Strap In, Vision Pro Owners
Amazon + Anthropic • Talking Apple + Claude • Dun3 • Meta's Apple Poach • Risks for NVIDIA & Microsoft • Grok 4 • NVIDIA $4T
Apple is apparently gearing up to launch an updated Vision Pro later this year, per Mark Gurman of Bloomberg. The big upgrade? Not the new chip, which will only be mildly useful without much content that can actually push the device, but instead a new head strap. I'm not kidding. They better be selling this thing separately for current Vision Pro owners and not forcing people to buy a new $3,500 paperweight or there will be riots in the streets. Well, amongst the few dozen Vision Pro owners in any given area.
I kid, but I don't. Clearly such an update would simply be a message to the market that Apple isn't abandoning their plans in the face of slow sales and developer adoption. They're not ready to unveil a real upgrade yet – that will come in 2027 – but the spec bump and head strap is throwing the proverbial bone out there.
Obviously – obviously – the first Vision Pro should have been released as a dev kit as I wrote 15 months ago. I know it would not have been very Apple-like to do that, but Apple needs to start doing some un-Apple-like things, IMO. They could have leaned on the messaging that this future piece of technology wasn't quite ready for prime time, but developers could get a glimpse of the future today if they're willing to pony up. And if a consumer really wants to try it, well, sign up for a developer account and have at it. That would have flipped the narrative entirely here. And honestly, I'm not sure the messaging would have been that much different from the way Steve Jobs used to call Apple TV a "hobby".
All this new strap will do is further showcase how badly Apple messed up the ergonomics here. The fact that the Vision Pro shipped with two straps in the box already said all you needed to know. Now we'll apparently have a third option – after Apple implicitly endorsed other third-party options as well. The entire package is entirely too inconvenient for most people to even consider using on a regular basis. This better be one hell of a strap.
• Written on an M4 MacBook Air 💻
• Sent from London, England 🏴 Notebook
🤖 Amazon Going Deeper with Anthropic
This article is sort of framed oddly, suggesting that Amazon is "weighing another multibillion-dollar investment in Anthropic" when actually, we know from a 10-Q filing that they're due to invest "an additional $2.7 billion by Q4 2025". So I'm not sure what they're "weighing" here as this money was already committed in the form of convertible notes in different tranches. I suppose it's possible that they could have already invested the $2.7B between that filing (in May, for the quarter ended March) and now, but it's highly unlikely. So perhaps they're thinking about investing more on top of that commitment? Or maybe lines were just crossed and this reporting on that (already known) commitment. Regardless, one tidbit here is that the tech giants are capped "to keep them well below owning more than a third of Anthropic". Google currently owns 14% and my backtracking math has Amazon currently at 22.4% (before any new investment), so both have some room to grow if they want to put more money in (and undoubtedly insane valuations which will make growing their ownership stakes much from here difficult). One more wildcard: what if Apple were to invest? They pulled back from the OpenAI deal last year, but if they're really considering Claude as a Siri replacement/supplement... [FT 🔒]
🗣️ Apple/Claude, Meta Superintelligence, AI Browsers
I was back on Alex Kantrowitz's Big Technology Podcast to talk about all of this and more. And notably, we talk through the Apple/Claude potential partnership for Siri and yes, if it could portend an Apple investment into Anthropic... the startup is undoubtedly too expensive for Apple to buy at this point – nor would Amazon (or Google) want/allow that, but Anthropic needs money. Apple has money. Sidenote: I'll be back on Big Technology Podcast on a monthly cadence going forward. You can find it on Spotify here or Apple Podcasts here. [YouTube]
🪱 Dune: Part Three
Look, Dune Messiah would have been a cooler name but it also perhaps signals Denis Villeneuve's intent not to stick to that particular book and to go beyond it. At the same time, he talks about how he doesn't view this as a trilogy because he considers the first two to really be one film – so it's weird to call this one "Part Three", unless you consider this to also be a part of one (now really) long film? At least no one is going to be confused like they may have been with another Anya Taylor-Joy movie, Furiosa ("A Mad Max Saga"). Also, note the fact that it won't be shot entirely with IMAX cameras (like Christopher Nolan's The Odyssey will for the first time in a feature film) but clearly a lot of it will be. By the time he's filming his Bond, might he be ready to go all-in on IMAX? (Probably not – but again, certainly a good chunk of it!) [Variety]
💰 Meta Poaching Apple's Top LLM Guy
Mark Gurman and Riley Griffin were able to pull a few more details about the package it took to lure Ruoming Pang from Apple to Meta. "More than $200M" over a few years, largely in stock tied to performance of Meta's own stock (and probably some other metrics). But yes, there are likely large "signing bonuses" involved in these deals depending on how much equity the talent would be walking away from. In Pang's case, that's obviously just Apple RSUs and not startup equity, but you can see how such cash payouts can rise quickly. It all sounds in the range of a 'Godfather' offer. "Apple didn’t try to match the offer..." Yeah I mean the only question is if their counter would have been off by two or three orders of magnitude. [Bloomberg 🔒]
📈 The Risks for Microsoft and NVIDIA as They Near $4T
When the author wrote about the two companies racing one another to hit the milestone, I doubt he thought that NVIDIA would cross it in less than a week. Such is the state of AI exuberance. And while Microsoft has clearly ridden the wave as well – to the tune of their highest P/E ratio in 20 years – their risks and rewards are clearly less binary than they are with NVIDIA. They also have more levers to pull, which is what we may be seeing now with the layoffs even in a time of seeming strength as their revenue-per-employee is far below that of say, Google. Poor Apple, just literally not in this conversation right now. They very nearly hit $4T themselves last December, but now are nearly a full $1T behind their old friend/foe NVIDIA and a good $600B behind their old foe/foe Microsoft. The first company to hit $1T and $2T and $3T will likely be the 3rd to hit $4T – if they do at all (come on, they will, but it's rough out there right now!). All that said, if the AI market falters, NVIDIA and Microsoft are falling, fast. While Apple will likely be sitting pretty thanks to their inaction and ineptitude in AI. [WSJ 🔒]
Spyglass
💻 Escalated, the AI Browser Wars Have – Quickly
Right after Dia, Perplexity's Comet is here – with OpenAI inbound...
Loose Leaf
To me, the biggest news about Linda Yaccarino stepping down as CEO of Xitter is that Xitter still had a CEO. I honestly thought that when they merged with xAI that whole charade sort of went away. Kudos on lasting two years, I guess. That is about double what anyone guessed. [Axios]
OpenAI's first "open" model (weights, not source) could be here as soon as next week. And yes, it may add even more pressure to the Microsoft relationship – will it be on every cloud soon? [Verge]
As expected, Mistral is in talks to raise up to $1B, likely from MGX and others. Another deal that will make it harder for Apple to make a deal if they wanted to this path (if they even could)... [Bloomberg 🔒]
While there's no word on Mistral's valuation, SpaceX seems poised to hit $400B, which would put them close to ByteDance (pre-TikTok sale) in terms of highest private valuation. Well ahead of OpenAI. For now. [FT 🔒]
It's not just Meta that is poaching from OpenAI, OpenAI is also poaching from Meta – but really, mainly Elon companies with talent from Tesla and xAI. I'm sure OpenAI was devastated to have this news leak via internal Slack messages. [Wired]
We're getting a live-action Hot Wheels movie, thus moving us one step closer to the prophecy of the Kool-Aid Man busting through a brick wall on the silver screen. [Variety]
In other computing headwear news, Meta has formalized their investment in EssilorLuxottica – $3.5B into the publicly-traded firm for a stake just under 3%, which they hope to bring up to 5%. Most importantly, this seemingly formalizes the fact that Google/Samsung/Apple/Snap/etc won't be working with their brands on any competing products... [Bloomberg 🔒]
'Max' is now officially 'HBO Max' again with app icons being updated around the world. No need to download a new app this time (which was another amazing moment of stupidity in all this). And now the countdown is officially on until they're 'HBO' again. [Verge]
I Quote…
"I would expect Grok to discover new technologies that are actually useful no later than next year, and maybe end of this year. It might discover new physics next year… Let that sink in."
– Elon Musk, talking a big game (obviously as usual) with the release of Grok 4. Can xAI back up these claims better than Meta was able to back up Mark Zuckerberg's big claims about Llama 4? Or will Elon be next to step in with $500M pay packages and 'Godfather' offers?
(The early results seem promising for Grok 4 – but they also did for Llama 4! You can try it out yourself for the low-low price of $300/month.)
I Spy...
In July 2015, 10 years ago, NVIDIA's market cap was $10B. Yesterday, it hit $4T.