Apple says the M4 includes “up to” four high-performance CPU cores, six high-efficiency cores, and a 10-core GPU. Apple’s high-level performance estimates say that the M4 has 50 percent faster CPU performance and four times as much graphics performance. Like the GPU in the M3, the M4 also supports hardware-accelerated ray-tracing to enable more advanced lighting effects in games and other apps. Due partly to its “second-generation” 3 nm manufacturing process, Apple says the M4 can match the performance of the M2 while using just half the power.
As with so much else in the tech industry right now, the M4 also has an AI focus; Apple says it’s beefing up the 16-core Neural Engine (Apple’s equivalent of the Neural Processing Unit that companies like Qualcomm, Intel, AMD, and Microsoft have been pushing lately). Apple says the M4 runs up to 38 trillion operations per second (TOPS), considerably ahead of Intel’s Meteor Lake platform, though a bit short of the 45 TOPS that Qualcomm is promising with the Snapdragon X Elite and Plus series. The M3’s Neural Engine is only capable of 18 TOPS, so that’s a major step up for Apple’s hardware.
Apple’s chips since 2017 have included some version of the Neural Engine, though to date, those have mostly been used to enhance and categorize photos, perform optical character recognition, enable offline dictation, and do other oddities. But it may be that Apple needs something faster for the kinds of on-device large language model-backed generative AI that it’s expected to introduce in iOS and iPadOS 18 at WWDC next month.
The wait between the M1 and M2 and the wait between the M2 and M3 were each about a year and a half. With as few technical details as Apple has announced, it’s tough to know what to make of the faster turnaround between the M3 and M4. It could be that the M3 was behind schedule and the M4 was on time or ahead; it could also be that the M4 is a relatively modest architectural update to the M3. We’ll need to test the hardware ourselves to determine exactly how the M3 and M4 stack up to each other.
The fast introduction of the M4 makes it a little clearer why Apple might choose not to update devices like the Mac mini with an M3 chip. Either the M3 processor generation will be uncommonly short, or Apple plans to sell a mix of M3 and M4 devices this year.
Apple’s newest iPad Pro puts an M4 chip inside a thinner frame and is available in new 11-inch and 13-inch sizes, while also upgrading the screens on both to “tandem” OLED displays for more brightness.
Compared to the last iPad Pro, released in early 2022, Apple is highlighting how thin and light these new Pros are. The 11-inch model is 5.3 mm thick and weighs less than a pound, while the 13-inch is 5.1 mm, which Apple says is its thinnest product ever, at 1.28 pounds.
The tandem OLED design, dubbed Ultra Retina XDR, delivers 1000 nits at full-screen brightness, and 1600 nits at peak HDR, equivalent to a high-end Samsung TV. The screens are “nano-texture glass,” which is essentially a matte display finish.
What’s really big is the inclusion of an M4 chip, built from a second-generation 3-nanometer process. It’s the first time the iPad Pro has included a new Apple chip ahead of other devices. Apple is touting a 50 percent improvement over M2 performance and can deliver the same performance as M2 at half the power, or one-quarter the power in certain scenarios. Notably, the new iPad Air uses the M2; this year’s lineup puts more of a gap between the use cases and price/performance points of the various iPads, even if there is still some cross-over.
The chip’s neural engine got a very specific call-out; it’s capable of 38 trillion operations per second, and Apple says it’s the most powerful (consumer-level) AI computer sold.
The new iPad Pros get a 12 megapixel camera, along with LIDAR scanners and an adaptive flash that improves document scanning. Finally, the “selfie”/conferencing camera moves to the landscape position on the iPad Pro, making that ultra-wide 12MP camera far more useful and the iPad-as-laptop experience a good deal better. On that note, there are new Magic Keyboards, Smart Folios, and an Apple Pencil Pro available soon to attach to these new models.
The 11-inch iPad Pro starts at $999, and the 13-inch starts at $1,299, with 256GB storage to start. They’re both available for order today, with delivery next week.
This is a developing story and this post will be updated.
Apple today announced the first 13-inch iPad Air. The company is also releasing a revamped 11-inch iPad Air next week, meaning the tablet will be available in two sizes for the first time.
The 13-inch iPad Air has 30 percent more screen real estate than its smaller counterpart, so Apple is marketing it as being for multitasking and applications like iPadOS’ Split View. During its Let Loose event today, Apple said it decided to release a 13-inch iPad Air because more than half of iPad Pro users opt for the larger (12.9 inches versus 11 inches) size.
In addition to a larger screen, the 13-inch iPad Air will offer better sound quality than the 11-inch version, Apple says, due to it offering twice the bass. Both tablets have landscape stereo speakers and Spatial Audio support.
Both iPad Airs will come with an M2 chip that Apple claimed will give the devices 15 percent faster CPU performance, 25 percent faster GPU performance, and a 40 percent faster neural engine. With improved memory bandwidth as well, the new iPad Airs will be 50 percent faster than M1-based predecessors, Apple says.
Apple also says the new chip will make the tablets more appropriate for AI-based tasks, like visual look-up, subject lift, and live text capture.
The new Airs will also have a front-facing, ultrawide 12MP camera that supports Apple’s Center Stage feature for keeping users in a central field of view during video calls. Apple moved the camera positioning to one of the tablet’s longer edges to appeal to the way people tend to use tablets for video calls.
Enlarge/ The new iPad Air cameras support 4K video and 240-fps slo-mo.
Apple
Apple’s new thin tablets will come in four colors.
Enlarge/ Apple’s new iPad Airs in blue, purple, starlight, and space gray.
Apple
The 13-inch iPad Air will start at $799 with 128GB of storage, and the 11-inch iPad Air will start at $599, with storage going up to 1TB. The tablets will be available starting May 15.
Enlarge/ The promotional image for Apple’s May 7 event.
Apple
On May 7, Apple will host a product announcement event at 9 am ET. Labeled “Let loose,” we expect it will focus on new iPads and iPad accessories.
We won’t be liveblogging the stream, but you can expect some news coverage as it happens. Below, we’ll go over our educated guesses about why Apple might be doing this.
Why hold an event now?
It’s unusual for Apple to host an event shortly before WWDC. New products debut at that event all the time, so if it’s just a faster chip and a nicer screen for the iPad Pro and iPad Air, why not wait until June?
We’re not completely sure what the answer is, but we can make educated guesses.
Apple has been criticized by commentators over the past few weeks for three things. First, iPads have not been selling well. Second: While the Vision Pro introduced a new product category that may grow over time, its initial launch didn’t sell that well. Lastly, Apple is perceived by many as way, way behind on generative AI tech, which is already transforming other companies. There have also been rumors that Microsoft might announce a new silicon that will be fiercely competitive with Apple Silicon for AI tasks, and Microsoft’s chips could be announced at a planned event between May 7 and the start of WWDC.
Given all that, our best guess is that Apple wants to focus its messaging—and the time window for that messaging—on the right targets without muddying the message by trying to address everything at once.
Talking about the iPad’s challenges while also preempting Microsoft with a new chip announcement could be Apple’s focus for this event. Getting those things out of the way now would allow WWDC and its new operating system announcements to focus heavily on AI, which is the bigger question the company is looking to answer.
What new iPads might look like
Given that an Apple Pencil is in the event’s promotional image—a much more explicit hint than Apple usually provides—there’s no question iPads will be a focus.
As is often the case these days, we have a plethora of leaks, supply chain reports, and, of course, insider reporting at sites like Bloomberg and The Information to give us a rough idea of what to expect from Apple’s new hardware.
iPad Pro
It’s likely that the star of the show will be a significant redesign of the iPad Pro for the 11-inch and 12.9-inch sizes.
Both are rumored to get OLED displays, a huge step up over the LCD display in the current 11-inch iPad Pro. The 12.9-inch iPad Pro has a MiniLED display, which competes directly with OLED in the consumer TV space, so it won’t be as big a leap for that device, but we can still expect better contrast and richer colors.
Enlarge/ The Apple Park campus in Cupertino, California.
Apple’s earnings report for the second quarter of the company’s 2024 fiscal year showed a slide in hardware sales, especially for the iPhone. Nonetheless, Apple beat analysts’ estimates for the quarter thanks to the company’s rapidly growing services revenue.
iPhone revenue dropped from $51.33 billion in the same quarter last year to $45.96 billion, a fall of about 10 percent. This was the second consecutive quarter with declining iPhone revenues. That said, investors feared a sharp drop before the earnings call.
Notably, Apple’s revenue in the region it dubs Greater China (which includes China, Taiwan, Singapore, and Hong Kong) fell 8 percent overall. The company fared a little better in other regions. China’s economy is slowing even as China-based Huawei is taking bigger slices of the pie in the region.
Globally, Mac revenue was $7.5 billion compared to last year’s $7.12 billion. Other products—which include the Watch, AirPods, Apple TV 4K, HomePod, and the new Vision Pro headset—were down to $7.9 billion from last year’s $8.76 billion, despite the fact this quarter included the launch of the Vision Pro.
iPad revenue was also down, at $5.6 billion from $6.67 billion. Apple is expected to launch new iPads next week, which suggests that those updates are needed to achieve the company’s business goals.
The rosiest revenue category was services, which includes everything from Apple Music to iCloud. Its revenue was $23.9 billion, up from Q2 2023’s $20.91 billion.
The company also announced authorization of $110 billion for share purchases.
Enlarge/ John Schmidtlein, partner at Williams & Connolly LLP and lead litigator for Alphabet Inc.’s Google, arrives to federal court in Washington, DC, US, on Monday, Oct. 2, 2023.
Halfway through the first day of closing arguments in the Department of Justice’s big antitrust trial against Google, US District Judge Amit Mehta posed the question that likely many Google users have pondered over years of DOJ claims that Google’s market dominance has harmed users.
“What should Google have done to remain outside the crosshairs of the DOJ?” Mehta asked plaintiffs halfway through the first of two full days of closing arguments.
According to the DOJ and state attorneys general suing, Google has diminished search quality everywhere online, primarily by locking rivals out of default positions on devices and in browsers. By paying billions for default placements that the government has argued allowed Google to hoard traffic and profits, Google allegedly made it nearly impossible for rivals to secure enough traffic to compete, ultimately decreasing competition and innovation in search by limiting the number of viable search engines in the market.
The DOJ’s lead litigator, Kenneth Dintzer, told Mehta that what Google should have done was acknowledge that the search giant had an enormous market share and consider its duties more carefully under antitrust law. Instead, Dintzer alleged, Google chose the route of “hiding” and “destroying documents” because it was aware of conflicts with antitrust law.
“What should Google have done?” Dintzer told Mehta. “They should have recognized that by demanding locking down every default that they were opening themselves up to a challenge on the conduct.”
The most controversial default agreement that Google has made is a 21-year deal with Apple that Mehta has described as the “heart” of the government’s case against Google. During the trial, a witness accidentally blurted out Google’s carefully guarded secret of just how highly it values the Apple deal, revealing that Google pays 36 percent of its search advertising revenue from Safari just to remain the default search tool in Apple’s browser. In 2022 alone, trial documents revealed that Google paid Apple $20 billion for the deal, Bloomberg reported.
That’s in stark contrast to the 12 percent of revenue that Android manufacturers get from their default deals with Google. The government wants the court to consider all these default deals to be anti-competitive, with Dintzer suggesting during closing arguments that they are the “centerpiece” of “a lot” of Google’s exclusionary behavior that ultimately allowed Google to become the best search engine today—by “capturing the default and preventing rivals from getting access to those defaults.”
Google’s lawyers have argued that Google succeeds on its merits. Today, lead litigator John Schmidtlein repeatedly pointed out that antitrust law is designed to protect the competitive process, not specific competitors who fail to invest and innovate—as Microsoft did by failing to recognize how crucial mobile search would become.
“Merely getting advantages by winning on quality, they may have an effect on a rival, but the question is, does it have an anti-competitive effect?” Schmidtlein argued, noting that the DOJ hadn’t “shown that absent the agreements, Microsoft would have toppled Google.”
But Dintzer argued that “a mistake by one rival doesn’t mean that Google gets to monopolize this market forever.” When asked to explain why everyone—including some of Google’s rivals—testified that Google won contracts purely because it was the best search engine, Dintzer warned Mehta that the fact that Google’s rivals “may be happy cashing Google’s checks doesn’t tell us anything.”
According to Schmidtlein, Google could have crossed the line with the Apple deal, but it didn’t.
“Google didn’t go on to say to Apple, if you don’t make us the default, no Google search on Apple devices at all,” Schmidtlein argued. “That would be suicide for Google.”
It’s still unclear how Mehta may be leaning in this case, interrogating both sides with care and making it clear that he expects all his biggest questions to be answered after closing arguments conclude Friday evening.
But Mehta did suggest at one point today that it seemed potentially “impossible” for anyone to compete with Google for default placements.
“How would anybody be able to spend billions and billions of dollars to possibly dislodge Google?” Mehta asked. “Is there any real competition for the default spot?”
According to Schmidtlein, that is precisely what “competition on the merits” looks like.
“Google is winning because it’s better, and Apple is deciding Google is better for users,” Schmidtlein argued. “The antitrust laws are not designed to ensure a competitive market. They’re designed to ensure a competitive process.”
Proving the potential anti-competitive effects of Google’s default agreements, particularly the Apple deal, has long been regarded as the most critical point in order to win the government’s case. So it’s no surprise that the attorney representing state attorneys general, Bill Cavanaugh, praised Mehta for asking, “What should Google have done?” According to Cavanaugh, that was the “right question” to pose in this trial.
“What should they have done 10 years ago when there was a recognition” that “we’re monopolists” and “we have substantial control in markets” is ask, “How should we proceed with our contracts?” Cavanaugh argued. “That’s the question that they answered, but they answered it in the wrong way.”
Seemingly if Google’s default contracts posed fewer exclusionary concerns, the government seems to be arguing, there would be more competition and therefore more investment and innovation in search. But as long as Google controls the general search market, the government alleged that users won’t be able to search the web the way that they want.
Google is hoping that Mehta will reject the government’s theories and instead rule that Google has done nothing to stop rivals from improving the search landscape. Early in the day, Mehta told the DOJ that he was “struggling to see” how Google has either stopped innovating or degraded its search engine as a result of lack of competition.
Closing arguments continue on Friday. Mehta is not expected to rule until late summer or early fall.
Enlarge/Assassin’s Creed Mirage returned to the earlier games’ focus on stealth assassinations in a historical urban environment.
Ubisoft
Apple has spent the last year trying to convince gamers that they can get a console-like, triple-A experience on the latest iPhones. The newest test of that promise will be Ubisoft’s Assassin’s Creed Mirage, which now has a release date and pricing information.
Mirage will land on compatible iPhones—the iPhone 15, iPhone 15 Plus, iPhone 15 Pro, and iPhone 15 Pro Max—on June 6, according to Ubisoft (though the App Store listing says June 10.) That coincides pretty closely with Apple’s annual developer conference, so we’d expect it to get a shoutout there. Ubisoft’s blog post also says it will come to the iPad Air and iPad Pro models with an M1 chip or later.
The game will be a free download with a 90-minute free trial. After that, you’ll have to pay $50 to keep playing, which is pretty close to what the game costs on PC and consoles. It will support cross-progression, provided you sign into Ubisoft Connect. Ubisoft Connect is not exactly beloved by players, but it’s nice to be able to take your saves back and forth between other platforms if you can stomach it.
That cross-progression feature is key because the game launched several months ago on other platforms, so players interested in it probably already have made some progress in the story, if they haven’t finished it already.
Mirage is well over a dozen mainline games into the franchise, but it’s a smaller, more focused game than 2018’s Odyssey or 2020’s Valhalla. While those games expanded the franchise away from its stealth roots to become more of a full-fledged The Witcher 3-like open-world RPG experience, Mirage goes back to the old style of gameplay. It originally started as DLC for Valhalla but was expanded into a full game.
It won’t be the first triple-A game to hit the iPhone 15 and later, though the list has been short so far. A couple of Resident Evil games have made their way to phones (Resident Evil 4‘s remake and Resident Evil Village), and Apple has also managed to get respectable ports of No Man’s Sky, Death Stranding, and Baldur’s Gate 3 to Apple Silicon Macs.
When we tested the Resident Evil titles on the iPhone 15, we found that the graphics and performance were quite respectable—perhaps comparable to what you’d get on a PlayStation 4 Pro, a mid-range gaming laptop, or a Steam Deck—but that the touch controls never seem to cut it, so you’ll want to use a controller. iOS supports the latest PlayStation, Xbox, and Nintendo controllers, as well as attachable controllers like the Razer Kishi. Mirage will also support those controllers.
Enlarge/ An iPhone in Standby mode, charging wirelessly on a desk.
Apple
If your iPhone’s alarm hasn’t woken you lately, it seems you’re not alone: Apple has confirmed to Today that a software bug is to blame, following user complaints on TikTok and other social platforms.
Apple is “aware of an issue causing some iPhone alarms to not play the expected sound,” according to the report and “is working on a fix.” The company’s official statement didn’t go into more detail on what caused the bug or why it seems to affect some users but not others.
These sorts of bugs usually relate to some kind of time change; one circa 2010 iOS alarm bug was caused by Daylight Saving Time, and another cropped up in the first two days of 2011 when alarms suddenly stopped working for the first two days of the year (for whatever reason. they began working properly again on January 3 without any kind of software update). Daylight Saving Time in 2024 kicked in all the way back in mid-March, so it’s hard to say whether the problem is related to the change this time around.
If you aren’t affected by the bug—my alarms have been working fine—you can still keep this bug in your pocket for when you’re late for something for another reason.
Starting in March with the release of iOS 17.4, iPhones in the European Union have been subject to the EU’s Digital Markets Act (DMA), a batch of regulations that (among other things) forced Apple to support alternate app stores, app sideloading, and third-party browser engines in iOS for the first time. Today, EU regulators announced that they are also categorizing Apple’s iPadOS as a “gatekeeper,” meaning that the iPad will soon be subject to the same regulations as the iPhone.
The EU began investigating whether iPadOS would qualify as a gatekeeper in September 2023, the same day it decided that iOS, the Safari browser, and the App Store were all gatekeepers.
“Apple now has six months to ensure full compliance of iPadOS with the DMA obligations,” reads the EU’s blog post about the change.
Apple technically split the iPad’s operating system from the iPhone’s in 2019 when it began calling its tablet operating system “iPadOS” instead of iOS. But practically speaking, little separates the two operating systems under the hood. Both iOS and iPadOS share the same software build numbers, they’re updated in lockstep (with rare exceptions), and most importantly for DMA compliance purposes, they pull software from the same locked-down App Store with the same Apple-imposed restrictions in place.
Apps distributed through alternate app stores or third-party websites will have to abide by many of Apple’s rules and will still generally be limited to using Apple’s public APIs. However, the ability to use alternate app stores and browser engines on the iPad’s large screen (and the desktop-class M-series chips) could make the tablets better laptop replacements by allowing them to do more of the things that Mac users can do on their systems.
Though Apple has made multiple changes to iOS in the EU to comply with the DMA, EU regulators are already investigating Apple (as well as Google and Meta) for “non-compliance.” Depending on the results of that investigation, the EU may require Apple to make more changes to the way it allows third-party apps to be installed in iOS and to the way that third-party developers are allowed to advertise non-Apple app store and payment options. Any changes that Apple makes to iOS to comply with the investigation’s findings will presumably trickle down to the iPad as well.
Of course, none of this directly affects US-based iPhone or iPad users, whose devices remain restricted to Apple’s app stores and the WebKit browsing engine. That said, we have seen some recent App Store rule changes that have arguably trickled down from Apple’s attempts to comply with the DMA, most notably policy changes that have allowed (some, not all) retro game console emulators into the App Store for the first time.
In the world of AI, what might be called “small language models” have been growing in popularity recently because they can be run on a local device instead of requiring data center-grade computers in the cloud. On Wednesday, Apple introduced a set of tiny source-available AI language models called OpenELM that are small enough to run directly on a smartphone. They’re mostly proof-of-concept research models for now, but they could form the basis of future on-device AI offerings from Apple.
Apple’s new AI models, collectively named OpenELM for “Open-source Efficient Language Models,” are currently available on the Hugging Face under an Apple Sample Code License. Since there are some restrictions in the license, it may not fit the commonly accepted definition of “open source,” but the source code for OpenELM is available.
On Tuesday, we covered Microsoft’s Phi-3 models, which aim to achieve something similar: a useful level of language understanding and processing performance in small AI models that can run locally. Phi-3-mini features 3.8 billion parameters, but some of Apple’s OpenELM models are much smaller, ranging from 270 million to 3 billion parameters in eight distinct models.
In comparison, the largest model yet released in Meta’s Llama 3 family includes 70 billion parameters (with a 400 billion version on the way), and OpenAI’s GPT-3 from 2020 shipped with 175 billion parameters. Parameter count serves as a rough measure of AI model capability and complexity, but recent research has focused on making smaller AI language models as capable as larger ones were a few years ago.
The eight OpenELM models come in two flavors: four as “pretrained” (basically a raw, next-token version of the model) and four as instruction-tuned (fine-tuned for instruction following, which is more ideal for developing AI assistants and chatbots):
OpenELM features a 2048-token maximum context window. The models were trained on the publicly available datasets RefinedWeb, a version of PILE with duplications removed, a subset of RedPajama, and a subset of Dolma v1.6, which Apple says totals around 1.8 trillion tokens of data. Tokens are fragmented representations of data used by AI language models for processing.
Apple says its approach with OpenELM includes a “layer-wise scaling strategy” that reportedly allocates parameters more efficiently across each layer, saving not only computational resources but also improving the model’s performance while being trained on fewer tokens. According to Apple’s released white paper, this strategy has enabled OpenELM to achieve a 2.36 percent improvement in accuracy over Allen AI’s OLMo 1B (another small language model) while requiring half as many pre-training tokens.
Enlarge/ An table comparing OpenELM with other small AI language models in a similar class, taken from the OpenELM research paper by Apple.
Apple
Apple also released the code for CoreNet, a library it used to train OpenELM—and it also included reproducible training recipes that allow the weights (neural network files) to be replicated, which is unusual for a major tech company so far. As Apple says in its OpenELM paper abstract, transparency is a key goal for the company: “The reproducibility and transparency of large language models are crucial for advancing open research, ensuring the trustworthiness of results, and enabling investigations into data and model biases, as well as potential risks.”
By releasing the source code, model weights, and training materials, Apple says it aims to “empower and enrich the open research community.” However, it also cautions that since the models were trained on publicly sourced datasets, “there exists the possibility of these models producing outputs that are inaccurate, harmful, biased, or objectionable in response to user prompts.”
While Apple has not yet integrated this new wave of AI language model capabilities into its consumer devices, the upcoming iOS 18 update (expected to be revealed in June at WWDC) is rumored to include new AI features that utilize on-device processing to ensure user privacy—though the company may potentially hire Google or OpenAI to handle more complex, off-device AI processing to give Siri a long-overdue boost.
Last year, Apple introduced the ability to set multiple timers at once in the Clock app on its various platforms.
“We truly live in an age of wonders,” deadpanned Apple’s Craig Federighi in the company’s official presentation, tacitly acknowledging the gap between the apparent simplicity of the feature and the amount of time that Apple took to implement it.
The next version of iPadOS may contain another of these “age of wonders” features, an apparently simple thing that Apple has chosen never to do for reasons that the company can’t or won’t explain. According to MacRumors, iPadOS 18 may finally be the update that brings a version of Apple’s first-party Calculator app to the iPad.
Calculator was one of the very first iPhone apps that shipped with the iPhone back in 2007 but was mysteriously and inexplicably absent from the iPad when it launched in 2010. It’s also the very last of those original missing apps to find its way to the iPad’s home screen—Stocks, Clock, Voice Memos, and Weather had all made the jump previously, with the Weather app coming as recently as 2022.
It’s not that the iPad is incapable of calculating; the Spotlight search feature can already handle basic off-the-cuff math and conversion questions, and third-party calculator apps like PCalc, Numerical², Calcbot, and innumerable free-to-download no-name calculator apps have stepped up to fill the gap. But it was never clear why Apple decided against shipping a first-party Calculator app with the iPad, when it had shipped one with every iPhone since 2007 and every Mac since 1984.
The new Calculator app should be more than just a straightforward port of the current iOS or macOS app. Apple is apparently planning a small overhaul of the Calculator app for macOS 15 with a history tape for tracking past calculations, a resizable window, and an updated round-button design that more closely imitates the iOS version. The iPad and macOS versions of many of Apple’s apps share a lot of code these days—Stocks, Voice Memos, News, Home, Weather, Clock, and others share essentially the same design and layout in both operating systems—so it’s a fair bet that this redesigned Mac app and the newly introduced iPad app will be the same software.
At least one developer of a prominent iPad calculator seemed undaunted by the news that his app could be Sherlocked this fall.
“Yes, I saw the MacRumors article,” wrote PCalc developer James Thomson on his Mastodon account, responding to no one in particular. “Yes, it’s fine.”
Bloomberg’s Mark Gurman thinks that Apple’s M4 chips for Macs are coming sooner rather than later—possibly as early as “late this year,” per a report from earlier this month. Now Gurman says Apple could completely skip the M3 generation for some Macs, most notably the Mac mini.
To be clear, Gurman doesn’t have specific insider information confirming that Apple is planning to skip the M3 mini. But based on Apple’s alleged late-2024-into-early-2025 timeline for the M4 mini, he believes that it’s “probably safe to say” that there’s not enough space on the calendar for an M3 mini to be released between now and then.
This wouldn’t be the first time an Apple Silicon Mac had skipped a chip generation—the 24-inch iMac was never updated with the M2, instead jumping directly from the M1 to the M3. The Mac Pro also skipped the M1 series, leapfrogging from Intel chips to the M2.
But if the M4 does come out by the end of 2024, it would be a much faster turnaround than we’ve seen for other Apple Silicon chips so far. Roughly a year and a half passed between the introduction of the first M1 Macs in late 2020 and the first M2 Macs in the summer of 2022; about the same amount of time passed between mid-2022 and the late-2023 introduction of the first M3 Macs. If Apple holds to a more typical 18-month gap between the first M3 Macs and the first M4 Macs, there’s still plenty of time for an M3-based Mac mini refresh to be released.
Apple last updated the Mac mini in January of 2022, replacing the M1 model with an M2 version and introducing a new variant with an M2 Pro chip that included more Thunderbolt ports, better external display support, and better CPU and GPU performance. Most of Apple’s desktops—both Mac minis, as well as the Mac Studio and Mac Pro—are still using Apple’s M2 chips, while all of the laptops and the iMac have gotten an M3 refresh at this point.
Gurman’s previous reporting on the M4 suggests that it will be an “AI-focused” chip series, which probably means that it will beef up the processors’ Neural Engine to power the on-device generative AI features that are expected to come with iOS 18 and Apple’s other major operating system updates this year. Apple already has a head start on the PC ecosystem in this respect—all of the M-series chips and A-series chips going all the way back to 2017’s A11 Bionic have included a version of the Neural Engine. Intel and AMD’s processors have only begun to include similar neural processing units (NPUs) within the last year or so.
Gurman hasn’t reported on the M4 series’ specifications, but he has said it will include at least three performance tiers: a base model codenamed “Donan,” a midrange version codenamed “Brava,” and a high-end model codenamed “Hidra.” It remains to be seen which of these chips would replace the Pro, Max, and Ultra processors in current-generation M2 and M3 Macs.
