Affiliate Disclosure
If you buy through our links, we may get a commission. Read our ethics policy.

JP Morgan raises Apple stock target to $245 on AI announcements

JP Morgan has raised its Apple price target

Analysts at JP Morgan say that Apple's launch of AI features will prompt users to upgrade their iPhones, and has raised the price target by $20.

JP Morgan has previously expected that Apple's AI features would eventually drive an upgrade cycle, but its analysts were previously sure this wouldn't happen until the 2025 release of the iPhone 17 range. It was convinced enough that in April 2024 it dropped the target price to $210.

Even though it has increased that price since, it has only now taken the target higher than it was during the latter part of 2023. Specifically on the strength of Apple Intelligence and its expected appeal, JP Morgan has now set its Apple target price to $245.

The analysts say they do still expect to see a peak with the launch of the iPhone 17 range, but that the upgrade cycle will now start with the iPhone 16 series in September or October 2024. Overall, JP Morgan expects Apple to sell 250 million iPhones across calendar year 2025, and 275 million the following year.

JP Morgan continues to expect significant hardware updates in the iPhone 17 range to support AI functions. Its analysts believe that AI adoption across all smartphones will help accelerate demand for an AI-based iPhone.

The analysts estimate that Apple's iPhone revenues will reach $269 billion in financial year 2026. This is 34% higher than its own $201 billion prediction for the financial year 2024.

All of which marks a significant change from recent months when the majority of analysts were predicting serious issues for Apple over its declining sales in China. JP Morgan has consistently disagreed with that assessment, instead believing that Apple would recover from its China concerns.

Part of that belief in Apple's continued success was due to its growing Services revenue. Now JP Morgan says it is only raising its Services estimates modestly.

Yet it does also believe that AI features provided by third-party developers will increase Services revenues over the next few years. Similarly, the analysts say they expect to see AI features also driving upgrades to the iPad and the Mac.

One more reason for the analysts positive projections is Apple's focus on privacy and on-device AI. They say that this is an advantage to Apple because rivals are having to rely on cloud solutions, and so are going to incur greater costs.

JP Morgan is not the only firm to see Apple as once again being a good buy for investors. Following the WWDC announcements, there was a brief drop in Apple's stock price — but that quickly reversed itself.

Over the first day after the announcements, Apple's stock price climbed out of the brief 1.9% dip, and instead surged ahead 7.26%. Its stock price that day of $206.03 was Apple's 52-week high at the time.

Apple Intelligence is coming to the iPhone with iOS 18, or at least starting with that. The updates are expected to continue into 2025.



4 Comments

blastdoor 15 Years · 3594 comments

China is a major threat, as are laws targeting big tech companies around the world. 

But... Apple's unique PCC approach to AI seems potentially very valuable. It means even more revenue could come from services, with the server hardware backing up those services sitting in the US. Much easier to build server hardware in places other than China, too. 

And I sure would rather buy AAPL than NVDA. NVDA is vastly overpriced. Apple isn't the only huge company looking to avoid paying NVDA their 80 percent margins/ 

I predict the era of fabless chip design firms making huge $$ is coming to a close. The profits will go to the foundries (TSMC, Intel, Samsung) and the big firms that can afford to do their own internal chip design (Apple, Amazon, Microsoft, Google). Companies like Nvidia and AMD are going to get squeezed. 

mpantone 18 Years · 2254 comments

Nah, foundries have a lot of overhead costs compared to fabless chip designers. It is not getting easier to run foundries, it is getting harder which is why the semiconductor industry has seen a massive consolidation over the past twenty years.

Fabless designers have plenty of business. It's not like anyone is expecting Walmart, FedEx or JPMorganChase to internally design their own chips. 95% of the Fortune 500 will buy AI accelerators (and other related datacenter hardware) from other providers. And remember that Jensen said these systems are really about 70% software. NVIDIA benefits right now from having the most active software development infrastructure around. It's not just about running an synthetic AI benchmark.

Sure, maybe Company B's AI accelerator is cheaper than NVIDIA's. But if it doesn't have a software ecosystem that allows developers to create, it doesn't matter what it scored on ____ benchmark. We are seeing that right now with NVIDIA's 98% AI datacenter marketshare. It's not because the other companies' products can't post good scores, it's about the software surrounding it.

That 98% marketshare will be whittled away over time, everyone knows that. But if NVIDIA can maintain their technological lead, they can command a premium for their offerings and walk away with the lion's share of the industry profits. We see that on the consumer AIB graphics card market. AMD Radeon cards often score very closely to NVIDIA GeForce cards in pure rasterization benchmarks but once you throw in other features (ray tracing, machine learning super-sampling) and software quality, there's a hands down winner.

It's not just about how many cores you can cram onto a die. And foundries like TSMC and Samsung don't offer development ecosystems like NVIDIA Omniverse. That's where much of NVIDIA's value lies.

libertyandfree 11 Years · 192 comments

 If AI really takes off and is a big benefit to the consumer we will all be subscribing to some AI service within the next few years and this could adds $10’s Billions of additional revenue per year. It wasn’t that long ago the majority had no cell phone.them we had low cost feature phones and now everyone is paying $50 per month for unlimited service for every member of their family above 12 years old. 

blastdoor 15 Years · 3594 comments

mpantone said:
Nah, foundries have a lot of overhead costs compared to fabless chip designers. It is not getting easier to run foundries, it is getting harder which is why the semiconductor industry has seen a massive consolidation over the past twenty years.

You just did a nice job of explaining one of the two reasons that foundries are going to be more profitable than the design firms. There aren't a lot of foundries and the barriers to entry are high. That's one condition for earning high profits. The other is to have huge demand for your products, which it seems the foundries have (and will have more of). When nobody else can do what you do and there's a big demand for what you do, you win. 

Fabless designers have plenty of business. It's not like anyone is expecting Walmart, FedEx or JPMorganChase to internally design their own chips. 95% of the Fortune 500 will buy AI accelerators (and other related datacenter hardware) from other providers.

95% of the Fortune 500 will pay for AI services. That doesn't mean they're buying chips from Nvidia. More likely, it means they're paying for AI services from Microsoft, Amazon, or Google. And 95% of the Fortune 500 don't care what chips those service providers are running in their data centers. So the only ones who need internal chip design are Microsoft, Amazon, Google, (and for consumers and maybe some slice of small business and some small slice of enterprise, Apple). While fabbing cutting edge chips is harder than ever, designing cutting edge chips is easier (not easy in an absolute sense, just easier in a relative sense) than ever. Automated design tools and lincesable IP make chip design vastly easier today than it was in the 1990s. So those big firms can, and are, designing their own chips like never before. 

And remember that Jensen said these systems are really about 70% software. NVIDIA benefits right now from having the most active software development infrastructure around. It's not just about running an synthetic AI benchmark.
Sure, maybe Company B's AI accelerator is cheaper than NVIDIA's. But if it doesn't have a software ecosystem that allows developers to create, it doesn't matter what it scored on ____ benchmark. We are seeing that right now with NVIDIA's 98% AI datacenter marketshare. It's not because the other companies' products can't post good scores, it's about the software surrounding it.
That 98% marketshare will be whittled away over time, everyone knows that. But if NVIDIA can maintain their technological lead, they can command a premium for their offerings and walk away with the lion's share of the industry profits. We see that on the consumer AIB graphics card market. AMD Radeon cards often score very closely to NVIDIA GeForce cards in pure rasterization benchmarks but once you throw in other features (ray tracing, machine learning super-sampling) and software quality, there's a hands down winner.
It's not just about how many cores you can cram onto a die. And foundries like TSMC and Samsung don't offer development ecosystems like NVIDIA Omniverse. That's where much of NVIDIA's value lies.

Nvidia isn't the only company that knows how to do software, and while CUDA lock-in exists, it's nothing like Wintel lock-in once was. I'd say CUDA has about as much lock-in power as the old proprietary big iron operating systems, like System/360. IBM seemed unbeatable at its peak, but how's System/360 doing relative to Linux in the data center? Companies like Apple, Microsoft, Google, and even Amazon absolutely know how to do software as well as Nvidia. And they are operating at a scale that makes the investment worthwhile, even if only for internal purposes. Also, the core of AI is just a bunch of math, and you can't patent math and there are a lot of ways to do math. 

Note that I'm not saying Nvidia is going out of business. I'm just saying their sky high profit margins are going to fall a lot, and so is their stock. They aren't inherently a trillion dollar company. And within 10 years, I bet Intel surpasses them in market cap. I guess the caveat I'd put on that prediction is that if Nvidia were to buy Intel now, then that would put them in a position to eventually be worth their market cap, because then they'd have something that's really valuable -- the ability to fab their own chips. I doubt Nvidia leadership has the vision to buy Intel, though.