time?
Inklings is a newsletter featuring links and commentary from M.G. Siegler on timely topics found around the web.
💫 OpenAI Missing Projections – On one hand, you could read a startup missing internal projections as the company either not being great at extrapolation (more understandable during fast growth) or setting targets that are too optimistic (perhaps to help woo investors). These are internal numbers for a reason; they aren't held to the same level of scrutiny as, say, those of a public company. On the other hand, OpenAI is an awfully mature company at this point having raised more money than any other in history, with an experienced team, and one that is obviously trying to go public sometime soon. Plus, as far as we know, they haven't missed on such projections in the past, so... It's not a great look, but it's made worse by their response which includes the highly scientific measurements such as "firing on all cylinders" while calling the report "prime clickbait" – both of which, of course, are non-denial denials to a report, which has now been backed-up. Even the partners, whose stocks are getting hit by proxy at the moment can't seem to align on the right rebuttals here. The actual actions of the company perhaps remain more indicative here – killing "side quests", altering deals to expand their market, and yes, bringing on some help in the marketing/PR front. [WSJ 🔒]
💨 Amazon's 'Quick' Superagent – The branding for their desktop-based AI assistant isn't bad, but confusing. 'Amazon Quick' definitely sounds like a fast delivery offering, not their version of Claude Cowork. Then again, 'Q' didn't really work for them – and was weird for other reasons – so I guess just add some letters? More importantly, Amazon doesn't have the best track record when it comes to software. Almost all of it whether on the web, desktop, or mobile, quite frankly, sucks. So the hope will have to be that their approach – multiple models, obviously their own 'Nova' models and presumably Claude, but presumably not GPT, at least not yet – and their AWS connections will win the day. For $20/month, it better be fast. [Information 🔒]
Cost-Per-Click AI Ads – It feels like OpenAI has little choice but to go down this path already for ChatGPT – they need to start showing results to advertisers and this is the way you traditionally show results. But I'm still quite skeptical this model will work all that well in AI. There's a very fundamental problem: such AI services, at least to date, have removed most of the need to click. That's how they've been disrupting Google, of course. With web search, users are trained to click. With AI, users are trained to not need to click. So, yeah. Good luck getting to that $100B+ advertising business – which would be the overall biggest driver of the business at that point – in a few years this way. I just think they'll have to find a new model and methods for advertising in AI. Just like Google had to do for Search and Meta had to do for the feed. [Digiday]
"Stargate has now had three or four permutations, I don’t know what it is right now. I can’t tell you. Maybe it never really existed in the first place."
– An unnamed "someone familiar with Microsoft's thinking" which, of course, sounds a lot like someone who works at Microsoft and is undoubtedly chuckling right now over "Project Stargate". As it, of course, actually started life as an initiative between OpenAI and Microsoft, before "The Blip" derailed, well, pretty much everything between the two.
Soon, Stargate was reborn/repackaged as a $500B mega-project unveiled at The White House between OpenAI, SoftBank, and Oracle. Why? Because obviously OpenAI as a massively unprofitable startup couldn't raise the debt required to finance such a project on their own. Yet they probably needed to in order to ultimately compete with Google.
But, well, the best laid plans and all that – which, to be clear, this was not. Now we have Microsoft picking off Stargate scraps and "Stargate" itself morphed into just a catch-all PR term for OpenAI data center deals. "I don't know what 'Stargate' means at this point," says another "involved" buy unnamed source. So say we all.
Below, members of The Inner Ring will find thoughts on:
• Meta's Reverse Hackquisition
• Microsoft & OpenAI's New Deal
• and more...
2026-04-28 19:06:37
Back in 2012, I wrote a post comparing Apple's method of product development to the opening (and closing) of Christopher Nolan's 2006 film, The Prestige. The opening dialogue of that film explains the title:
Every great magic trick consists of three parts or acts. The first part is called “The Pledge”. The magician shows you something ordinary: a deck of cards, a bird, or a man. He shows you this object. Perhaps he asks you to inspect it to see if it is indeed real, unaltered, normal. But of course… it probably isn’t. The second act is called “The Turn”. The magician takes the ordinary something and makes it do something extraordinary. Now you’re looking for the secret… but you won’t find it, because of course you’re not really looking. You don’t really want to know. You want to be fooled. But you wouldn’t clap yet. Because making something disappear isn’t enough; you have to bring it back. That’s why every magic trick has a third act, the hardest part, the part we call “The Prestige”.
My post was entitled, "Apple’s Magic Is In The Turn, Not The Prestige" and it was specifically about how while some viewed the unveiling of the iPhone 5 at the time as boring, they were missing the point. The point, per my title, was "the turn". How Apple does what it does to transform the ordinary, a smartphone, into something extraordinary, the iPhone.
Here I am, nearly 14 years later, thinking about that movie and that point again with regard to Apple.1 But this time, the context is a bit different. If you'll forgive the turn of phrase, the key this time isn't "the turn" but yes, the Ternus. And to take it a step further, while John Ternus' focus has been on "the turn" over his nearly 25 years at Apple, ultimately the key, as he takes the reigns as CEO, may actually be in "the prestige"...
I already wrote up my thoughts about Tim Cook and his tenure at Apple – including a few months ago when the news of his stepping down as CEO was just rumored, but seemed likely, soon. Now my thoughts turn to Ternus, and his path going forward – which will be Apple's path going forward.
The reality is that a lot of that path, just as it was for Cook when he took over for founder Steve Jobs in 2011, is already locked-and-loaded. All companies working with hardware have to have roadmaps that are set years ahead of time. And development cycles require product work to start at least a couple years in advance. With Apple, that's often even longer because of their complex supply chain and early access to new technologies. That is to say, Apple's pipeline of products are probably locked in for the next couple of years, at least.
Some may slip, of course – as has been the case with Apple in particular over the past few years as AI has upended not only their software, but their overall product strategy. But for the most part, we know what we're going to get this Fall and next Fall too – and a lot of that is obvious even without leaks. It's the years after that where things start to get more interesting as it relates to Ternus. Because he will start making calls this Fall when he takes over that will set those wheels in motion...
Of course, leading up to his ascension, it sure seems like Ternus already has been doing that, at least a bit. With the MacBook Neo, Ternus got a lot of credit – and, of course, "stage" time – for ushering in that product. While it seems obvious with the benefit of hindsight that it would be a huge hit, it was also quite antithetical of Apple's normal approach, certainly with regard to the Mac lineup.
That's why I immediately framed it as the smartest announcement Apple had made in a long time. It signaled a sea change when it came to their strategy. For the past many years I had been worried that Apple had become completely calcified when it came to changing their ways – which seemed like it could be particularly problematic in the Age of AI, where everything is moving at light speed. That obviously proved to be true, but shakeup of the AI divisions aside, it was this MacBook Neo that signaled the possibility of real change for Apple.
The company has famously been run by leadership that for the most part had been there for decades. This type of stability obviously has huge benefits, but also some pretty big downsides, as I believe we saw play out with AI. And Ternus was in this boat too as someone who has been at Apple for that quarter century – nearly half his life!
Steve Jobs may have famously told Tim Cook not to focus on what he would do, just to do what's right (for Apple, when he took over), but the truth is that Apple has spent the past 15 years of Cook's tenure largely doing what it thought Jobs would do anyway. And for the most part, this was the right call! Stock prices may deceive, but sales figures do not. But now that we're on to the next generation with Ternus and a leadership team that is naturally going to turn over as many of them are at or near retirement age, there truly is an opportunity to "think different".
Ternus will say all the right things right now about staying the course that Apple is already on. But again, he has little choice at the moment, that path is pretty much set. Sure, there will be choices here and there – in particular with partnerships around AI and elsewhere – but the product pipeline is pretty locked. The question is really what Ternus chooses to set in motion for 2028 and beyond.
Again, a lot of that is in process too, but in such a way that things can and do get tweaked or prioritized or cancelled. And new projects can get started depending how Ternus and his team read the tea leaves.
Cook, famously, had a bit of a rough time in this regard. Yes, there's the Apple Watch – but per above, that project was undoubtedly actually kicked off while Jobs was still alive. AirPods were likely truly post-Jobs, but I mean, they are EarPods without wires. They're great! But not exactly an entirely new product category. The car project was probably more nuanced, as the Apple Board was said to have been discussing it when Jobs was still around, but "Project Titan" definitely started under Cook. Of course, it never really went anywhere! And then there's the Vision Pro. Certain parts and technologies may have dated to Jobs, but it was clearly a Cook package. And well, it's perhaps the most to-be-determined product Apple has ever launched. It hasn't looked good to date, but they could morph it over time into something solid, not unlike the Apple Watch. We'll see.
Really, the only slam dunk Cook can take credit for is Services (well, and his Chinese supply chain, if you want to consider that a product!).2 Yes, some of those – notably iTunes and the App Store – kicked off under Jobs too. But it was Cook who put most of the wheels in motion (thanks largely to Eddy Cue) to get the division to where it is today – to the point where it will someday overtake even the iPhone itself in terms of revenue.
Anyway, my point is that while Cook couldn't exactly "coast", he could coast far more than Ternus will likely be able to. Especially as the aforementioned wave of AI threatens to upend every corner of technology.
In the near term, I believe Ternus and Apple are going to be in a great position with the iPhone because it's still the most-used and best device. That means that it's going to be the most-used and best device for AI too. But it will be under assault in a way it hasn't since its earliest days. Both with new AI hardware coming and, I imagine, new smartphone entrants built more specifically for this moment.
But even if Apple remains sitting pretty with the iPhone, they cannot afford to be sitting still. They still need to be funneling the proceeds from the device into the What's Next™ product. This is obvious, but again, Apple hasn't really been able to do this post-Jobs. They have a great satellite constellation of products, but nothing that moved Apple from the Mac business to the iPod business to the iPhone business, as Jobs did. Nearly 20 years later – Cook's entire tenure – and Apple is still the iPhone business.
Again, it's a great business to be in! But even it won't last forever. It still very well could last as the key business driver throughout Ternus' tenure too, but I mean, just assuming he gives the same 15 years of service, that will take us to 2041. Yes, the years are short, but the decades are long in tech. That's a long ways away. Especially when you think about it on the timescale of AI.
So what is Ternus to do? I don't know, that's above my pay grade. But at a high level, I do think he needs at least an iPad-level new hit product. And perhaps an iPod-level one in terms of relative game-change. I'm not saying Apple will die without one, but I am saying Apple will likely be seen as dated and in decline without one.
I do think the 'iPhone Ultra' – the artist formerly rumored to be the 'iPhone Fold' – will be a hit for Apple. But at the end of the day, it's an iPhone. The smart glasses? Certainly Apple could make them work, and perhaps piss off Meta in the process. But are they going to be a massive stand-alone product? I suspect that even in the success state they'll be viewed as one of the iPhone "satellite" products. Ditto with AirPods with cameras and any sort of AI pendants, obviously.
What's the next truly new product? Maybe full AR glasses? Perhaps taking cues from both smartglasses from below and the Vision Pro from above? You could certainly see that. But there's also a lot to be determined, both technologically and societally between now and then.
And wouldn't it be more fun for Ternus to truly come out of left field with something? Again, like the iPod. What category can Apple zero in on to truly revolutionize? As much as I still might like a true Apple Television set, that ain't it.
Robotics is probably the most obvious path here. It plays both to Ternus' hardware background and to our AI future. And beyond the rumored table top devices, Apple already has public research products that look... well, awesome. Very Apple, but also very Apple-of-the-future. If Apple can make Wall-E a reality, can you think of a better love letter to Steve Jobs?3
Apple looking around the landscape and taking what Tesla has (at some point) done with Optimus, and others have done and swooping in to do it "right" would be the Apple playbook, of course. Is that feasible by 2040? Again, we'll see...
Regardless, it feels like it's time for Apple to take some risks again.
Lastly, I would just note that while Ternus' focus and strength has been on "the turn", he will need to learn a new trick to cement his legacy atop Apple. Again, it's something Cook never quite nailed. The Prestige.
Cook as a presenter was fine. He got better as time went on, but that was also because Apple made it easier for him as time went on by taking away the live studio audience! I mean, to be fair, the pandemic really did that. But Apple, unlike their peers, never brought it back. They need to. Ternus needs to. He needs to tap into the energy of Apple's ecosystem again with live events. I suspect he will – perhaps even with his first iPhone event in the fall.
Ternus is obviously no stranger to MC'ing such events – again, digitally. Beyond the aforementioned MacBook Neo unveil, he's actually been "on stage" for years now – to the point where I called his continued presence out as interesting even before the rumors that he might be the next CEO started... But he should aim to forge his own type of presence on stage, unveiling new products.
I'm not saying he has to be Steve Jobs. It remains too tall and order, and really, he should want to be his own person in this – again, this is about Apple's future, not about the past. Others like Evan Spiegel at Snap and Jensen Huang at NVIDIA show a potential path here. Ternus, while an engineer, is not nearly as stilted as Mark Zuckerberg has been (though also obviously better over time), so he has a real shot at pulling off "the prestige" at such events.
While the key to Apple remains "the turn", Ternus nailing "the prestige" would give everyone, from the employee base to the user base to Wall Street, confidence in the path forward. As I see it, Ternus' tall task is twofold: he needs both a new trick and to reveal it, as well as Apple's old tricks, in a way that conveys the true magic of the products. As The Prestige taught us, "You have to bring it back."
1 That movie is perhaps also top of mind as I'm about to record a podcast about Christopher Nolan's film as it approaches its 20th anniversary... ↩
2 Maybe Apple Silicon, but that really launched under Jobs with the 'A4' chip in the original iPad. ↩
3 I mean, hopefully without the destroyed world and morbidly obese human population that's unable to move from their seats. But you know, the optimistic version of that. ↩
2026-04-28 03:36:07
"The Clause" is dead. Long live, the clause.
That's the easy read of this out-of-the-blue update to the agreement between OpenAI and Microsoft. I say "out-of-the-blue" because the two sides had just spent months hammering out an updated agreement that was finalized just six months ago. This is an update to that update.
Of course, it wasn't really out-of-the-blue. This update is all about letting OpenAI offer their models on AWS. For proof of that, look no further than Amazon CEO Andy Jassy tweeting, "Very interesting announcement from OpenAI this morning." As he notes, there's an AWS event in San Francisco tomorrow. So yeah, they wanted to get this squared away before that.
Why would Microsoft agree to that? Clearly, OpenAI finally agreed to kill "The Clause". Long a main point in the relationship between the two, it basically gave OpenAI a position of power over Microsoft because it allowed them to pull back their AI if and when they achieved "AGI". Of course, this was also a main sticking point in that aforementioned last update to their agreement, but while Microsoft successfully defanged it at that point – notably by ensuring OpenAI's nonprofit board could no longer unilaterally declare AGI achieved, but also by maintaining rights to some models and technology even after AGI is achieved – it didn't fully go away. Now it's going away. It's simply replaced by a date: 2032.
That date was also a part of the last negotiations, but this makes it cleaner. It's hard to tell if it's some sort of acknowledgement by OpenAI that AGI might not be here by then, or that Microsoft thinks it may be here sooner – or that it really no longer matters either way, at least for business purposes. Regardless, what matters here is OpenAI having the ability to offer up their products on other clouds. This is one big way in which Anthropic has started to eat their lunch. Now fresh clouds are back on the menu.
Probably. Sort of. Maybe.
The weirdest part of the announcement – which is a mere 242 words, and because there can't be any sort of announcement between these two without a weird part, it's just the rule – is the following:
Microsoft remains OpenAI’s primary cloud partner, and OpenAI products will ship first on Azure, unless Microsoft cannot and chooses not to support the necessary capabilities. OpenAI can now serve all its products to customers across any cloud provider.
That's seemingly a pretty important "and" there. In other words, at least in this telling, Microsoft still has a right of first refusal of sorts here. It's not enough for them to not be able to support something that OpenAI wants to do with some new offering on another cloud provider, Microsoft has to explicitly say they will not support it, making it seemingly okay if OpenAI looks elsewhere.
In practice, this probably means that Microsoft will continue to get first look rights to all of OpenAI's models in Azure, but over time, Microsoft will be okay with those models trickling down elsewhere.
I can't wait to see what trips up this new clause! Because you just know something will. OpenAI will work on something with Google or Amazon that they'll say Microsoft can't technically offer – perhaps around TPUs and/or Trainium chips? – and Microsoft will get pissed. Again.
But for now, clearly Microsoft signed off on OpenAI's deal with Amazon here – otherwise we wouldn't get that Jassy tweet – though kind of a dick move not to acknowledge Microsoft paving the way here? Then again, Microsoft had been threatening to sue over the new OpenAI/Amazon deal, believing it violated their exclusive OpenAI rights. So...
It's still somewhat surprising that Microsoft is giving in here seeing as it's Amazon, which controls the only cloud bigger than their own. But it's also sort of a win/win. Beyond the aforementioned removal of The Clause, Microsoft still owns upwards of 27% of OpenAI (perhaps more like 25% after the dilution of the last mega round). OpenAI getting access to AWS definitely boosts the value of OpenAI because it should massively boost the business of OpenAI. Again, not only against Anthropic, but overall as the company tries to go public.
For all the bickering between OpenAI and Microsoft over their deal for years now, Microsoft would rather own their percentage of a larger number rather than a smaller one. On some level, it's just math. Yes, even between these two star-crossed lovers.
And yes, the two will have to keep working together – at least for the next 6 years or so.1 Though you have to believe Microsoft will appreciate even more distance from their old bedfellow. Especially as they work towards AGI/Superintelligence fully on their own. And they'll certainly enjoy not having to send back a 20% cut to OpenAI any longer – and you can imagine Wall Street will like that too, ahead of Microsoft's earnings later this week.
OpenAI will still be paying their 20% cut for sales of services hosted on Azure, but they get to pull that end date forward a couple years, to 2030. [see: update] And it's also now apparently capped at some level, but they're not sharing what that level is. This is a smaller win for OpenAI which had wanted to bring this rev share down, and it will be a better story for the IPO roadshow.
Anyway, I'm no divorce lawyer, but that's my read. While OpenAI was somewhat free to see other people under their last updated agreement, this gives them much more freedom (provided Microsoft doesn't fall back in love with OpenAI and/or isn't spiteful of their new relationship potential – famous last clause). And Microsoft, in turn, gets to rip-up that pesky AGI prenup once and for all.
One more thing: while this actually doesn't seem to have anything to do with the trial between Elon Musk and OpenAI/Microsoft, it's interesting timing giving that it just kicked off today. Might some of this come up? In particular, the AGI bit?
Update April 29, 2026: Per a new report by Aaron Holmes and Sri Muppidi for The Information on the backstory in these negotiations, apparently the 20% revenue cut OpenAI shares back to Microsoft may come from overall revenue, whether or not Azure is involved (so yes, money made via Amazon, Google, etc). That's sort of wild (and presumably the reason for the cap) – and great for Microsoft!
M.G. Siegler
M.G. Siegler
M.G. Siegler
M.G. Siegler
M.G. Siegler
1 I think we can easily predict when Microsoft may dump their stake in OpenAI – though they might do so in chunks before then depending how the IPO goes! ↩
2026-04-27 22:00:07
Just when you thought OpenAI and Microsoft couldn't possibly further consciously uncouple, they find a way, it seems. This is just breaking as I publish this – and there's seemingly some weird/vague wording in here – so I'll have more thoughts later. But it definitely seems like something OpenAI wanted – to be able to fully date other clouds. While Microsoft gets even more distance and to keep more money.
Weird timing too with the trial in which OpenAI and Microsoft are defendants vs. Elon Musk just kicking off (more below). Perhaps related?
Update: some thoughts!
M.G. Siegler
Inklings is a newsletter featuring links and commentary from M.G. Siegler on timely topics found around the web.
💰 Google's (Up To) $40B Anthropic Investment – You know how I know that Anthropic is freaking out over being compute-constrained? In the past week, they've now agreed to sell (up to) $65B worth of equity to the investors which are already their two largest backers in Amazon and Google. And they're doing it – well, at least the first tranches – at the price of the last funding round, even though they have offers to raise at far higher valuations already. This will mean Google has around $13B invested into Anthropic now, and up to $33B if milestones are hit. Amazon, meanwhile, also has $13B invested now, and up to... $33B. These numbers align so nicely that they don't seem like coincidences. But then again, Amazon has more money at lower entry prices, so their ownership stake should still be higher than Google's. This would buy Amazon another 1.3% now and Google another 2.5%. [NYT]
📱 The OpenAiPhone? – First, are we sure this isn't a chip meant for another piece of OpenAI hardware? Obviously, Ming-Chi Kuo would check on that but still, the report is weird enough. Not because the idea is bad – as I've stated before, of course Amazon and everyone else would still want to control their own smartphone as it means controlling their own destiny. (This is why I suspect Meta, the most annoyed about the Apple/Google stranglehold in the space, will try again at some point.) OpenAI, on the verge of their own hardware, would obviously want such control for the same reasons. The difference, of course, is that they're a "startup", one that's trying to focus after years of "side quests". A smartphone would be the ultimate – and ultimately expensive – side quest. Maybe it's just something to explore if the true AI device – undoubtedly tethered to the iPhone – is a massive hit? Mainly I want to ensure we see that device, with so many billions yet to be burned on just model training and inference alone. [@mingchikuo]
💬 XChat – While I think I was asking for a stand-alone Twitter DM app in the past, that was at least a dozen years ago. I don't really understand why it exists now. There are way too many chat apps and this is seemingly just a carbon-copy of direct messages already within Xitter, but more buggy (I guess more iOS-native). Further, Elon has long talked up the notion of X as the "everything app" which seems more complicated when you need three or four apps for the everything app. Speaking of, it sounds like the X Money app is coming shortly too – though perhaps "a day late and dollar short". Zero surprise there. Porn-y names aside, this is like watching Meta's obsession with bundling and unbundling their apps, but in slow motion. Yes, XChat app hit #1 in the App Store for a day or so, but obviously you can do that on the back of the parent app with millions of users. The fact that it's fallen into the 20s just a couple days later seems like a bad sign. Who is this app for? Even the security folks are skeptical of that angle. As is Grok! [TechCrunch]
Below, members of The Inner Ring will find thoughts on:
• Elon vs. OpenAI, the Trial
• China Blocks Meta's Manus Deal
• and more...