AI: Surveying mid-year AI Options for Mag 7 and beyond. RTZ #769
...Meta and Apple current AI book-ends
It mid-point in this frenzied and inconclusive AI year. And it feels like a few minutes to closing time. With the people at the bar are fast surveying their options. Recalibrating their filters and choices.
The AI Tech Wave is accelerating in its hardware and software exponential bottom up Scaling from ‘first principles’ technology perspectives. Both in the US and China. Especially as both are hoping to dip in the same global AI talent pool where half of them are from China. Directly and indirectly.
Amongst the Mag 7, Meta seems the most eager and anxious to sort out and makes it choices. Apple the most relatively relaxed. From the outside.
Microsoft ought to be more anxious than Meta given that the public markets have discounted more AI goodness for its stock than most other Mag 7. Especially given their relative non-legal leverage in their ‘partnership’ with OpenAI. OpenAI has all the AI IP cards. Microsoft is not even in the top one hundred without them. Regardless of their ‘above them, below them, around them’ legal agreements. Courts do not move on Internet time, let alone AI time.
Amazon and Nvidia are the most relatively secure in the middle. Amazon with its market leading Cloud data center position, now increasingly AI enabled.
Nvidia with its solid position as the purveyor of AI GPUs and Data Center infrastructure of choice to all the combatants. Lubricated by copious amounts of open source CUDA and other growing AI libraries. Increasingly covering Physical AI domains like robots, autos, and local AI processing.
Google in my view is also in good shape internally, despite external investor anxieties (more on this in a bit).
That leaves the runt of the litter, xAI/Grok/Tesla, with the energy of the fiercest pit bull in Elon Musk.
The one with the biggest market cap premiumsrelative to its auto peers, and its tech/AI peers. Markets for now love the meme-fied tail winds of its Robotaxi, Optimus robot dreams. Customers painting Elon ‘Tom Sawyer’ Musk’s white fence still on social media and their patronage.
The two of the seven most in the news this week of course have been the ones I started this essay with above: Meta and Apple.
Meta’s founder/CEO just outlined his revamped AI team and strategy bolstered with billions in fresh and existing capital.
As Bloomberg outlines in “Zuckerberg Debuts Meta ‘Superintelligence’ Group, More Hires”:
“Meta CEO Mark Zuckerberg announced a major restructuring of the company's artificial intelligence group, including a commitment to developing AI "superintelligence" that can complete tasks as well as or better than humans.”
“The new group, called Meta Superintelligence Labs (MSL), will be led by Alexandr Wang, former CEO of Scale AI, and will include existing teams focused on large language models, AI products, and Fundamental AI Research.”
“Zuckerberg plans to spend "hundreds of billions" on AI projects and research in the years to come, and has hired 11 new researchers and engineers from companies like OpenAI, Anthropic, and Google to join the MSL team.”
As John Hammond in Jurassic Park would say, Zuck has ‘spared no expense’.
Besides $70 plus billion a year in AI data center and related capex, it’s billions more on ramping up the AI Talent:
“Meta will spend “hundreds of billions” on AI projects and research in the years to come, Zuckerberg has said, though the Facebook founder also expects that many firms will likely overspend on AI in an effort to avoid missing the wave.”
The urgency comes from the risk of ‘underspending’ vs the potential losses of ‘overspending’:
“There’s a meaningful chance that a lot of the companies are overbuilding now,” he said last summer. “But on the flip side, I actually think all the companies that are investing are making a rational decision, because the downside of being behind is that you’re out of position for, like, the most important technology for the next 10 to 15 years.”
And the ‘poaching’ is a unique version of Zuck’s ‘Build vs Buy’ strategy, which he has effectively used already to date to build the Meta of today:
“In addition to the newly structured team, Zuckerberg announced 11 new hires for the group, including researchers and software engineers from OpenAI, Anthropic and Google. The group includes former DeepMind researchers Jack Rae and Pei Sun; several OpenAI researchers, including Jiahui Yu, Shuchao Bi, Shengjia Zhao and Hongyu Ren; and Anthropic’s Joel Pobar, a software engineer who previously worked at Meta for over a decade. Bloomberg previously reported some of these hires.”
Apple on the other hand is taking a relatively more relaxed approach on the other side of the Mag 7 spectrum.
Again, Bloomberg reports separately in “Apple Weighs Using Anthropic or OpenAI to Power Siri in Major Reversal”:
“Apple is considering using artificial intelligence technology from Anthropic PBC or OpenAI to power a new version of Siri, potentially sidelining its own in-house models.
The company has talked with both Anthropic and OpenAI about using their large language models for Siri, and has asked them to train versions of their models that could run on Apple's cloud infrastructure for testing.
If Apple moves forward with using third-party models, it would represent a significant reversal and an acknowledgment that the company is struggling to compete in generative AI, and could allow Apple to offer Siri features on par with AI assistants on Android phones.”
The article also notes that the company has considered acquisitions like Perplexity, that I wrote about a few days ago. The whole Bloomberg piece is worth reading, but it’s most notable for the company NOT mentioned, Google.
Yes, the Alphabet/Google that currently pays Apple over $20 billion a year to keep Google Search as the default across Apple’s 2.4 billion plus users of iPhones/Macs/iPads and other devices for their daily surfing habits. And hopes to strike a deal extending it to Gemni AI Search, as expressed by CEO Sundar Pichai.
And yes, the Google that I mentioned briefly on their AI positioning within the Mag 7 discussion above.
The Google in a relatively great place in this AI Tech Wave ‘race’ as I’ve said for some time. It’s the one in the strongest place up and down the AI Tech Stack chart below.
From its own TPU processors, to its own half a dozen plus applications serving bilions directly for their Chrome browsing needs, Gmail email needs, Productivity Apps needs with Docs, Drive and more, Entertainment needs with its YouTube global goliath, and much more. All tied up with an industry strong and leading AI Gemini family of models, large, medium and small.
Axios summarizes it all well in “Tech giants play musical chairs with models”:
“There are five consumer-tech giants — but only three leading AI foundation models.”
“Why it matters: When the dealmaking music stops, someone's going to be left out.”
“Driving the news: Apple is talking with both Anthropic and OpenAI about using their foundation models to power Siri, after in-house efforts to upgrade Apple's voice assistant have faltered, Bloomberg reported yesterday.”
They summarize the landscape well at 50,000 plus feet:
“The big picture: Every big player in tech is working on their own foundation models — the biggest and most ambitious large language models that fuel ChatGPT and all the other services at the heart of the generative AI revolution.”
“OpenAI, Anthropic and Google seized the high ground early and have stayed ahead of the pack, both in scale, innovative advances and subtle refinements.”
“Important runners-up range from Elon Musk's xAI to France's Mistral and China's DeepSeek.”
All items covered broadly and deeply here at AI: Reset to Zero. And the accelerating AI innovations and iterations still at work:
“Most of tech's five trillion-dollar giants already have a match in the foundation-model game, but there's constant movement, and the music is still playing.
Google is the only one of the giants that has built its own top-tier model.”
Its takeaway on Google that I wholeheartedly agree with.
Going on to where Microsoft finds itself today after an iconic start:
“Microsoft tied its future to OpenAI early in the game with a gigantic investment and a commitment to deploying OpenAI models to the vast installed base of Microsoft users.”
“But the alliance has frayed, and the two companies are locked in high-stakes negotiations for an amicable breakup.”
“Microsoft also has its own in-house foundation model project, but that has yet to surface.”
And then onto Meta’s current AI frenzy:
“Meta got a later start and bet heavily on an open-source strategy with its Llama model family.”
“But disappointment in the progress made by the most recent Llama flagship model has forced a rethink, per the New York Times.”
“The Times reported last week that Meta execs had discussed "de-investing" in Llama, though the company denies that.”
“Meanwhile, CEO Mark Zuckerberg has poached a number of OpenAI researchers and added ScaleAI founder Alexandr Wang and former GitHub CEO Nat Friedman to his roster.”
“Zuckerberg announced yesterday these new hires would lead a unit called Meta Superintelligence Labs that will bring together all of Meta's foundation model work.”
Their take on Amazon and AWS (Amazon Web Services) to date:
“Amazon has invested in its own families of models to offer its cloud customers, chiefly Nova and Titan.”
“But for its effort to improve the popular but aging Alexa voice assistant, Amazon found that Nova alone couldn't handle the job and pulled Anthropic's Claude in as well.”
“Amazon has also made multiple investments in Anthropic.”
Coming back to Apple, where I disagree strongly with their opening statement:
“All this means that Apple's choices are limited.”
But with some possible reasons why Apple may not continue with Google, beyond the current Google/Search Legal Remedies legal case:
“Apple wouldn't turn to Google — not only because both companies are defending giant antitrust lawsuits that might make a deal perilous, but also for historical-cultural reasons.”
“Apple never forgave Google for building Android, though it still takes Google's billions for making Google search the iPhone default.”
Then to Apple’s choices remaining at the bar, approaching closing time.
“That leaves OpenAI and Anthropic — both of which Apple has explored partnering with, per Bloomberg.”
“Siri already lets users route questions to OpenAI's ChatGPT.”
“But a team tasked with evaluating Apple's external options found that Anthropic's Claude was the best candidate for the broader Siri upgrade, Bloomberg reported.”
It’s the classic Apple ‘Build vs Buy’ dynamic, which I reviewed recently.
“Yes, but: Apple could still decide to redouble its internal efforts instead.”
“The company has a long history of avoiding shipping half-baked products and letting projects take as long as they need to become what Apple thinks of as "great."
“But the pace of AI change is putting that strategy to its toughest test yet.”
And a wider aperture back to the whole bar:
“Zoom out: Driving and shaping all these firms' deals and choices is the war for AI talent, with both giants and startups desperately throwing money at a relatively small number of researchers.”
“Each company hopes that its team will be able to deliver on the astronomical promises executives have made about AI's transformative benefits and ultimate profits.”
“But not everyone can win — and each of the tech industry's previous waves has had only one or two victors.”
That is a decent survey of the bar scene at this mid-point in 2025 going into the July 4th weekend.
But the good news to always keep in mind, is that the bar will more likely be open again tomorrow. And it’s so early in this AI Tech Wave, despite the daily countdown to 'closing time’. Stay tuned.
(NOTE: The discussions here are for information purposes only, and not meant as investment advice at any time. Thanks for joining us here)