Sunday, October 16th 2022
TSMC Cuts Back CAPEX Budget Despite Record Profits
Another quarter, another record breaking earnings report by TSMC, but it seems like the company has released that things are set to slow down sooner than initially expected and the company is hitting the brakes on some of its expansion projects. The company saw a 79.7 percent increase in profits compared to last year, with a profit of US$8.8 billion and a revenue of somewhere between US$19.9 to US$ 20.7 billion for the third quarter, which is a 47.9 percent bump compared to last year. TSMC's 5 nm nodes were the source for 28 percent of the revenues, followed by 26 percent for 7 nm nodes, 12 percent for 16 nm and 10 percent for 28 nm, with remaining nodes at 40 nm and larger making up for the remainder of the revenue. By platform, smartphone chips made up 41 percent, followed by High Performance Computing at 39 percent, IoT at 10 percent and automotive at five percent.
TSMC said it will cut back its CAPEX budget by around US$4 billion, to US$36 billion, compared to the earlier stated US$40 billion budget the company had set aside for expanding its fabs. Part of the reason for this is that TSMC is already seeing weaker demand for products manufactured using its N7 and N6 nodes, as the N7 node was meant to be a key part of the new fab in Kaohsiung in southern Taiwan. TSMC is expecting to start production on its first N3 node later this quarter and is expecting the capacity to be fully utilised for all of 2023. Supply is said to be exceeding demand, which TSMC said is partially to blame on tooling delivery issues. TSMC is expecting next year's revenue for its N3 node to be higher than its N5 node in 2020, although the revenue is said to be in the single digit percentage range. The N3E node is said to start production sometime in the second half of next year, or about a quarter earlier than expected. The N2 node isn't due to start production until 2025, but TSMC is already having very high customer engagement, so it doesn't look like TSMC is likely to suffer from a lack of business in the foreseeable future, as long as the company keeps delivering new nodes as planned.
Sources:
TSMC, via @dnystedt
TSMC said it will cut back its CAPEX budget by around US$4 billion, to US$36 billion, compared to the earlier stated US$40 billion budget the company had set aside for expanding its fabs. Part of the reason for this is that TSMC is already seeing weaker demand for products manufactured using its N7 and N6 nodes, as the N7 node was meant to be a key part of the new fab in Kaohsiung in southern Taiwan. TSMC is expecting to start production on its first N3 node later this quarter and is expecting the capacity to be fully utilised for all of 2023. Supply is said to be exceeding demand, which TSMC said is partially to blame on tooling delivery issues. TSMC is expecting next year's revenue for its N3 node to be higher than its N5 node in 2020, although the revenue is said to be in the single digit percentage range. The N3E node is said to start production sometime in the second half of next year, or about a quarter earlier than expected. The N2 node isn't due to start production until 2025, but TSMC is already having very high customer engagement, so it doesn't look like TSMC is likely to suffer from a lack of business in the foreseeable future, as long as the company keeps delivering new nodes as planned.
40 Comments on TSMC Cuts Back CAPEX Budget Despite Record Profits
The cost of designing a chip is rising. The cost of building a fab is rising sharply (see Moore's Law, not the First in this case but the Second). Not just chip designers, even governments won't be able to afford multibillion subsidies every year to each of the big three. I won't be surprised of one of the three (Samsung?) abandons leading edge development in a few years. Most foundries that existed a decade or two ago have already done that, and it happened when fabbing was far less capital intensive.Hm, in many cases we consumers are still paying for these chips. It's monthly payments though, not $500 at once.
Intel hit the wall a while back now tsmc.
Get ready to use 2-3kw PSU or maybe they might start bundling with a generator.
In the past when you go to a smaller node, the higher number of chips per wafer gives a cost per wafer manufacturing benefit that is enough to cancel out inflation / complexity of development cost per chip.
N5 was where cost per die went up slightly vs N7. i.e. that cost to manufacture went up slightly, whereas in the past it would go down. So there's no longer a cost benefit to going to a new smaller node.
Like everything, this is going to bifurcate the market. The 'haves' will buy expensive devices with N3 / 3nm or smaller class chips in them, while cheaper devices will have N5 class chips in them.
Apple is already bifurcating their lineup with iPhone 14 using the N5 node A15 SoC, while the iPhone Pro uses the N4 (still 5nm class) node A16. We also see it with Zen 4 pricing.
We're going to see that again next year, only the high end iPhones will have 3nm class N3 node A17. Everything else will have the 5nm class N4 node A16. I think it will stay like that for a long time.
Just as an example, 31% of Nvidia's income is from the sale of gaming video cards. An even small fraction of that 31% comprises customers that have the choice to reduce purchase frequency. That is because a portion of GeForce cards are purchased by prosumers / professionals for various work and those individuals do not have the option to reduce purchase frequency in many cases.
If recent / upcoming product launches are any indication, customers buying "budget" options are paying prior high-end prices.
For a very long time we get to enjoy higher performance for same money every year, but the recent slowdown in tech advancement from the foundries is putting an end to that. We used to get a few foundries who are able to compete at the cutting edge, now TSMC is the only one at the cutting edge. No one comes close so they get to dictate the price of the latest and greatest node. Anyone who wants the price friendly option will need to move to the previous nodes (7nm or 10nm) where competition is still alive.
On the other hand, manufacturing costs for any given node probably go down over time, we just don't know how fast and by how much.
But that said, not only is development cost on smaller nodes near exponential, the actual 300mm wafer cost is going up so much that the per-chip cost is now going up as well.
Look at row 8 specifically, the per chip cost went from $2433 at 90nm to 233 at 7nm, and now is reversing with 5nm. This is just the cost to manufacture, development cost and capital to build the smaller nodes is ramping up as well. This is why I'm saying, we hit the inflection point with 5/7nm nodes.
It will cost more per chip to make from now on. That cost will be passed to the consumer.