Developer conference season is over, culminating with what has, over the last 15 years, been the granddaddy of them all: Apple’s Worldwide Developers Conference (WWDC).
On to the update:
From the Wall Street Journal:
Apple Inc. introduced software changes Monday that aim to further expand its reach beyond the iPhone, including by boosting its payment system and car connectivity. The tech giant used the kickoff of its annual Worldwide Developers Conference to preview coming changes to its software for laptops, smartphones and tablets. It was a mix of morale booster for third-party developers that help fuel the digital ecosystem for its devices and an opportunity to preview how its own software will change later this year when officially rolled out.
If that seems like a rather boring lede, well, it was a rather boring keynote. Start right at the top: I’ve often noted that I always pay close attention to what a chief executive has to say at the beginning of their keynotes — particularly for their developer conferences — as it often lays out the overall strategic direction for the company. CEO Tim Cook, though, didn’t really say anything at all, and handed things off to Senior Vice President of Software Engineering and traditional WWDC emcee Craig Federighi for a CGI-transport through Apple Park and a dive into iOS. That was the pattern throughout: cutesy CGI Craig, and a completely disconnected launch into another part of Apple’s operating system family.
This is, of course, how Apple has long organized WWDC: iPhone OS, iPad OS, macOS, etc., each get their own section of the keynote. What was striking about yesterday’s keynote, though, was that these divisions seemed like arbitrary placeholders for the unveiling of cross-platform features. For example, the iOS segment included an expansion of Live Text capabilities:
A new premium sports offering for Apple News:
Shared photos libraries:
The biggest news in the Watch OS segment was that the Fitness app was coming to the iPhone; the headline feature in the macOS segment, Stage Manager, was most notable because it also came to iPad OS.
The macOS segment continued with the rejiggered Spotlight:
A section on gaming:
Then it was time for iPad OS, and new collaboration features built on top of Messages:
A new collaboration app called Freeform:
An update to GameCenter:
Tim Cook came on to wrap up the keynote, announcing that all of the company’s new OS’s were available as betas today:
And there it is: it turned out Cook was plugged into the theme all along. The real story of this keynote was not the updates to iOS, macOS, watchOS, or iPadOS (tvOS and the rumored realityOS were conspicuous by their absence); rather, this was the annual update to Apple OS.
This makes sense: while iOS seems new to people who remember the launch of OS X, it’s worth noting that we are further away from the launch of watchOS (7 years) than iOS was from the launch of OS X (6 years). iOS itself is up to version 16! These are mature products that don’t really have much to add; it makes sense that whatever innovations Apple can come up with are further up the stack, specifically in the services layer that ties their products together.
Though primarily a software-focused event, Apple’s WWDC keynotes are often stage for an interesting hardware announcement or two as well, and this year Apple did not disappoint. At the company’s biggest Mac-related keynote of the year, Apple unveiled the M2, their second-generation Apple Silicon SoC for the Mac (and iPad) platform. Touting modest performance gains over the original M1 SoC of around 18% for multithreaded CPU workloads and 35% in peak GPU workloads, the M2 is Apple’s first chance to iterate on their Mac SoC to incorporate updated technologies, as well as to refresh their lower-tier laptops in the face of recent updates from their competitors.
With the king of the M1 SoCs, M1 Ultra, not even 3 months behind them, Apple hasn’t wasted any time in preparing their second generation of Apple Silicon SoCs. To that end, the company has prepared what is the first (and undoubtedly not the last) of a new family of SoCs with the Apple Silicon M2. Designed to replace the M1 within Apple’s product lineup, the M2 SoC is being initially rolled out in refreshes of the 13-inch MacBook Pro, as well as the MacBook Air – which is getting a pretty hefty redesign of its own in the process.
The launch of the M2 also gives us our first real glimpse into how Apple is going to handle updates within the Apple Silicon ecosystem. With the iPhone family, Apple has kept to a yearly cadence for A-series SoC updates; conversely, the traditional PC ecosystem is on something closer to a 2-year cadence as of late. M2 seems to split this down the middle, coming about a year and a half after the original M1 – though in terms of architecture it looks closer to a yearly A-series SoC update.
Apple’s CPU progression for the M1 family was quite straightforward, at least once you realize that you actually have to start with the A14 chip the company built for the iPhone. Specifically, Apple designed four component cores — a high performance CPU core, high efficiency CPU core, GPU core, and neural engine core — and then repackaged them in various ways to make up the M1 family:
|high perf||high eff||GPU||NE|
The M2 is mostly following the same pattern, but with the A15 as its base:
|high perf||high eff||GPU||NE|
This suggests a lineup that will end up looking something like this:
|high perf||high eff||GPU||NE|
The Extreme is obviously speculative; something, though, needs to power the promised Apple Silicon-based Mac Pro, and this would certainly fit the progression.
That aside, there are some interesting differences between the M2 and the M1 that are worth noting, particularly the higher GPU count and, in what I suspect is a related development, an increase in the maximum RAM from 16GB to 24GB. It appears it is the latter that enabled the former; the M2 is the first Apple chip to use LPDDR5 RAM, which operates at significantly higher speed and thus provides significant extra bandwidth. That extra bandwidth is gobbled up by those extra GPU cores, which are likely the primary reason why the M2 is physically larger than the M1.
Still, despite these extra cores, the M2 has equivalent battery life to the M1, thanks to both design improvements and also the fact it — like the A15 — is built on TSMC’s N5P process (which, as I noted last fall, offers either 7% more speed or 15% more efficiency, or some combination of the two). The M2 also has some of the dedicated media modules first created for the M1 Pro.
What is interesting to consider is what process the M3 will be built on; to answer this the obvious place to start is with this fall’s A16 iPhone chip: my assumption is that Apple will use TSMC’s third-generation 5nm chips, which, confusingly enough, are being marketed as N4. Unlike N5P, N4 does offer an increase in transistor density, helping it deliver more meaningful performance and/or efficiency gains.
At the same time, it’s already June, and Apple only just announced the M2: while I suspect part of that delay has been driven by a combination of parts shortages and COVID lockdowns (i.e. the chips were ready but the re-designed MacBook Air wasn’t), the fact remains that the company still needs to roll out the rest of the line. It may make more sense to simply skip the A16.
This makes even more sense when you remember that TSMC’s 3nm process is coming online at the end of this year; 3nm is a huge leap over 5nm, and should provide massive increases in performance and efficiency to every chip built on that process. The first Apple product to be built on TSMC’s 3nm process will likely be the A17, and I would expect the A17 to be the foundation for the M3 series, which will probably launch in 4Q 2023 (because 3nm was a bit delayed, Apple will not be the launch customer for 3nm; on the flip side, the company will likely get as much volume as it wants when it is finally ready to go).
Credit where credit is due: after last year’s WWDC I opened my Article by noting that Apple had killed 3rd-party weather apps; Jacob Wolman argued that Apple was actually prepping a new weather API. One year later and WeatherKit is a thing and Wolman was right.
WeatherKit, which provides weather data to developers at a much cheaper rate than the Dark Sky service Apple acquired, is a great example of how Apple can impact privacy in a way that benefits everyone: having cheaper weather data decreases the incentives for 3rd-party weather apps to sell user data.
There are also encouraging signs about SKAdNetwork improvements; on the flipside there were no big privacy/business-model-destroying announcements during the keynote, like the feared IP-proxying-by-default that would kill fingerprinting for conversion measurement. It remains to be seen what WWDC sessions will bring, but for this year at least it looks like the ad industry can stay focused on continuing to recover from 2020’s WWDC announcement of ATT.
This Update will be available as a podcast later today. To receive it in your podcast player, visit Stratechery.
The Stratechery Update is intended for a single recipient, but occasional forwarding is totally fine! If you would like to order multiple subscriptions for your team with a group discount (minimum 5), please contact me directly.
Thanks for being a subscriber, and have a great day!