Fall 2013 Apple predictions
September 03, 2013
Hey! This is an old post. It reflects my views and the state of the world in 2013, but is likely to be outdated. You’ll find me on Twitter
, building Nozbe
, and making engineering blueprints
If you’re an Apple nerd, I’m sure you’re as excited as I am about this fall. They used to make product releases throughout the year, but not this time. Aside from minor tweaks and the Haswell MacBook Air, we’ve got nothing so far.
I thought it would be interesting to take what we know — the leaks, the rumors, external roadmaps and Apple’s historic choices — and try to predict what they release, with what features, using what technologies and for how much.
Let’s jump right in:
I had a hard time figuring out what iPhone 5S could realistically feature to be significantly better than the 5. All previous models had a very strong selling point: iPhone 3G replaced the original iPhone, was cheaper and had a 3G radio; 3GS was much faster; 4 had a stunning Retina display; 4S was much faster and had a better camera; and iPhone 5 was again much faster and also much nicer. What’s 5S’s selling point?
If you remember the iPhone 3GS keynote, S was meant to stand for speed. And it did matter, because iPhone 3G was seriously under-powered. And so was iPhone 4. But the 5 isn’t, as it also doubled the CPU and GPU performance. There just isn’t a need to make it faster, not yet.
“But if it’s not faster, what’s going to make it worth choosing over iPhone 5?” — so I thought. As a nerd I’m biased towards performance and overestimate its impact on purchasing decisions. In reality, most people don’t care, as long as it’s not slow.
Sure, there needs to be something that makes 5S better, but it doesn’t need to be a whole lot — iPhone is optimized for subsidized markets, and when you’re signing up for a $2000 phone contract, the upfront difference between a $100 phone and a $200 phone isn’t that big. I imagine that for most people new colors, a better camera and the mere fact that it’s a new phone are enough.
And for everyone else, for seemingly everyone in the internet, nothing Apple does this year will keep them from being very disappointed. Just like with 4S and with the 5. And yet 5S will be quickly announced to be the best-selling iPhone ever. Just like 4S and 5.
Either way, here are my predictions:
- $650 ($200 subsidized)
- new colors: Gold/Champagne, Steel/Graphite. (Bright, playful colors in 5C and metallic, luxury colors in 5S — makes sense, right?)
- new 128GB model (at $950). The market for such a model isn’t huge, but that’s okay — it’s easy to just solder a different Flash chip on the logic board. And why not? Apple is doing it already with iPad, so why not do it with their flagship phone?
- better camera and dual-LED flash for better low-light pictures
- fingerprint sensor for secure device unlocking and perhaps mobile payments
- world phone: instead of selling 3 iPhone versions, there will be just one (thanks to a new Qualcomm chip)
- There have been rumors about A7 being a 64-bit CPU. I don’t buy it. It will happen eventually, but it doesn’t seem like ARMv8’s advantages outweigh the more complex architecture’s overhead yet
- greater battery life. If iPhone 5S isn’t going to be faster, it could be less power-hungry. iPad and MacBook Air already have fantastic battery lives, but iPhone lags behind. Many people barely go through the day on battery and even when you do, it’s subconsciously crippling the way you use it. Apple can do better. Here’s how:
- A7 will move to the 28nm process
- Buy power savings with die area — even at the same process node you can achieve better performance and/or power consumption by building smarter, more specialized circuits at the expense of larger die area (and, hence, cost). That’s one of the ways Haswell allows for the stunning 12-hour battery life in MBA, and it’s a trick Apple could pull with iPhone 5S.
- Make the logic board smaller so you can fit a bigger battery
- IGZO display. It uses a new semiconductor material in the transistor layer of the LCD. IGZO has a very high carrier mobility, so such a display needs less power to drive it. And since you can make the wires thinner (which makes the panel more transparent), less power is needed for the backlight too (it’s a big deal with high-resolution displays). It’s even better: it doesn’t leak charge as much, so you can make pauses between screen updates. tl;dr: 50-60% less power consumption of the display. It’s unclear whether it’s possible for Apple to get IGZO displays this year at such massive scale, though. I’m betting no.
- Panel self-refresh. Even when you’re looking at your home screen or reading an article, the GPU is still busy updating the screen content every 17ms. By embedding a small amount of RAM into the display itself and giving it the capability to refresh itself, we’re enabling the SoC to stay in the sleep state for longer, saving power.
Last year’s phone, sold for $550 (or $100 subsidized), in 16GB version only. We know how it works.
Here’s the problem with selling a two-year-old phone: it’s going to be slow. iPhone 4 is still a very nice, high-quality, premium phone, as it was in 2010. But its guts are subpar. They were fine when the phone was released (and we were too much in awe of the Retina display to complain), but now the A4 gives iPhone 4 bad rep.
The point is: a fast phone with cheaper enclosure, display and camera is a better deal than a high-quality phone that’s simply slow.
- C as in cheap. $450/free in US. Elsewhere: $350. (Crazy? So I thought, but Apple is already doing it with iPhone 4 — in Poland, for example, the it costs ~$370, $80 less than you’d expect. Same story in other countries I’ve checked.)
- C as in colors. White, Red, Yellow, Green, Blue…
- A6 processor
- 512MB of RAM, like in iPod Touch (but a downgrade from iPhone 5’s 1GB)
- 4″ IPS display, although a cheaper one (like in iPad mini), and with on-cell touchscreen (the in-cell technology found in iPhone 5 is said to be expensive)
- cheaper camera, most likely the same 5Mpx sensor found in iPod touch
Both are out. The 4 is now old and slow, and while 4S is still a pretty good phone, how awkward would it be to sell a nice, glass and steel (albeit old) phone cheaper than a plastic one (even if they were able to)? Also, it would suck to still sell old devices using the 30-pin connector and the 3.5″ display.
iPod shuffle and nano
Other than the possible update of enclosure colors to better match those of iPhone 5C, I don’t expect any changes here.
It’s ridiculous that iPod shuffle only has 2GB of storage, given that 3rd generation Shuffle had a 4GB model and the price difference for Apple is literally $1-2, but since they haven’t done a 4GB Shuffle in 4 years, they probably won’t do one this year either.
To be discontinued. It’s been 6 years since its release. When Classic was introduced, iPhone was still a novelty available only in US and Android as we know didn’t even exist yet. World changed so much since then.
Today, there is only one good reason for its existence: it has more space than any other iPod by a long shot (160GB vs 64GB on the top iPod touch) while being fairly cheap ($250 vs $400 for the top Touch). Still, how many people want 160GB of music in their pocket (or car)? I think not many, especially considering how much better the Touch (and any modern smartphone) is at everything else. Yes, there still will be people who want this much, but it’s a tiny and shrinking market for Apple.
I was very surprised last year that Touch carried the same display as the iPhone 5, since it’s a very expensive part. What I think Apple should have done is use a lower-quality panel (still IPS, just not as good) and a standard touchscreen. The device would be a bit thicker and the screen would be a bit worse, but it would still be great, and Apple would be more willing to put those saved bucks into, say, upgrading RAM to 1GB.
If they haven’t done that last year, though, they won’t downgrade the device this year. They could if they were to release a new, redesigned model with other features better, but that’s not going to happen either.
I’d also love to see the silly Loop dropped. They didn’t include it in the 16GB model (which was released later), so it’s probably going to be dropped with the next redesign. (But again, Apple wouldn’t make a downgrade like that now without a good justification).
And given that, not that much will change:
- like other iPods, the color scheme might be changed slightly
- A6 processor
- new 128GB model, at $500. Again, not a huge market, but probably big enough. It would make a pretty good alternative to the Classic, though a fairly expensive one.
One note as for the 16GB model. I don’t know if you remember, but iPod touch used to cost $200. After last year’s upgrade, though, it starts at $230 (most likely because of the expensive display). It’s unusual and rather awkward for Apple to have the 16GB model $70 cheaper than the 32GB one (and not $100). I think Apple would like to get back to the old price and will keep selling last year’s 16GB Touch to achieve that. Only the 32, 64 and the new 128GB models will get a new CPU.
- replaces iPad 4
- updated to the current design style: black anodized aluminium on the black model, shiny chamfered edges, different edge curvature on the back, narrower edges on the sides
- thinner and considerably lighter (currently, the big iPad weighs 650g, which is more than twice as much as iPad mini; I think 450g might be achievable)
- A7X SoC
- 2GB of RAM, up from 1GB. It will make switching between apps much faster and the overall responsiveness better. Because of Retina iPad’s giant resolution, apps take up a lot of space in memory (textures are heavy), so the more RAM there is, the longer apps’ processes will stay in there, ready to be sprung back to life
- iPad will move from a glass touchscreen to a film touchscreen (like they did with Mini). It will help reduce thickness and weight and will bring the display a bit closer to the surface
- move to IGZO would be awesome, and seems likely. Although the price difference will be more significant here than with the iPhone, iPad is relatively low in volume, making adoption of the new technology easier.
Will be discontinued. It now makes up less than 10% of all iPad units sold. Not surprising — Mini is much nicer and $70 cheaper (and if you want a big screen, iPad 4 is way better).
One note though, Apple will probably keep selling iPad 2 to educational institutions (big iPad is probably a better tool for the job than Mini, and Apple has a long history of providing otherwise-discontinued or even-lower-end models to education).
The problem Apple faces in the tablet market is that its horizontal competitors — Amazon and Google — sell them at cost. (Amazon is trying to make money on content and Google is trying to gain foothold for Android on tablets.) Look at the new Nexus 7. It has 2× the CPU performance, 4× as much RAM and 2.5× as many pixels as the current iPad mini. Oh, and it costs $100 less. That is some serious competition.
Granted, specs were never what made people buy Apple products, but still — the new Nexus has a nicer display, it’s faster and at that price, it just makes iPad mini look bad.
Hardware is where Apple makes its money, though, so we’re not going to see a Mini that would compete on price and specs. Instead, we’ll see two models: one competing on price, one competing on specs.
last year’s iPad mini
- price in the $250-270 range
- A5 processor, 512MB of RAM
- 16GB model only
- no rear camera. Sounds crazy? They’ve done it before with the 16GB iPod touch, and hey — it’s a goddamn iPad, not a camera!
- Silver+black color only (again, like the cheapest iPod touch)
iPad mini with Retina display
- slightly thicker, slightly heavier, slightly lower battery life (same tradeoffs that affected the big iPad after going Retina)
- A6X processor, or some version of it (it might be shrunk to 28nm and/or under-clocked to fit in this tiny machine’s power envelope)
- 1GB of RAM
- IGZO? Of the three devices where I mentioned this technology, Retina iPad mini needs it most, but its larger volume and lower margins (compared to the big iPad) are making it hard. Still, my bet is yes.
Retina MacBook Pro
- CPU: update to Haswell
- Quad core option for 13″ model would be nice, but won’t happen. There is no quad-core mobile Haswell part that would have an adequate GPU and a TDP low enough for the 13″ rMBP
- 13″: Intel Iris Graphics 5100. 2x the performance of the previous HD Graphics 4000.
- 15″: Iris Pro 5200 (perhaps even a custom, even-higher-binned part). It will replace the GeForce 650M, yielding a slightly lower performance (which will make some people really, really angry), but a considerably better experience for everyone (lower power consumption, no weird issues with fans suddenly spinning up, etc.)
- 802.11ac. No difference unless you own a 802.11ac router, but it’s a big step forward
- super fast PCIe solid state drives. 800MB/s, almost 50% faster than previous-generation SATA III drives
- insane battery life. 12 hours will be easy. Even more than that will be achievable (especially on Mavericks)
- Thunderbolt 2, capable of driving a 4K (think: 27″ Retina) display
- Slightly optimistic price predictions:
- 13″: $1400/$1600
- 15″: $1900/$2500
- If Apple does get rid of the discrete GPU on the 15″ model, there’s going to be a lot of free space inside of the machine. (Look at how much space the CPU, GPU, PCH and their power management take up.) But at the same time, the power requirements go down, so there’s no need to put an even heftier battery inside. I think they’ll make the device even thinner and lighter to take advantage of that. (And who knows, maybe they’ll add some new crazy feature I can’t think of?)
To be discontinued. Remember the plastic white MacBook? You could see its end on the horizon when Apple released the second generation of MacBook Air in 2010 and priced it starting at $999. MacBook was killed next year when MBA was updated to Sandy Bridge.
Expect the same strategy with MacBook Pro. When the Retina MBP is updated to Haswell this fall, the classic MBP will be discontinued. It’s about time.
Here’s my problem with the current iMac: I think it’s quite pathetic to pay $1300 for a computer, in 2013, and not get an SSD. Worse than that, get a 5400rpm hard drive. That’s actually a downgrade from 2009-era iMac, which had a faster 7200rpm hard drive. Yes, a 2.5″ HDD allows for a thinner enclosure and 5400rpm means that it doesn’t introduce vibrations to iMac’s delicate structure — but at what cost? Making a desktop computer thinner (as if it made a difference on the desk) at the expense of worsening the already biggest performance bottleneck in the computer deserves slapping a “form over function” label on it.
So you want an SSD-equipped iMac. The cheapest option you can get now is a 1TB+128GB Fusion Drive for $250 extra. I think it’s unacceptable.
I’d love to see the Fusion Drive standard in every iMac, but that’s not going to happen in the current form — the cost is prohibitive. Honestly, though, who needs a terabyte of space on their computer? Some people do, but I bet most don’t even use half of that space. And you don’t need 128GB SSD to get a big boost in performance. All you need is to put your OS, apps and their configuration files there. 128GB is nice, but when you also have a hard drive for everything else, 64GB is plenty. As a user of a Mac Mini super-charged with a 64GB SSD, I can attest that.
And given that:
- I (somewhat optimistically) predict that Apple will put a 500+64GB Fusion Drive standard in every iMac. What about the cost? The price of a 500GB hard drive and 64GB SSD is about $45 higher than the price of a 1TB HDD (on Amazon; Apple gets them cheaper). I think iMacs carry margins high enough to swallow up the extra cost
- extra capacities: 1TB+128GB Fusion drive at $150 extra, 256GB SSD at $200 extra, 512GB SSD at $500 extra.
- PCIe SSDs
- Haswell, 802.11ac, Thunderbolt 2
- GPU: upgrade to GeForce 7xx. I don’t them bouncing back to AMD this year (it happens every few years), and high-end Intel Iris Pro GPUs are not available for high-TDP desktop parts
- Haswell, HD Graphics 5000, 802.11ac
- I’d love to see the Fusion Drive standard in all models, but Mac mini has tighter margins than iMac. I’d be okay if it meant that all models were $50 more expensive, but Apple won’t do that. They might put Fusion Drive in the high-end model, but the $600 one will only feature a 5400rpm HDD
- PCIe SSDs where appropriate. It would be fine if they put previous-generation SATA III drives here, but I doubt they would be any cheaper for Apple, given the scale on which they’re producing the PCIe ones for MacBook Air. With things like that, Apple tends to unify all models.
- it’s been 3 years since Mini’s last redesign. It used to have an optical drive back then! It’d be nice to see Mini become even smaller, but it will probably happen next year, when Intel Broadwell brings its power requirements further down
Can’t wait to see how much of this I got wrong.
I did my best to try to guess Apple’s strategy, but let’s not forget how good they are at surprising, delighting (and sometimes disappointing) us.
And not that I would mind. I love surprises — don’t you?
Am I missing something? Did I screw up the facts? Let me know.
Big thanks to Rinat Khanov for reading my drafts
Published September 03, 2013.