“Apple still has a team working on its own internal models that it could take advantage of in the future. But some Apple leaders hold the view that large language models will become commodities in the years to come and that spending a fortune now on its own models doesn’t make sense.”
“If Apple leadership truly does believe LLMs will become commodities, then the company’s AI success will depend less on bespoke new models, and more on its ability to control the hardware, software, and services that AI runs on.
And for now at least, those are the very areas that Apple’s at its best.”
Really? IMHO, it depends on whether Satellite internet can be as good as cellular network and Apple’s foolishness not to launch a Satellite device or make iPhone a capable satellite phone.
The stars of the report were iPhone sales, which increased 23% year on year thanks to the new iPhone 17 model, and Apple’s business in Greater China, which surged 38% to $25.5 billion.
After all, the analyst noted, the prices of the memory chips used in Apple products are currently through the roof, and supply is severely constrained, because the chips are also in high demand for all the AI data centers being built.
AI and datacenter are causing many things to be expensive. Electricity, water, GPUs, memory chips,…
Talking about memory chip prices.
He argued that Apple is constrained at this point, and it’s hard to lay out a timeline when supply and demand will balance, given key bottlenecks in advanced chip nodes and limited supply-chain flexibility.
Customers don’t feel like they “need AI today,” Munster said. “But investors are yearning for more details.”
Until Apple provides more substance around AI features, timing and monetization, Munster said, investors are likely to remain in a wait-and-see mode — even as the company posts strong financial results.
I’ve counted five waves from the $243.42 level, so I expect a retracement soon. However, I’m not sure how high the price will go before the pullback begins.
“We don’t believe this is a function of strength in the PC market because non-Mac PC lead times show no recent sign of a positive inflection,” Woodring posited. “Instead, we believe this [is] a function of Mac-specific demand strength boosted by the popularity of OpenClaw (aka Moltbot), which is meant to run locally on-device (for privacy and performance). While strength across the Mac lineup should likely drive some revenue upside in the March quarter, we’ll keep watching lead times to see how sustainable this trend is.”
Abel gave Apple and these other three companies extremely high praise, not only saying they should compound over time, but also that they are businesses that Berkshire understands well and ones in which it has “a high regard for their leaders…”
…the most telling takeaway this year was about what Berkshire won’t be selling. Abel indicated investors should expect only “limited activity” in the company’s biggest equity holdings, including American Express , Coca-Cola , Moody’s , and iPhone maker Apple (NASDAQ: AAPL).
Apple (AAPL) still makes up more than 20% of Warren Buffett’s (BRK) stock portfolio. Therefore, it’s likely that BRK will continue selling AAPLs until its share drops to around 10%.
“Last October Apple announced the M5 chip, embedding a neural accelerator in every GPU core. This wasn’t marketing theater. Apple’s own benchmarks show the M5 running a 30-billion-parameter model—capable of drafting legal briefs, debugging code, synthesizing research—in under three seconds on a standard MacBook Pro. Eighteen gigabytes of memory. No internet. No subscription. No API key. The trick is an architectural advance called Mixture of Experts: The model carries 30 billion parameters but fires only three billion per query, like a law firm where all the partners are on retainer but only the relevant specialist answers your call.”
“Logic suggests the price maker in the AI stack will be whoever is closest to the end user. In the consumer market, Apple dominates; its AI-powered software is expected to be running in most if not all of its devices by 2032. New devices with “Apple Intelligence” will locally handle the tasks, email summaries, writing assistance, photo editing, translation and search, that would otherwise generate cloud inference calls.”
