Apple’s 2025 TSMC Chip Order Could Hit $60B, Surpassing Intel’s Annual Revenue

[fansink]

Consumer electronics giant Apple can order as much as $60 billion in chips from TSMC, believe analysts, and allow the Taiwanese chip manufacturer to earn as much as NT$1 trillion in revenue in 2025. The analysis is based on the assumption that TSMC's share in Apple's products is growing, and the firm's fabrication plants in Arizona are also expected to play a key role. Apple CEO Tim Cook has repeatedly stressed that his firm is TSMC Arizona's largest customer, and TSMC's leading-edge 2-nanometer chip fabrication process is expected to play a key role in increasing the value of Apple's orders to TSMC.

Apple's Orders To TSMC Will Increase Due To Newer Manufacturing Technologies, Believe Analysts

TSMC's first quarter of 2025 revenue marked a sizable growth in the firm's 3-nanometer revenue over the year-ago figures. During the period, 22% of the firm's revenue came from 3-nanometer technologies, accounting for a 13 percentage point jump from Q1 2024's 9%. At the same time, revenue contribution from 5-nanometer process technologies dropped by one percentage point, while revenue contribution from 7-nanometer process technologies dropped by four percentage points.

The jump in 3-nanometer orders is primarily due to Apple's orders as the firm often uses TSMC's latest chip manufacturing technologies for processors used in its smartphones. Now, analysts quoted by Taiwan's Economic Daily believe that 2-nanometer orders from Apple could touch NT$1 trillion in 2025 or roughly equivalent to $33 billion.

TSMC earned $90 billion in revenue in 2024, and the sources quoted by the publication add that the revenue growth will also stem from TSMC increasing its presence across all of Apple's products.

If the orders materialize, Apple's share of TSMC's revenue could grow significantly this year. The analysts believe the firm bought NT$624 billion of chips from TSMC in 2024. TSMC does not disclose customer revenue details, but Apple is long thought to be the firm's largest customer. A growth in orders to NT$1 trillion in 2025 would be 60% in percentage terms.

The NT$1 trillion figure is the high end of the range, as analysts expect the 2025 figure to range between NT$800 billion and NT$1 trillion. The midpoint would mark 44% annual growth, while the low end would mark 28%.

They believe that whether a high end is achievable depends on the speed at which TSMC's Arizona plants picks up production and the production ramp up of the 2-nanometer process technology. TSMC's 2-nanometer manufacturing process is the most advanced in the world. It uses newer nanosheet transistors over the previous technologies' FinFET designs.

TSMC is competing with Samsung and Intel regarding the 2-nanometer technology. Intel also aims to start producing its 18A manufacturing process in 2025. Under the leadership of new CEO Lip-Bu Tan, Intel is seeking a share of the contract chip manufacturing industry, where TSMC is the dominant player and holds the highest market share.

Apple Could Order A Whopping $60 Billion In Chips From TSMC In 2025 Believe Analysts

Apple could order as much as $60 billion in chips from TSMC this year beleive analysts to account for double digit percentage growth.

wccftech.com

 

[Daniel Nenni]

Apple has been 25% of TSMC's revenue for a long time. Apple is taking a beating in China so I do not see how Apple could be ordering big this year. QCOM and MediaTek make up a big part of the smartphone business as well. I think the +20% Others is interesting. +7% for HPC seems low. Automotive is surprising. 2Q25 will be interesting, absolutely.

 

[fansink]

Apple has been 25% of TSMC's revenue for a long time. Apple is taking a beating in China so I do not see how Apple could be ordering big this year. QCOM and MediaTek make up a big part of the smartphone business as well. I think the +20% Others is interesting. +7% for HPC seems low. Automotive is surprising. 2Q25 will be interesting, absolutely.

Apples sudden increase from it's traditional 25% to 45% may be Ai related

TSMC is on track (so far) to beat last years YOY revenue increase; FY25 could get very close to NT$4T (~US$130B)

 

[siliconbruh999]

TSMC is projected at 125-130 Billion in revenue. Apple increases to let say 30% it would be $39 billion still

 

[fansink]

TSMC is projected at 125-130 Billion in revenue. Apple increases to let say 30% it would be $39 billion still

My earlier projections don't include Super Hot Runs (SHR), which are very hard to forecast.

SHR orders typically incur a 30%-40% premium over standard orders.

Between Apple, Nvidia, Intel, and AMD, all seemingly wanting to "strike while the iron is hot", who knows how many SHR's there'll be in 2025.

TSMC’s leading-edge foundries may be running at near full SHR capacity for the rest of 2025.

 

[siliconbruh999]

My earlier projections don't include Super Hot Runs (SHR), which are very hard to forecast.

SHR orders typically incur a 30%-40% premium over standard orders.

Between Apple, Nvidia, Intel, and AMD, all seemingly wanting to "strike while the iron is hot", who knows how many SHR's there'll be in 2025.

TSMC’s leading-edge foundries may be running at near full SHR capacity for the rest of 2025.

Hot lots are impossible to forecast especially with TSMC. They have the power they can charge double.

It's WccFtech news article so it automatically makes it unreliable.

 

[fansink]

Hot lots are impossible to forecast especially with TSMC. They have the power they can charge double.

It's WccFtech news article so it automatically makes it unreliable.

After Intel's CFO David Zinsner’s admission of limited (~zero) 18A external customers, and knowing that companies like Apple knew that 18A was dead a long time ago (it's their business to know), makes it clear that Apple would use this opportunity to bury their competitor as deep as reasonably possible

You'll see, everyone who stands to gain from this mishap will do everything they can to exploit Intel's misstep, to their advantage, TSMC, Apple, Nvidia, AMD, and Qualcomm, are all digging burial holes

 

[siliconbruh999]

I don't doubt about it that everyone is looking to bury Intel but you are underestimating Intel AMD was in Worse position and it came back.

 

[fansink]

I don't about it that everyone is looking to bury Intel but you are underestimating Intel AMD was in Worse position and it came back.

AMD’s comeback was on TSMC’s reliable nodes

Intel lacks that partner, facing steeper yield and market share challenges

Until they fully join the dark side they’re vulnerable

 

[siliconbruh999]

AMD’s comeback was on TSMC’s reliable nodes

You are making it sound like their design team didn't do anything

Intel lacks that partner, facing steeper yield and market share challenges

There is not that big of a yield challenge but market share challenges are getting bigger each day.

Until they fully join the dark side they’re vulnerable

LoL I would rather have them otherwise TSMC would jack the prices to Moon.
Let's not forget TSMC is not infallible

 

[count]

AMD needed a fundamental business model change to come back. AMD would not have come back if they remained an IDM.

I think Intel could come back if they change their business model.

 

[Xebec]

Apple has been 25% of TSMC's revenue for a long time. Apple is taking a beating in China so I do not see how Apple could be ordering big this year. QCOM and MediaTek make up a big part of the smartphone business as well. I think the +20% Others is interesting. +7% for HPC seems low. Automotive is surprising. 2Q25 will be interesting, absolutely.

Re: Automotive, FWIW Tesla is using N2 for Full Self Driving / Robotaxi chips (AI5), and is selling millions of cars using the chips on N4 (HW4/AI4). Tesla sold ~ 1.8M vehicles last year, so 2.5M (including Taxi) seems right for 2026.

I think we'll see more legacy auto start to use advanced tech too -- VW is partnering with Rivian that is probably buying more advanced chips than other legacy automakers. I think Ford is also pushing to do more integration with vehicle systems than before - maybe not on the advanced nodes, but probably moving 'up' to better legacy nodes than before = new business.

 

[Daniel Nenni]

Re: Automotive, FWIW Tesla is using N2 for Full Self Driving / Robotaxi chips (AI5), and is selling millions of cars using the chips on N4 (HW4/AI4). Tesla sold ~ 1.8M vehicles last year, so 2.5M (including Taxi) seems right for 2026.

View attachment 3170

I think we'll see more legacy auto start to use advanced tech too -- VW is partnering with Rivian that is probably buying more advanced chips than other legacy automakers. I think Ford is also pushing to do more integration with vehicle systems than before - maybe not on the advanced nodes, but probably moving 'up' to better legacy nodes than before = new business.

I remember when I first saw Tesla on SemiWiki 10+ years ago. I then noticed on LinkedIn that some of my AMD contacts worked there including Jim Keller. Now all of the car companies are on SemiWiki and building design teams so they can chase Tesla.

Back then Automotive was maybe 4-5% of TSMC's mature node revenue. The pandemic really motivated automotive to make their own chips but AI will push that even further with leading edge silicon. Exciting times!

 

[Daniel Nenni]

AMD needed a fundamental business model change to come back. AMD would not have come back if they remained an IDM.
I think Intel could come back if they change their business model.

AMD was fabless and floundering with Globalfoundries. AMD did not see real progress against Intel until they did two things: Started working with TSMC on an exclusive basis granting them access to the TSMC inner circle and acquiring Xilinx which had an excellent foundry team.

So it wasn't just a business model change, it was a partner strategy with TSMC and building a better foundry team Xilinx. And yes the design team did an excellent job but without TSMC that would not have been possible, my opinion.

 

[Daniel Nenni]

I don't about it that everyone is looking to bury Intel but you are underestimating Intel AMD was in Worse position and it came back.

I agree. Also, if it was a CEO vs CEO competition Intel would win today. Lisa Su destroyed Pat Gelsinger but Lip-Bu Tan is a level above Lisa. Watch him and learn, a master's class in pivoting a company. Lisa has been AMD CEO for 10+ years, Lip-Bu will do it in less than 5 years, absolutely.

Go ahead, bet against Lip-Bu, I double dog dare you.

 

[Xebec]

AMD was fabless and floundering with Globalfoundries. AMD did not see real progress against Intel until they did two things: Started working with TSMC on an exclusive basis granting them access to the TSMC inner circle and acquiring Xilinx which had an excellent foundry team.

So it wasn't just a business model change, it was a partner strategy with TSMC and building a better foundry team Xilinx. And yes the design team did an excellent job but without TSMC that would not have been possible, my opinion.

I am curious if there is any more backstory on this as to the timing of AMD divesting it's fabs.

Going fabless feels like something they should have done earlier (i.e. writing on the wall for their business shrinking vs. cost of new fabs.. + they were already working with TSMC for Radeon). Was going fabless a last minute reaction to "our balance sheets are in pain", or did they try to shop their fabs out for a while and then ended up with GF as the partner?

 

[Xebec]

I agree. Also, if it was a CEO vs CEO competition Intel would win today. Lisa Su destroyed Pat Gelsinger but Lip-Bu Tan is a level above Lisa. Watch him and learn, a master's class in pivoting a company. Lisa has been AMD CEO for 10+ years, Lip-Bu will do it in less than 5 years, absolutely.

Go ahead, bet against Lip-Bu, I double dog dare you.

What's the over/under on Lip-Bu vs. Jensen Huang?

 

[tedc]

I am curious if there is any more backstory on this as to the timing of AMD divesting it's fabs.

Going fabless feels like something they should have done earlier (i.e. writing on the wall for their business shrinking vs. cost of new fabs.. + they were already working with TSMC for Radeon). Was going fabless a last minute reaction to "our balance sheets are in pain", or did they try to shop their fabs out for a while and then ended up with GF as the partner?

Global Foundries was AMD's fabs and manufacturing before they were spun out in 2009. As part of the divestiture, AMD had made commitments to use GF fabs and arguably suffered as GF failed to keep up with the leading edge.

Once GF stopped developing their 7 nm processes, AMD was able to modify their purchase agreements to (more) freely choose TSMC for leading edge processes.

 

[Daniel Nenni]

What's the over/under on Lip-Bu vs. Jensen Huang?

As a CEO for Intel? I pick Lip-Bu even though I am a big Jensen fan. I was rooting for Intel to buy Nvidia before BK came in so Jensen could be CEO. That would have been something. Today though I say Lip-Bu for the Titanic sized turnaround.