I'd love to be a fly on the wall during these negotiations. Just to see what the heck they even talk about, what information each party comes armed with, and what kinds of concessions get offered.
I mean it isn't just a "we'll lower this" I bet there's lots of juicy confidential future roadmap stuff, things like "Ok we'll help you invest in this _____ tech and machinery but you have to knock off half a cent" or who knows what!
TSMC: you'll pay _____ or you can go to the other fab, oh right, you'll pay _____
--------
vendorSellingToBigBox: we would like _____ per unit of our product
bigBoxBuyer: that's nice. we'll pay 40%______ per unit. you'll sell extremelyLargeNumberOfUnits, and make it up in volume. oh, if you don't sell a minimum number, you have to buy all of the unsold items back from us. you'll also get the worst shelf space available compared to your competitors. i'm sorry you thought this was a negotiation.
The TSMC fabs aren't idling. Between AMD, Intel, nVidia, Broadcom, Google, Amazon, etc. They've got customers with more demands for their chips than they have capacity to provide.
If apple doesn't go to them, I think the only other option is Samsung? But I don't know that Samsung's fabs are open to third parties.
Well, the facts are a little different. Apple is 25% of the volume on the latest nodes and Intel just had an awkward conversation with TSMC saying "Oh, those orders, no, we actually meant the year after", so Apple very much has TSMC "by the balls".
Samsung has produced parts for others for a Very Long Time, most famously NVIDIA (not "nVidia" since many decades) use(d) them.
> If apple doesn't go to them, I think the only other option is Samsung?
Shorter-term that's true.
Apple is well known at times to think longer term on logistics, strategy. If push comes to shove, there are two additional options for Apple.
1) Political. Taiwan not being China depends on the support of the US. Apple is a economic security prize for the US, a strategic asset. Apple can get powerful politicians in the US to lean on Taiwan, which can dictate TSMC's behavior to an extent. This would prompt increased compromise from TSMC (3-4% hike, not 6%).
2) Apple starts investing into fabs and builds a TSMC competitor from the ground up, perhaps in alignment with ASML (Europe very much wants an improved semiconductor context) and a few US partners (build the fabs in the US and Europe for national security purposes). Apple is one of the few companies where you have to take this potential seriously, they can safely bankroll it entirely by themselves if necessary.
It can get worse. At one point Walmart mispriced a product then came after my company for the loss. That was an awkward moment. Edit: even worse, now customers expect the lower price, eroding your pricing power.
Walmart will also do this intentionally to their vendors, or engage in what I call temporal arbitrage.
It hurts Walmart as a business very little to stop selling Vendor X goods on their shelves when there are substitutes available. However, for Vendor X it's often the case that Vendor X cannot adjust their output as quickly as Walmart can take the product off shelves. That action, taking them off the shelves for even a short period of time while Walmart presses for renegotiating terms can result in layoffs, loss of revenue could result in lines of credit drying up. It's an existential threat to a company because they can't scale up and down at the speed Walmart demands.
If I recall correctly there's a record of this happening with Rubbermaid, and they responded to Walmart's demands for prices by cutting costs, cutting corners with quality, and ultimately sullied their brand name.
You should, because this kind of behaviour ripples through so many companies, that I wouldn't be surprised if you actually not on the receiving end of things.
Retailer lowballs the manufacturer. The manufacturer then needs to lowball someone too, probably starting from its employees. Also the supplier who can be anything from local hardware store to some IT giant.
I think another upshot is that it ends up making all consumer products "lowest common denominator" consumer products. It's a vicious cycle:
Step 1: Company has a great product. Walmart begins stocking it nationwide, perhaps promoting it to increase their marketshare. That product's revenues explode, they expand, everything seems great.
Step 2 to N: Then Walmart threatens to pull from shelves unless they can meet the cost of $substituteGood. The company cuts costs to make it work. Rinse and repeat.
Step N+1: The product is no longer great.
This is exactly what has happened with Rubbermaid and with countless other consumer goods.
There's a book called The United States of Wal-Mart that details a lot of the damage that Wal-Mart has done. I wouldn't be surprised if there's a follow up for Amazon.
Apple is famous for pre-buying and covering upfront costs for new node processes at TSMC. TSMC has the best tech, but apple has the biggest purse. Even apple deciding to stay on existing nodes and not fund the latest one can probably affect their finances, especially when apple is 25% of their income
You can actually deduce the vendors that have influence on the big box buyers if you look carefully enough. And some of the vendors will tell Walmart/Costco to take an airborne intercourse course against rolling pastries.
Others will sell a white-label or generic brand to them. After all, it's just business.
Look for large big box stores that conspicuously have a product line absent, or only have a very defined subset of the product available.
For awhile iPhones were just completely missing from some larger stores, but it seems Apple has the upper hand there now and they're back in business. And even now Apple holds a very right reign on the prices charged (Best Buy seems to have the most leverage here, as they often sell new MacBooks for a bit lower than the Apple Store). This is one of the reasons Apple started the stores; gives them leverage against the retail giants.
There are a number of brands you'll never see at Costco; they don't play the game.
Supply chain shortages means you used to see an item, but it is not available at this time. Vendor arrangements means you've never seen Del Monte can corn at Costco, but you've seen Kirkland can corn that tastes just like Del Monte's.
> TSMC: you'll pay _____ or you can go to the other fab, oh right, you'll pay _____
It wouldn't be easy for Apple to just go to another fab. They'd have to find another vendor with the capacity and then they'd have to tweak things to match the new fab's design rules. Not trivial for something like M* chips. I suspect Intel would be happy to have Apple in their nascent foundry biz, but not sure they would have the capacity right away and also suspect the design would require a good amount of tweaking to get it in production in an Intel fab.
That's the point of the comment I made. TSMC knows that Apple can't just jump ship to another vendor. There isn't another vendor that can do what Apple needs. Currently. As long as that is true, Apple is stuck at TSMC.
Apple could say that they will accept the 6% increase, but then lower their investment in new fabs 6%. I honestly don't know which number is bigger.
Remember that Intel's fabs are waiting in the wings and that I would bet dollars to donuts that Intel is talking with Apple.
TSMC's position in that context is less strong than you'd imagine, especially since Apple has to be thinking that they want to diversify away from Taiwan and China to avoid "Russia risk" of having their supply shut off to zero suddenly when China does something crazy.
No way they could switch in under 5 years. TSMC provides design guidance and knowledge that will be deeply embedded in the design of their chips. It’s not like a print shop where you can send over a pdf.
I agree with the diversity angle though. Tim Cook seemed really slow to recognize it, but the last few years apple has accelerated shifting manufacturing. India will now be joining China in producing their latest phones (14), which is a big first.
Murdoch e-mails are nice in that they nicely highlight the multiple fronts nature of negotiating where you can often find compromise by shifting how you approach the problem.
The way Jobs closes it is quite unusual however because in this case NewsCorp were late movers. Apple had already reached terms with all their competitors and could put on the table a take it or leave it offer which Murdoch couldn’t refuse.
I was part of the team in charge on negotiating the termination of an on-going contract involving a potential settlement in the low teens millions of dollars between two huge industrial companies once. Obviously it’s nothing as big as a deal between Apple and TSMC but still not an insignificant sum.
I was surprised by how straightforward the whole thing went. Basically we had prepared our arguments: we knew the minimum payment we could justify, the maximum we thought the contract allowed them to ask, what we were aiming for and how much we estimated going to court would cost us and them. They had done the same. There were a couple of back and forth where one party made an offer, the other countered with another one explaining why what was asked didn’t apply and why it should rather be like so and so. At some point, we said that’s our last offer we won’t move anymore. They decided to agree and that was it.
What you witnessed proves bike shedding is real.
I'm in much smaller league when it comes to budgets, but lets just say I had longer arguments about a budget for a filing cabinet than a half year software project.
> I was surprised by how straightforward the whole thing went
Executives value their own time over anything else, and as such - they don't find it valuable to haggle over small details in lieu of just getting the deal done.
An even better equivalent would be the Apple-Microsoft negotiations in 1997, around Microsoft Office, IE-as-default, non voting shares, etc. I think TechEmails published a bunch of these.
Yeah absolutely. Another huge thing I got some window into at Qualcomm is technical validation. So Apple is going to be looking for proof that the chips deliver and TSMC is going to be pushing back against high expectations on that front.
Qualcomm is absolutely brutal as a vendor. They pursue % of revenue deals instead of selling chips at $Y. They can only get away with it for lte/5g because Broadcom and others totally dropped the ball.
> Most likely middle managers/directors emailing each other.
Pretty much like sovereign nations negotiating treaties.
Which is an apt metaphor, because neither Apple nor TSMC has the ability to force the other to do anything.
So, negotiations become a crawl to find a mutually-acceptable local optimum. Which ends up with both getting things they see as strategic priorities, while compromising on things they care about less.
Most of the dance is around figuring out what the other's strategic priorities actually are, and which concerns are secondary.
Or they have mutually-exclusove strategic priorities, in which case there is no deal.
Having worked at a company that got an Apple deal... It usually isn't a level bargaining table... And it isn't unusual for no profit to be made whatsoever on the Apple deal, because merely the rumour that Apple has signed a deal with your company has so many other benefits.
TSMC is in a stronger position than most suppliers, because they have really strong demand for their product and don't need the prestige that an Apple deal offers.
But Apple needs TSMC. Not saying you’re wrong, but TSMC also has leverage. TSMC doesn’t want to price Apple out, but at the same time, Apple can’t just pack up and go to a competitor. At the nodes they’re working at, going to a competitor could require a complete relayout with new design rules.
Apple 100% does need TSMC at this point and there is leverage both ways.
I suspect if Apple really wanted to they would probably start to produce their products in two lines like Nvidia did, with cheaper phones produced on something like Samsungs 4nm and the higher end phones on TSMCS 3nm.
This has the added benefit of diversifying their supply chain (to a degree), giving TSMC less leverage over Apple and potentially reducing the production cost of the lower end phones.
> I suspect if Apple really wanted to they would probably start to produce their products in two lines like Nvidia did, with cheaper phones produced on something like Samsungs 4nm and the higher end phones on TSMCS 3nm.
Wouldn't be the first time Apple has used multiple vendors for their SOC. The 6s for example had SOCs made by both Samsung (14nm) and TMSC (16nm).
No, but they can use their massive war chest and expertise to ramp up a competitor or become one themselves over the next decade or two. With massive corporate ships like TSMC and Apple it takes a long time to change course but they work on longer timelines so a big negative shift over a long time scale is an alarming signal.
Building your own semiconductor fab with good yelds that are profitable is monumentally more complex than all of those things combined. Why do you think only Intel and Samsung are the ones left at the cutting edge anf even they are struggling to keep up with TSMC?
If you think Apple can just buy their way into fab success, you don't understand the complexities of semiconductor manufacturing.
> If you think Apple can just buy their way into fab success
No one is saying they can "just" buy their way into fab success. However, I think they do have the cash and clout to make it happen on a decade long timeline, possibly in league with Intel/Samsung or the US Government, or all three, etc. Obviously that would be enormously costly and risky, but any move to eventually do so is a threat to TSMC, so even if they don't do it they have a negotiating position with TSMC right now because the relationship is enormously valuable.
> Sure, but how are they going to produce iPads and iPhones before then? They can't have more than a year's worth of CPUs stockpiled.
From TSMC chips. Basically if TSMC insists on raising prices to levels above what Apple is prepared to pay they can threaten to move away from TSMC eventually. If TSMC refuses to budge then Apple will have to pay the higher prices in the short to medium term and then expend capital to not be beholden to TSMC over the long term. So then the question is does TSMC value the relationship over the long term or do they value the higher prices in the short to medium term? That could be Apple's negotiating position.
Apple can't easily go to a competitor. They have the cash on hand to play kingmaker though. They could go to one of TSMC's competitors and make them into the new TSMC. TSMC's current position is largely due to Apple fronting huge sums of cash for them to develop new nodes. They get guaranteed sales at a generous price for the first few million chips off the line. There's few other players, if any, that could offer TSMC a deal like that.
Apple could with some pain replace TSMC but TSMC can't really replace Apple. While Apple is a demanding partner they pay their bills on time and can and do front R&D money. Once Apple's deliveries are made TSMC now has a money printer on the new process node Apple paid for.
I don't understand this comment: even if there were times in past years when AMD CPUs were significantly less performant and more energy-hungry than Intel CPUs, Macs with AMD CPUs that would have sold at least well enough to keep the Mac alive as a product line.
I don't think you can just drop a pile of cash and create a competitor to TSMC. Even some state actors can not achieve that. I think TSMC has leverage here, and it will have for long time to come.
Of course you can. China is doing it right now without ASML even. There is nothing special about TSMC here. It's just funding and time. If Apple wanted to I wouldn't doubt they could reach cutting edge tech level within 5 to 10 years. It would cost enormous amounts of money and would probably be a very bad business move. But technically there is no problem.
China is proof that you can not just create competitor to TSMC. China throws billions for decades and it still behind TSMC. They already lost chip war, that's why they are trying to catch up in some other new technologies - nuclear (fusion), AI, space etc.
China is proof you can just create a competitor of TSMC. It's simply a matter of poaching their underpaid talent.
> China throws billions for decades
PRC's first pseudo-serious semi push was 2014 big funds, so far they've spent about 50B, over 8 years. That's less than annual R&D of western semi sector. And basically for first 5 years it wasn't a very serious effort until Trump's Huawei sanctions. PRC semi poached about ~3000 semi engineers from TW, a few hundred top talent from TSMC, and that was enough to realize the recent 7nm breakthrough on DUV in the last couple years once PRC semi started to get serious. US companies wouldn't have any issue accessing western equipment and US could absolutely pressure TSMC / TW into further semi capitulation. So maybe dropping cash and add political sticks to brain drain TSMC. The secret isn't to recreate TSMC, but to paperclip all their people and get to to build US-SMC state side.
> lost chip war
That remains to be seen, so far there hasn't been any sector PRC couldn't successfully indigenize in relatively short order once central gov started turning the screws. It might not be leading edge, or even commercially competitive, but it will likely be enough. They could also destroy critical East Asian semi supply chain before US can reshore in next few years and be the only country with the most comprehensive indigenous semi supply chain even if on older nodes. Still lots of ways chip war can play out.
China is proof that you can do it within a decade. China at worst started a decade ago and for the majority of that time wasn't pushing hard, until Trumps sanctions hit. Now they have 7nm DUV. I don't see how this is anything but a huge success for China.
It may take time, and it may take more money than people want to spend, but I don't see how it can cost much more than half a trillion dollars, since you could just buy TSMC for less than that.
1) TSMC is not for sale. It's of national importance to Taiwan.
2) There is only one TSMC. If you could just buy one for 500bn the EU would have bought 3 already. Building your own TSMC from scratch takes a decade, if not longer.
The Chinese government, whose cash piles are even larger, has been unsuccessfully trying to do that for decades (i.e., produce semiconductors on the Chinese mainland high-end enough to go into a notebook computer that can compete on the global market).
They are not desktop or laptop chips though. Targeted at miners. If you are willing you could probably reverse engineer them and do some fun stuff. But they are definitely not general purpose.
They’re not quite equal. Apple needs TSMC but not newest node tsmc. TSMC needs apple to fund newest node. But apple can’t walk away entirely, because they need something from TSMC.
Apple is probably weaker because it’s likely someone else can fund the newest node for TSMC, just might take more work.
I'm half expecting Intel to start producing chips for others in the coming years. However I doubt that it's as simple as sending Intel the blueprints for the M3, there has to be some re-tooling on both ends I believe.
The conspiracy theorist in me wonders if they’ll join a secret lobby agitating for war between China and Taiwan, if Apple can’t get 6% off TMSC then no-one can. Then build a US based plant subsidized with taxpayer money.
How is this approach working out for China? Not snark, I genuinely want to know what resident experts think of the progress made by throwing vast sums at SoTA node fabs.
Clearly there's a level that Apple wouldn't pay (e.g. 100000000% hike), so there are definitely options up to and including cancelling plans to buy such chips from TSMC.
"The United States spends more on national defense than China, India, Russia, United Kingdom, Saudi Arabia, Germany, France, Japan, and South Korea — combined."
!?
And $200B is inside the consensus estimate of $200B to $300B. Wow.
Not to forget, Apple also controls the app store and billions of devoted fans with iPhone with push notifications /g. Seriously I am wondering when will we see a corporation go to war in a kind of futuristic Opium War sort of battle.
I'm not a fan of comparisons like that because they don't account for cost or standard of living between counties, just raw exchange rates. It would be especially pronounced when including countries like India and Russia.
You're right, but when the numbers are so lopsided on the US' favour, there's no way in hell that cost of living can account for such a stark difference.
I mean, the US has 20 aircraft carriers and the rest of the world has 28. The proof is in the pudding.
I have doubts that a fly posseses sufficient neuronal capacity to parse human language. Furthermore I'm also unsure if the hearing system of a fly would be sophisticated enough to capture the nuances of spoken language.
My reading of this is that Apple is happy to wait a bit on next-gen given the macro conditions. However if they just admitted that it would be a big hit to image. So instead they are looking to create scapegoats.
The most plausible scapegoat is the cutting-edge one with a price hike because there is maximum room for others to speculate on what isn’t right/good enough, and there is minimum risk of giving up their spot in line.
By the time competition is up on it, Apple can return to their most familiar location: somewhat behind on the bleeding edge of tech, but ahead on integration and UX.
It’s also possible that Apple is going to take fab in-house but I’m not aware of any signs of preparation and something like that seems like it would be impossible to keep secret.
>By the time competition is up on it, Apple can return to their most familiar location: somewhat behind on the bleeding edge of tech, but ahead on integration and UX.
Seems like a good tradeoff. Is the next years best TSMC chip really that much better than last year's? I doubt it.
> N3 technology will offer up to 70% logic density gain, up to 15% speed improvement at the same power and up to 30% power reduction at the same speed as compared with N5 technology.
That is a more modest bump than some previous nodes, but still seems worth a 6% increase in price depending on how yields shake out.
All those "up tos" can sometimes add up to "actually not much".
And if Apple is currently a "generation ahead" they can spend that lead-time in other ways. They've done it before out of necessity, and I wouldn't be surprised if Apple has phone features they're holding back to "gloss over" a release that doesn't really have much more CPU.
No. It's impossible to trade (even infinite amounts of) money for calendar time or supply access.
Any moves Apple made now would wind their way through ASML et al., through fab construction and qualification, and then finally adapting their chips to a new process.
They could pull the trigger on the Power to Intel because they'd spent a decade+ of contingency planning, OS/app support, and acquisitions.
Can we be sure Apple hasn't been quietly building a fab or at least fab expertise somewhere in the bowels of Cupertino? Do we have enough insight into ASML to be sure they haven't purchased even older fab equipment?
fabs take years to build. The only real option Apple would have is some kind of joint venture with Intel and maybe try to bring in the US Department of Defense or US government in general to push the whole chip national security talking point and get more funding
TSMC has a market cap of $356B and maybe $30B in cash. Apple has a market cap of $3.25T and over $200B in cash. Surely you are overlooking a possible option here, which is Apple acquiring TSMC. I don't see it as likely, but it is an option.
I really doubt that would be allowed to go through due to the conflicting interests it would create regarding chips produced for all other competitors (AMD, Google, Qualcomm etc). It would be seen as similar to the NVIDIA buyout of ARM.
Well, I think that's their leverage (if they have any). They kick the can, wait, take the hit, and then in the meantime make it clear that next time they'll have their own fab. In that scenario TSMC is looking at demanding a 6% hike vs jumpstarting competitors in other countries.
Maybe it's unrealistic but it wouldn't be the first time something like that happened in a market. The political conditions aren't unfavorable either.
> bring in the US Department of Defense or US government in general
This is a guaranteed route to introducing even more delays, cost overruns, schedule overruns, onerous requirements, intrusive oversight, and about a trillion meetings and studies along the way.
If you want to go fast, stay far far away from the USG.
> Apple has rejected these offers. [...] This may be one reason why we reported that the upcoming M2 Pro and the M2 Max for updated MacBook Pro models would be manufactured on the 5nm process instead of 3nm; both entities might not have reached a middle ground for price agreement.
This makes no sense?
M1 / M1 Pro / M1 Max / M1 Ultra were all manufactured on TSMC N5, even though volume production for N5P (improved N5) was already available by the time M1 Pro / Max were released. In fact, A15 (N5P) was released in September '21 whereas M1 Pro / Max (N5) were released in October '21 and Ultra was released in March '22.
M2 is manufactured on N5P, so it makes total sense that M2 Pro / Max / Ultra will also be manufactured on N5P. They're part of the same generation, so it makes sense to produce them on the same node.
This is also most logical from a financial standpoint: the base model M chips and the A chips are the smallest and therefore least expensive to produce on a new node with relatively poor yields and high prices. Once the node has matured a bit more (higher yields / lower prices), you manufacture the bigger chips (Pro / Max), and even later you print the big bois (Ultra)
TSMC charging 6% more for the 3nm process seems reasonable. Article notes Apple accounts for 25% of their revenue and says this gives them leverage over TSMC but aren't they the only 3nm player in town?
> charging 6% more for the 3nm process seems reasonable
I imagine the negotiations largely centre around timeline and quantity guarantees, as well as capability. Apple is cutting back iPhone Pro production [1]. If 3nm won't enable features that sell more iPhones, or if the market is shifting from a feature focus to one on costs, 6% more for 3nm may make sticking with five more profitable.
No, iPhone Pro demand is very strong. The regular iPhone demand appears soft or didn't ramp up. None is this is a surprise since early adopters would likely be the same who buy the Pros.
I think the iPhone 14 Plus was a huge miss for them. Anyone who wants a bigger phone probably also wants the Pro. Wouldn't be surprised if there is no iPhone 15 Plus.
I think this is way premature. The non-Pro iPhone is for the people who don't watch Apple keynotes and don't wait for the Apple Store to open up for pre-orders. It's for my mom, who will buy a new iPhone when she breaks her old one.
The iPhone 14 Plus hasn't even launched yet, so kind of hard to tell either way. I think it'll be one of the more popular phones for the walk-in crowd who only replaces their phone when their existing one breaks.
Yep, but TSMC don’t manufacture the main lithography machines: that’s typically solely or at least collaboratively from ASML in the Netherlands. So if Apple really wanted to, they could move gradually away. And given their volumes, this is leverage.
Uh it's not that easy at all? This is a decade worth of work right there. European manufacturers also don't really have big quantity manufacturing. SMIC has been moving gradually towards lower nm fabs. It's odd that people here nowadays think, oh you just get better lithography and we're done. This a multistep process to begin with. Do you think Intels only problem is that they can't buy ASML components? Not to mention that all European manufacturing "giants"(in demand and importance not in volume) like ASML have quite limited capacity. During Covid there was a shortage of machines that produce fibres needed to create N95 masks for example.
Out of those multistep process Apple has done very large part of designing Chip, designing OS to use chip optimally. Apple user market is not the one where they compare Mac with Wintel laptop. They find out M2 is crap in every way compare to Intel or AMD chip and then go on to buy a HP Pavilion or some such.
Besides with Perf/watt gap Mac/Iphone and rest Apple can wait it out for at least a year or maybe even longer.
They bought companies that have designed chips for a DECADE, and then that took another decade of experience. And even if Apple went out and bought an old fab or global foundries it will take a decade to build out the capacity with a worsening economic crisis I'm not sure how realistic that is. But sure, it is possible.
I guess you are just missing the point that Apple can wait out for a year or longer and keep using 5nm TSMC chip with more core, components or whatever. Once the supposedly "premium paying" clients buy the best 3nm chip, Apple can still be fine buying the rest. There is no personal ego hurt, if it still does not workout after 1-2 year of negotiation lawyers can go sign new contract for higher price. It is not like "Guys, TSMC asked for more money, so stop everything and work on setting up chip making factory in Tennessee however long it takes".
Also worsening economic crisis will hit TSMC much sooner than Apple.
Yep, right after they hire about 30,000 people long-term to manufacture the chips, and spend $20B on plants, which would take 5+ years to complete.
It's almost as if TSMC asking or an extra 6% when inflation is at an all-time high is reasonable, and Apple doesn't really have a viable alternative...
Might want to double-check your inflation info, there. We're not close to an all-time high. We're not even at 50% of an all-time high. And also the concept of inflation not being the same everywhere.
That would probably take years to set up, and considering Apple outsource manufacturing in general (Foxconn for assembly, Samsung for screens, TSMC for chips) it wouldn't necessarily fit their modus operandi.
On the other hand they have more money than they could possibly spend, so insourcing everything for vertical integration isn't out of reach.
Even posturing to begin that process would knock tens of billions off TSMC's market cap and spike its funding costs, to say nothing about complicating its contract negotiations with every other customer. They did it, after all, to Intel.
I also just think Apple has planned for these contingencies which prevents them from having no way out of TSMC. Tim Cook is a supply chain genius and no way do they sabre rattle without a realistic, even if slightly painful, second option.
I'm not sure I agree it's particularly good leverage since what happens is Apple ends up buying up all of TSMC's production capacity pushing out other customers.
There is guaranteed business either way, just Apple is more of a guarantee since they are existing.
The other bit is how much Apple invests in TSMC's R&D, which could be a huge loss if Apple moves.
But all this is likely just negotiations and Apple won't be making a move (where to anyway?). Maybe instead of 6% per chip Apple invests a bit more elsewhere in the company to keep production costs as low as possible.
That's a massive undertaking. We're talking many years and many billions of dollars and puts Apple too deep into manufacturing I'd think. Not to mention I believe there is a massive wait list for ASML machines that I doubt even Apple could get in front of.
But will they keep taking that fab capacity over many years and fund the R&D of the next process nodes by buying up guaranteed capacity ahead of time? Apple has been a good partner to TSMC and it would not be wise to replace them for short term gain.
TSMC is not hurting for cash and can easily shoulder the risks - or sell bonds if they'd like prefer to keep the warchest full. There is a long list of customers who'd kill for a chance to get (additional) capacity of the leading edge node process for years at a time (AMD, Nvidia, Google, Amazon, and the rest of the companies giving TSMC the other 75% of revenue)
> There is a long list of customers who'd kill for a chance to get (additional) capacity of the leading edge node process for years at a time
Is there? Seems like we're heading for a recession and reduced demand - at higher prices I wouldn't be surprised existing customers cut back on existing capacity.
> Is there? Seems like we're heading for a recession and reduced demand
Perhaps lower demand for consumer devices[1], but at Google and Amazon's scale, getting more efficient Tensor cores, ARM processors, and AI accelerators for their respective clouds will save them some OpEx by requiring less power (and cooling). It may also save CapEx as well, by reducing the need to build additional datacenters by getting +10% more compute[2] with the existing DC footprint, power and cooling budgets.
Server costs are amortized over long periods - longer than whatever recession may or may not happen. Also, if they can lock-in TSMC for years, they'll continue to reap benefits long after the recession is over and shift capacity to Kindles or Pixel devices when consumer demand recovers.
1. Which may explain Apple developing cold feet.
2. Back of envelope calculations. I do not know what node mixes Google and Amazon have in their data centers, but I'm guessing it averages-out between 10 to 15nm. How much can they benefit by replacing their oldest processors with 3nm parts?
They will, but I'd guess they probably won't be willing to pay the higher price either (I believe the only reason Apple has the first batch of N3 capacity in first place is because they outbid the competition). So that likely wouldn't leave TSMC any better off.
> I believe the only reason Apple has the first batch of N3 capacity in first place is because they outbid the competition
There are many more possible reasons why Apple had dibs, including being able to guarantee specific volumes, or committing to pay large non-cancelable amounts as a deposit, and could be as pedestrian as TSMC preferring to deal with a single large order rather than 2 or 3 less large ones as they work out the kinks in a bleeding edge process.
So I assume Moore's law has always meant that you get more transistors per dollar in spite of all the cost increases that come along with smaller nodes, but now I wonder, how long would that continue if Moore's law is breaking down?
Apple probably pays a premium on the 200 million pieces they buy. Since no one else apart from Samsung is making signficant money from Android, it'd be somewhat hard to cover the loss. PC volumes are lower, a lot of them are low-end, and Intel uses their own fab for most of their CPUs.
>> Can't they just threaten to sell the space to someone else?
You mean TSMC threaten? The only other huge players willing to pay a premium for state of the art node like that are GPUs, and they just got the whole crypto mining rug pulled out from under them so they may not be interested in paying top dollar any more.
My guess would be AMD; they'd love to have server processors even farther ahead than what Intel is able to fab themselves and the large margins for server processors are able to absorb TSMC's asking price.
My thought is margin is low on apple but volume is high. They also are a good partner helping to push the process forward. AMD and NVIDIA are probably more profitable than apple for tsmc but volumes are much lower.
I also think the ability to improve tier prices as fast as they have is directly related to apple’s volumes.
It would be interesting to see what would happen if TSMC could fill the capacity with other orders. Quite possible, as it seems to be the most sought chipmaker.
What would Apple do in that case? Samsung doesn’t come close to TSMC just yet.
They're probably both flexing. TSMC is so valuable because Apple's reserved high end high volume orders allows them to expand capacity, Apple probably also is a tester for new tech.
They'll both come to terms in some middle ground. Apple has history of such flexing to force other parties to play nice.
That's what I'm thinking. In Economics terms, this weird situation is called "bilateral monopoly" where both the buyer and seller have a monopoly of sorts since they're like the only unicorns in town.
But TSMC can delay 3nm until it has the clients. Enough of the "smaller" players will eventually want and pay for it. Apple is far more dependent on TSMC than TSMC is on Apple. Where else is Apple going to go to realistically? It would likely take 3-5 years to fund another competitor to bring it up to what is needed.
> Enough of the "smaller" players will eventually want and pay for it.
At the right price, everybody wants it, but TSMC wants a high price now, to recoup their investment on the new process. If they drop their price now, that will take a long time, and may even be never, as, over time, competition will drive down margins.
Apple may be fairly unique in having a “right price” that is a lot higher than that of other players while they also are willing to order hundreds of millions units.
> But TSMC can delay 3nm until it has the clients.
My impression is that they invested huge sums of money into the 3nm lines; you don't want to delay getting returns on such investments.
Considering that capital has a cost, it might be cheaper for them to forego the 6% increase (at least for now) and start returning some of their loans, rather than letting the interest accumulate. (Which Apple knows, of course, which is why they play hardball).
From what I hear apple's latest chips are already so far ahead of amd64 in performance and power consumption that they can easily wait a couple of years. I do not think that this is what they actually want to do, i am only saying that they can even afford that for a few years if they have to.
I am not sure that there is enough demand from smaller clients to 100% cover apple's usual capacity.
Apple's chips are better largely because they're the only one on TSMC 5. If apple went to say samsung they'd end up on a worse 5nm process while AMD takes up the 3nm capacity Apple just ditched.
The vast majority of Apple chips are phones, not Macs. Samsung is neck & neck and their top end phones are definitely competitive, performing better in many ways for many people.
Apple is about long-term. Short-term, yes, accepting the hike is the cheapest. However, it opens you up for future price hikes. Yeah, they can't just switch to Samsung or acquire and uplift GloFo, join venture with Intel also won't work. Long-term, though, they can start working on this.
That would mean having sub-optimal chips for 3-5 years in their products while Qualcomm/AMD could finally catch up.
TSMC having undisputedly best litography is advantageous to Apple. It means Apple can buy the whole production and nobody else will be able to compete with their chips. They will lose this if they will make the field more competitive by propping up a competitor.
Sure, but good luck with that endeavor. Samsung and Intel are struggling to keep up with TSMC despite decades of expertise in the field and very deep pockets, so starting from scratch even with infinite money isn't gonna be a picnic. And won't solve Apple problem in the short term.
If there is one thing modern Apple is famous for, it is to dislike any situation where they depend on one single external vendor for anything. Apple wants either a somewhat healthy competition (e.g. with batteries, screens or assembly where misbehaving suppliers can be replaced rapidly) or vertical integration - and their acquisition of Intel's mobile modem business shows they think in long terms anyway.
In any case, Apple investing into their own fab also makes sense from a geopolitical point. Apple is already beginning to diversify from China because of their shoddy covid policies causing delays all the time, and everyone can see the writing on the wall that says China is very high on the sanctions priority list of the West. Additionally, the threat of China invading Taiwan and TSMCs fabs getting blown up in a scorched-earth action or damaged by war is only growing bigger every day.
what he's trying to say is that the best companies in the world at chipmaking aren't keeping up with TSMC. Not because they don't have money to spend, but because the problem is really hard and there is a lot of room for 'not complete' failure.
You could be completely successful in making chips at small node sizes and have someone out-compete you, meaning your chips are less performant than the competition and you look worse, and it cost you tens of billions of dollars to get there. That's what samsung and intel have been experiencing for several years now.
One thing apple isn't famous for is manufacturing cpus. It's hard. They buy parts from other companies and pay yet another company to assemble them. There is no one to buy or fund to instantly catch up, and even if you did, next year you may be out again.
It's not an easy problem where throwing money means you win, and apple has the most money.
They struggle because Apple paid upfront for capacity that didn't exist. That allowed TSMC to buildup capacity and secure lithography machines from ASML.
Intel struggled because of piss-poor decisions by execs:
- Reduce stake in ASML
- Not use EUV for 10nm
- 14nm++++++++++
- That was fine (for some time) because Intel had no competition from AMD
- Thinking that 4 cores 8 thread is what consumer needs (if it weren't for AMD we'd still top of the line i7, or how they call it today i9, would still have 4 cores and 8 threads)
- Not seeing that new node is getting more and more expensive in terms of R&D and that pure-play fabs have gigantic advantage
When AMD Zen happened, Intel got caught with its pants down.
Samsung "struggling" because they couldn't outbid TSMC for EUV machinery (wonder where TSMC got money from...)
Capacity for cutting-edge nodes is limited purely on how fast EUV machinery can be delivered. Currently, it's limited by ASML production capacity.
I would suspect the strong dollar is impacting these negotiations. TSMC is probably trying to maintain margins and Apple is bulking at the larger than normal price increase.
On one hand Apple doesn’t really have any other partner to go to that can deliver on 3nm or anything equivalent. On the other hand, at this point, does it even matter? Sure, it would be nice, but is it worth the cost?
TSMC also needs to look at it and say Apple accounts for 25% of our business and the most lucrative and prestigious 25% at that. Is it worth potentially harming the relationship?
The article outlined at least one option that Apple has: stay on the 5nm node for a bit longer. It doesn't seem impossible at all that TSMC would cave/agree to only a 2-3% hike in 6 months. While a greater lead is I'm sure welcome, it's not like Apple isn't _extremely_ well positioned in every market they operate in, even on 5nm.
“N3 technology will offer up to 70% logic density gain, up to 15% speed improvement at the same power and up to 30% power reduction at the same speed as compared with N5 technology”
Kinda crazy to not pay 6% for these gains but that’s just me!
Only if your customers are willing to pay that much.
While Apple chips are highly sought after in the West, that's still a minor part of the worldwide market altogether (and while I'd be happy to go back to M1 Max, I couldn't care less for their glossy displays, shallow keyboards and sharp edges — not to mention incomplete Linux support).
... so? The customers they do have pay ridiculously high prices. Even a 6% rise of the full unit price would barely register, especially considering the increase in performance, which would probably easily justify 20%.
Same here. My Thinkpad doesn't look as cool but it's so much better to work on. Amazing keyboard, matte screen that even has touch, no sharp edges. And it has PORTS!
Apple solved these issues somewhat with the new MBP but the keyboard and screen still sucks for me. Their new SoCs are great though.
I dont think TSMC was saying N3 costs 6% more than N5, they were saying N3 costs 6% more than N3 previously did. Presumably it was significantly more expensive than N5 even before the price hike.
Given that Apple is very competitive even in 5nm I'd say that, yes, yes, they can stay at that process for as long as they want (which is probably +2yrs let's say)
So who would you be breaking up in this case and how do you “enforce anti-trust” laws on a Taiwanese company?
How would you break up Apple in this case? They can’t sell iPhones, iPads, Macs, Watches, monitors (their newest monitor has the same processor as the iPhone 11) and AppleTVs?
Would you break Apple up so each line of iPhones have to be a separate company?
Apple has more to lose here as they've bet their future on Apple Silicons currently only manufactured at TSMC. So it is interesting to see them still trying to arm-twist TSMC - would love to see their playbook as this is something they are known to do with all their suppliers.
My thinking is that Apple is also the only customer that could fund the development of a competitor. I doubt that Apple would want to run their own fab, but they could help fund the development of 3nm at either GlobalFoundries or Intel.
AFAIK TSMC is operating Apple’s machines. If they want, and the contract allows, they can ship the machines somewhere else, and operate them there themselves or with another partner.
It’s not like TSMC builds the machines. They just operate them, and we all know what happens to the man in the middle.
>It’s not like TSMC builds the machines. They just operate them
you clearly have no idea on chip manufacturing.
if foundry just operate machine then why GlobalFoundries licenses Samsung 14nm? it just operating machines, right? just operating machines, why you need to licenses the process? isn't ASML or other machine vendors cover everything you need? just operating machines required R&D on operating machines?
Nah they are TSMC's machines. Really expensive ones. Apple just has bought the production of certain fab before they are built for certain amount of wavers (or time?) for X amount of money. Now TSMC is saying that due to inflation that X needs to be made bigger and thus renegotiating the contracts.
(Also this paying upfront is really good for TSMC as they basically get interest free loan for building the fab)
Also I think there is misconception that you could just buy a EUV lithography machines from ASML and start making chips. That is just one part (a very expensive hundreds of millions costing part) of a multistep process to make a modern microprocessor. In general it takes 3 to 4 months to go from a bare silicon wafer to a packaged chip. This is also why losing power/contamination can be really costly at a semiconductor fab as you can lose 1/4 to 1/3 of a years total production capacity even if you get the fab back to running next week.
6% rate hike is pretty high especially as I've read recently a lot of the inflation costs have actually gone back into corporate profits instead of inflated products costs / labor costs. Add in high cost of USD on global markets.
Interesting scenario with both parties squaring off.
Is there a special name for situations where you have only one or very buyers AND sellers at the same time? Like a combination of monopoly and monopsony power? I wonder how prices are discovered in situations like these?
> I wonder how prices are discovered in situations like these?
You are reading about it! If a deal is met, the supply and demand curves presumably intersect, and a price is discovered.
Of course, with extremely low volume of deals and extremely high complexity, the risk of miscalculation and loss for either party is high. Apple overpaying and their customers not willing to pay, or TSMC underpricing and not being able to cover their costs of production to deliver the goods qt the agreed upon specifications. This is also a loss for Apple, since having a functional TSMC is the ideal, not forcing them to go out of business.
Being almost completely uninformed on the details of Apple/TSMC's dealings, I'm curious, would Apple ever position themselves to get into chip fabrication? I'm sure there are big reasons this is unlikely but it seems like if anyone could do it, it's Apple.
IT is billions of dollars to get a chip fab. I know of several companies that have looked into it (our core competency is manufacturing - why not add chips to the list of things we make and get out of these supply issues..), but the cost to get into it is far too high. Though I suspect some of those companies are regretting the decision over the past 2 years, they could have made the investment back just because we can make those chips you need for a price (to others who can't get needed chips at any price)
Honestly, I'm surprised they haven't started yet. If TSMC puts up too much of a fight, they might start internally considering how they would get into chip fabrication. They've already done everything else to great success (and to much surprise), so while the problem seems insurmountable, maybe Apple would be able to do what nobody else has been able to do in decades (aside from the Chinese with significant state investment).
If Apple isn't at least working on how to bring chip manufacturing in-house I'd be incredibly surprised. They have $48 billion in cash on hand, they at least have worked out a plan on how they would do it if needed.
It's risky because if they do it, they have to consistently be as good or better than TSMC, year after year. Makes more sense than trying to build a car though.
They only have to be "good enough" with some wins, Apple was able to survive the mediocre Intel performance and likely could survive some mediocre iPhone chip performance.
But they won't do it unless they need to, or they think they can get a major win, and even then likely though a subsidiary.
what does apple manufacture themselves? They don't even assemble their own products. It would be surprising if apple suddenly got into one of the most risky and expensive industries, not the other way around.
Looks like what this means is not that TSMC or Apple fold or even meet in the middle… the article implies that it’s sort of a stalemate that results in a short term lose-lose: products may get delayed or not be as good as they might otherwise be, investments get pushed out, etc.
Maybe Apple sees this (if we don’t pay, no one else will, so we don’t lose to another, but we don’t get the new fab) and thinks this is the optimal outcome. They are batting down the hatches prepared for a recession (and impacts of a super strong dollar), and cutting spending ahead of an expected demand drop.
Most of the comments seem to be reacting to the title rather than the content of the article.
Production of Apple chips by TSMC is not at stack. The price hike concerns the new 3nm process which is apparently costlier than usual. There apparently has been ongoing negotiation for a price hike for a bit which was rumoured to be 3% but seem to actually be 6%. Apple said they weren’t interested at this price point which means they probably won’t use 3nm for their next chips unless a different price agreement can be found.
They do but they also buy other chips from competitors sometimes, like they have some phone models that use Qualcomm and even Mediatek SOC/modem chips fabbed at TSMC.
Apple tends to be overly concerned about every half cent in the bill of materials. This is why many of their accessories have substandard (weak) processors and flopped. HomePod used a 4 year old chip which wasn't very responsive to voice commands.
1. Crazy to me to think apple can't absorb 6% price hike on chips. their margins are usually so high.
2. But its awesome that apple can say no to price hikes. This is good economics for end users and its a reason why capitalism and markets are often good and produce lower prices
Interesting that in point #2 you think the market helps lower prices, when in point #1 you establish that Apple doesn't lower the price of what they're selling regardless of how much they lower their own prices.
It seems to me that they did pay a lot to do R&D to develop computers, laptops, software, chip designs, etc.
You do not get rich writing unnecessary checks, but you do get rich writing checks towards investments that result in ROI. Such as paying TSMC to develop more advanced chips in exchange for retaining exclusivity for a certain period.
> 1. Crazy to me to think apple can't absorb 6% price hike on chips. their margins are usually so high.
What makes you think they'd eat that? Would consumers balk at or even notice a 6% increase in price on their next Iphone? Even if they did Apple could just say "Inflation sry"
I mean it isn't just a "we'll lower this" I bet there's lots of juicy confidential future roadmap stuff, things like "Ok we'll help you invest in this _____ tech and machinery but you have to knock off half a cent" or who knows what!