Apple ups orders for 5nm chips from TSMC, presumably for new iPhone

AppleInsider

Chipmaker TSMC has lost a significant order from Huawei, only to find Apple buying up all of its production capacity, for what is believed to be the "iPhone 12."

Apple has long been using supplier TSMC to manufacture its A-series processors, specifically 7-nanometer ones as used in iPhone XS's A12. Apple has also been known to be developing a 5nm processor with the company, and now industry sources report that orders have increased.

According to the China Taiwan Economic Daily, Apple has asked TSMC to produce almost 10,000 more processors in the fourth quarter of this year. The use of a 5nm processor and being produced in that quarter suggests that the order is for the A14 which is to be used in the iPhone 12.

The move reportedly follows Huawei's cancellation of approximately that same amount. While there is no indication from either TSMC or Huawei why the order was reduced, the phone maker is currently facing new criminal charges in the U.S., and has been substantially affected by the coronavirus.

TSMC is now said to have adjusted plans for its 5nm process because of the impact of COVID-19 on demand, though reportedly it remains on schedule.

The China Taiwan Economic Daily report citing unnamed industry sources, was first spotted by Chinese-language site IT Home.

danvdr

10,000? Seems like a pretty small number.

tmay

danvdr said:

10,000? Seems like a pretty small number.

I'm thinking 10,000 wafers. For a 300 mm diameter wafer, that's probably 20 to 30 million processors total, depending on die size and yield.

joeljrichards

I think we'll see variants of the A14 processor in more than just the phones so it may that Apple is thinking ahead to other refreshed lines: iPad Pros, iPads, Apple TV, and maybe MacBooks.

tht

tmay said:

danvdr said:

10,000? Seems like a pretty small number.

I'm thinking 10,000 wafers.

As a first order approximation, each wafer is 300 mm in diameter and a typical Apple iPhone SoC is 100 mm^2, those 10k wafers equate to about 7 million SoCs. There will be losses so it would be something like 5m instead. 

Apple will sell on order 40m to 50m A14 SoC models in cal Q4 2020, so not a lot, but this is on top of what Apple has already ordered. Always nice to have more capacity. 

Would be nicer if those wafers were for Mac ARM processors. 

avon b7

The term I have been hearing is '10,000 pieces of production capacity'.

At the start of the year Huawei stated its consumer division would have a difficult year. 

Then CBU sales were up in January, down in February and up again in March.

Then three days ago Huawei's CEO of CBU went on record as stating he expected growth rather than contraction for 2020.

That makes this rumour anybody's guess although given how the year has started, it is perfectly feasible. 

tjwolf

tmay said:

danvdr said:

10,000? Seems like a pretty small number.

I'm thinking 10,000 wafers. For a 300 mm diameter wafer, that's probably 20 to 30 million processors total, depending on die size and yield.

Completely agree, but the article specifically said "processors".  Proof-reading seems to no longer be part of the publishing process.

seanismorris

Devices don’t necessarily role out globally at the same time, so Apple having the ability to accelerate the rollout is handy.

They might also use the additional 10,000 for other devices like the ARM (A Series) MacBook.

ericthehalfbee

tht said:

tmay said:

danvdr said:

10,000? Seems like a pretty small number.

I'm thinking 10,000 wafers.

As a first order approximation, each wafer is 300 mm in diameter and a typical Apple iPhone SoC is 100 mm^2, those 10k wafers equate to about 7 million SoCs. There will be losses so it would be something like 5m instead. 

Apple will sell on order 40m to 50m A14 SoC models in cal Q4 2020, so not a lot, but this is on top of what Apple has already ordered. Always nice to have more capacity. 

Would be nicer if those wafers were for Mac ARM processors. 

The A13 is just under 100mm, but who knows how big the A14 will be (it SHOULD be smaller, but by how much we don’t know).

With Huawei cancelling I guess it means they can’t “pretend” they were the first to market with 5nm, like they did with 7nm).

Beats

avon b7 said:

The term I have been hearing is '10,000 pieces of production capacity'.

At the start of the year Huawei stated its consumer division would have a difficult year. 

Then CBU sales were up in January, down in February and up again in March.

Then three days ago Huawei's CEO of CBU went on record as stating he expected growth rather than contraction for 2020.

That makes this rumour anybody's guess although given how the year has started, it is perfectly feasible. 

I think you're on the wrong website.

Try iKnockoffInsider.com

CheeseFreeze

5nm... incredible how advanced the production process of these chips had become. Crazy.

tmay

avon b7 said:

The term I have been hearing is '10,000 pieces of production capacity'.

At the start of the year Huawei stated its consumer division would have a difficult year. 

Then CBU sales were up in January, down in February and up again in March.

Then three days ago Huawei's CEO of CBU went on record as stating he expected growth rather than contraction for 2020.

That makes this rumour anybody's guess although given how the year has started, it is perfectly feasible. 

There isn't a chance in hell of growth in 2020; that's just fantasy, and I don't believe a word about growth in March, when China, Huawei's largest market, was deep in the COVID19 pandemic.

tmay

tht said:

tmay said:

danvdr said:

10,000? Seems like a pretty small number.

I'm thinking 10,000 wafers.

As a first order approximation, each wafer is 300 mm in diameter and a typical Apple iPhone SoC is 100 mm^2, those 10k wafers equate to about 7 million SoCs. There will be losses so it would be something like 5m instead. 

Apple will sell on order 40m to 50m A14 SoC models in cal Q4 2020, so not a lot, but this is on top of what Apple has already ordered. Always nice to have more capacity. 

Would be nicer if those wafers were for Mac ARM processors. 

I agree with your numbers more so than mine, as I was extrapolating from an article I saw on SemiWiki, but not accounting correctly for the A series die size.

tmay

tjwolf said:

tmay said:

danvdr said:

10,000? Seems like a pretty small number.

I'm thinking 10,000 wafers. For a 300 mm diameter wafer, that's probably 20 to 30 million processors total, depending on die size and yield.

Completely agree, but the article specifically said "processors".  Proof-reading seems to no longer be part of the publishing process.

Editing/proofreading seem to be the red headed child of AI.

avon b7

tmay said:

avon b7 said:

The term I have been hearing is '10,000 pieces of production capacity'.

At the start of the year Huawei stated its consumer division would have a difficult year. 

Then CBU sales were up in January, down in February and up again in March.

Then three days ago Huawei's CEO of CBU went on record as stating he expected growth rather than contraction for 2020.

That makes this rumour anybody's guess although given how the year has started, it is perfectly feasible. 

There isn't a chance in hell of growth in 2020; that's just fantasy, and I don't believe a word about growth in March, when China, Huawei's largest market, was deep in the COVID19 pandemic.

I have no idea. 

Last year people were saying the same because of Trump and they recently released full year numbers that proved those people very, very wrong:

https://www.androidheadlines.com/2020/04/unstoppable-huawei-phone-shipments-up-40-million-in-2019.html

This year, it was Huawei who lowered the sales bar estimations only to increase them three days ago. 

It's impossible to know today how things will pan out. Too much is happening. 

snow66

It is going to be interesting to see how Apple takes advantage of their huge cash reserves over the remainder of the calendar year.  While some competition has been living on razer thin profit margins during a booming worldwide economy, Apple has used their 38% margin to create a rainy day fund large enough to continue full operation for years without selling a single iphone.  I would love to see Apple put the petal to the floor and speed up innovation at a time when much of the tech world are worried about how to survive until the world economy rebounds.

Going forward with 5nm is a start.

danvdr

snow66 said:

It is going to be interesting to see how Apple takes advantage of their huge cash reserves over the remainder of the calendar year.  

You're right. Apple is uniquely positioned to jump ahead of the competition.
Also, I'm guessing that Apple has been buying a whole bunch of AAPL.

mdriftmeyer

tht said:

tmay said:

danvdr said:

10,000? Seems like a pretty small number.

I'm thinking 10,000 wafers.

As a first order approximation, each wafer is 300 mm in diameter and a typical Apple iPhone SoC is 100 mm^2, those 10k wafers equate to about 7 million SoCs. There will be losses so it would be something like 5m instead. 

Apple will sell on order 40m to 50m A14 SoC models in cal Q4 2020, so not a lot, but this is on top of what Apple has already ordered. Always nice to have more capacity. 

Would be nicer if those wafers were for Mac ARM processors. 

Take your 7 million and multiple by .80 or 80% wafer viability and you get 5.6 million viable SoC A14 chips. The 7nm is just over 75% average yield efficiency and TSMC themselves stated 5nm is just at or slightly above 80% average yield efficiency.

prismatics

seanismorris said:

Devices don’t necessarily role out globally at the same time, so Apple having the ability to accelerate the rollout is handy.

They might also use the additional 10,000 for other devices like the ARM (A Series) MacBook.

I hope not. Apple going full ARM will kill the last 1% non-consumer of Apple. Me as a developer would certainly not welcome having to shift to other platforms. But eh, people will start to make Linux distros that are not shit finally

crowley

snow66 said:

It is going to be interesting to see how Apple takes advantage of their huge cash reserves over the remainder of the calendar year.  While some competition has been living on razer thin profit margins during a booming worldwide economy, Apple has used their 38% margin to create a rainy day fund large enough to continue full operation for years without selling a single iphone.  I would love to see Apple put the petal to the floor and speed up innovation at a time when much of the tech world are worried about how to survive until the world economy rebounds.

Not sure how much of an advantage cash reserves will be with interest rates near zero the world over.

avon b7

crowley said:

snow66 said:

It is going to be interesting to see how Apple takes advantage of their huge cash reserves over the remainder of the calendar year.  While some competition has been living on razer thin profit margins during a booming worldwide economy, Apple has used their 38% margin to create a rainy day fund large enough to continue full operation for years without selling a single iphone.  I would love to see Apple put the petal to the floor and speed up innovation at a time when much of the tech world are worried about how to survive until the world economy rebounds.

Not sure how much of an advantage cash reserves will be with interest rates near zero the world over.

Also, if Apple hasn't brought many new products to market already it is unlikely to do so now. It's been hoarding money for years and the competitors that matter aren't really operating on razor thin margins anyway. 

If some competitors fall by the wayside, their clients are more likely to go to other Apple competitors, further consolidating the market. 

That is without even considering the 'frenemy' aspect which means Apple needs some of its competitors to be able to get products to market, shoring up revenues for those competitors. 

There are a couple of products gaps in Apple’s business IMO but they are gaps where competitors already have established products.