If you remember my famous forecast from year 2010, that if Moore’s Law holds, then someone can clone any superphone of year 2010, and match those specs by year 2020, for the cost of $10. Superphones of year 2010 included the original iPhone 4, Samsung Galaxy, Nokia N8, Blackberry’s Bold, etc. And a superphone back then cost $600 without handset subsidy (ie clean, without contract). Excepting for local taxes and import duties, that price was also pretty well ‘universal’ globally. In year 2010, you’d be expected to pay $600 for a top-line smartphone by a major brand.
To be clear, right from the start, I emphasized, that Apple will not sell us a $10 iPhone in year 2020. I said it would be a clone-phone-maker, probably running Android or possibly one of the low-cost smartphone OS systems that were then in development. But 8 years ago, in year 2010, if you went into a mobile phone shop anywhere, and picked the top model, you’d get roughly the same specs, which were:
3G
Smartphone
Touch Screen
3.5 inch size screen
WiFi
Camera of 5mp
With LED flash
GPS
That was my spec list. If you did that phone today, you are essentially matching an iPhone 4 from year 2010. And that phone was considered a ‘superphone’ back then. A powerful internet-phone on both 3G and WiFi, a ‘real’ camera that was able to capture reasonably good images. A flash that let you take pictures in the dark. GPS for mapping. And yes, touch-screen obviously. And back in year 2010, a phone screen of 3.5 inches was considered ‘massive’ in its size. Today it is considered ‘tiny’.
MOORE’S LAW
So I applied Moore’s Law to the tech specs of the iPhone 4, in year 2010. And that law said that computer processing power doubles every 18 months. The corollary to Moore’s Law is that the same performance computer tech will cost half, in 18 months. And then I just ran 18-month cycles and cut the price of the ‘iPhone clone’ into half every 18 months.
This said that the price pattern for a new smartphone, matching 2010 flagship smartphones by spec, which cost $600 in year 2010, should cost as follows:
Summer 2010: cost $600 (starting point)
Winter 2011: cost $300 (was achieved)
Summer 2013: cost $150 (was achieved)
Winter 2014: cost $75 (was achieved)
Summer 2016: cost $38 (was achieved)
Winter 2017: cost $19 (is now due)
Summer 2019: cost $10
And based on this pattern, I boldly forecasted a $10 dollar price for a new smartphone, when sold in year 2020, that (roughly) matched the specs of what once was a superphone, in year 2010. Fair enough. And nicely, this gave us milestones, to see how this forecast was doing. I showed this slide since 2010, with always the latest period highlighted, as we followed how this forecast was turning out:
That forecast held through 2011, 2013, 2014, and even up to Summer 2016. I was able to find a new smartphone in the summer of 2016, that matched (or exceeded) the specs required, in India, for the local price of 2,443 India Rupees which was $36 US dollars. We were 2 dollars BELOW the target price for the forecast. The forecast looked good.
FORECAST NO LONGER HOLDS
Now I have some bad news. The forecast no longer holds. I have searched a dozen countries and phones of hundreds of brands, and cannot find a smartphone that matches those specs, at a price around $19. The nearest I’ve gotten is again in India. A smartphone by the name of Microkey, model E12, (new phone from Nov 2017) is nearly the specs we want. 4.5 inch (larger than required) touch screen Android smartphone, 8 mp camera (better), with LED flash, 3G and WiFi. But this phone does not have GPS. And it is available locally for $1,799 India Rupees. That is $28 dollars, not $19. We have now failed to match the predicted price path. We are 50% above where it should be.
This means, first off, that it is nearly certain, that by year 2020, we will not see anyone (yet) selling a new smartphone that matches the 2010 iPhone 4 in specs, and costing only $10. It looks like we’ll fall short from that amazing price point, BUT we ARE on track for a smartphone year 2020, that costs about $20. And considering the collapse in price from $600, then whether it is $20 or $10 doesn’t really matter. We will have achieved a computer into essentially every pocket, even in the Emerging World.
Note, that the price decline now is proceeding at a slower rate than it has been before. Part of that is the inevitable end to Moore’s Law which some say is upon us, others say will shortly arrive. But in this case, new smartphones costing under $30, we do have the issue of the non-digital elements to the cost of a phone. The plastic casing for the phone is not experiencing a Moore’s Law. Plastic does not become cheaper by half every 18 months. Similarly the battery. Batteries are getting cheaper but nowhere near as fast as Moore’s Law is lowering costs of computing. Also labor costs do not fall by Moore’s Law. The shipping and marketing costs do not fall by the rate of Moore’s Law. So the final non-subsidised non-contract price of a new smartphone model right now, if the bottom end is at $28, only a small part of that price is the electronics of the ‘smart’ and ‘phone’ part. The other elements that do not live by Moore’s Law are now a far larger proportion of this cheap smartphone, than were parts of the original iPhone 4 back in year 2010.
NEW PRICE PREDICTION FOR NEXT 5 YEARS
It WAS a wild ride, for 6 years. We DID match, exactly the predicted price path, up to the summer of 2016. Now that ride has ended. Smartphones are STILL becoming cheaper, but now the price shrinks more slowly. It literally was cut in half every 18 months since 2010, now it seems to be cut by one quarter, every 18 months (ie the price drop rate has fallen into half). This suggests, that going forward, we could expect this type of price pattern at the very lowest end of smartphone prices:
2017 Winter cheapest new smartphone $28
2019 Summer cheapest new smartphone $21
2020 Winter cheapest new smartphone $16
2022 Summer cheapest new smartphone $12
2023 Winter cheapest new smartphone $9
We should prepare for this pattern and now it is likely in year 2020 we won’t find $10 clones of the 2010 iPhone (by tech specs) and instead, that year we’ll find the first phone at $20, roughly matching the specs of our 2010 iPhone. And if this slower rate of price drop continues, then expect the first new $10 smartphone to be upon us around year 2023.
PS obviously we ignore the nonsense fake phone deals that never materialized. Secondly, these will not be ‘as good’ smartphones as what the original iPhone 4 or Nokia N8 or Blackberry Bold etc were in year 2010. The cheap clone ‘superphone’ will be very cheap plastic and may have all sorts of quality etc issues. But I am not considering discounted old models sold at prices to get rid of inventory. I mean it has to be a new model that year, that is sold at a fair price in its market. And while we won’t see the $10 smartphone yet in year 2020 we seem to be getting gosh-darned close to that, at $20. And that the $10 smartphone is likely postponed to year 2023. Still, only 5 years from now....
Even when a precise forecast isn’t coming perfectly to form, I am so happy to be able to revisit old forecasts to see how they are doing. And this one, once again, was a remarkably useful forecast in year 2010, when there were no smartphones really below the $200 level. To be able to give guidance to the industry, that prepare - in a decade, smartphones will cost a tiny fraction of what they were costing back then...
So for the record. I am now shifting my forecast. No $10 iPhone-clones in year 2020 (yet) but the new target is $20. And the $10 iPhone-clone is likely to arrive around year 2023.
@Abdul,
If only they would put in bigger batteries...
Posted by: Wayneborean | February 01, 2018 at 02:07 AM
Hi Jim
I'm currently modeling 1.55B smartphones total for year 2017. Would be 5% growth rate year-on-year. Not bad but nothing like it was in the hot years..
Tomi Ahonen :-)
Posted by: Tomi T Ahonen | February 01, 2018 at 09:36 AM
Hi Tomi,
5% sounds about right. What do you think of the prospects going forward? Is this a bad year or what we should come to expect.
Posted by: Jim Glue | February 01, 2018 at 03:37 PM
Remember, Moore's law is an observation that ever 3 years it is possible to move to a new process node that results in a doubling of transistor density.
Prior to approx 2008 this could be harnessed to double performance in 3 years, as the extra transistors could be used for things like speculative execution, and also made higher clock frequencies possible.
Since 2008 there has been a move to increasing the core count, but performance of conventional (non-parallel) software has not increased very much.
Sheer economies of scale are - IMO - the main driver in falling smartphone costs. Perhaps the main reason prices are unlikely to hit the $10 floor is that for most people - even in the developing world - $20 is cheap enough and gets the purchaser a better-than-iPhone4 spec phone.
Posted by: MH | February 15, 2018 at 10:23 AM
In Estonia, basic phones sell new for 15€. Used basic phones may cost even less.
The cheapest smartphone with Tomi's specs costs 37€ — with tax, but not delivery — from a seller with the easiest spec selection UI. It's an Allview P42 8GB (dual-SIM), with a 4" screen, but with only 512 Mb RAM and Android 6.0.
It really baffles me, why manufacturers skimp on RAM so much, even today.
At least there's an Allview P41 eMagic 8GB (dual-SIM), which has 1 GB of RAM, an goes for 49.95€.
In addition to these, four devices specced by Tomi were in the 40–50€ range, five between 50–60€.
What I can observe, is, that many (but not most) emerging smartphone markets have also reached saturation, so now it's down to specs.
The used phone market should be able to supply reasonably-specced phones for as much or even less.
Posted by: Mart | February 21, 2018 at 02:06 PM
War between google and apple?
Iphone x or google pixel xl 2?
Posted by: Healthifyme | September 07, 2018 at 03:35 PM
$10 iPhone! Is that joke or really?
Posted by: Mike Rooney | September 08, 2018 at 04:21 PM