The N2 (2 nm) family includes the A16 (1.6 nm) node, TSMC has described it as N2P with backside power delivery (which they are calling Super Power Rail or SPR).The cost to develop the N2 has been the highest in history at nearly $800 million. The wafer is expected to cost an eye watering $30,000 due to its complexity.
For reference, the 1.4 / 1.6 nm node is expected to cost $45,000+ per wafer.
So if your source (whatever it is) for these prices is accurate, then that’s a big price jump for SPR.
A14 (1.4 nm) also appears to have this two-step approach, with the first generation (A14) said to be without SPR and the second generation (A14P) said to be with SPR. I wonder if there’s a cost difference there as well.
By “third generation of the N2 node” do you mean A16? Because you’re right, backside power delivery is a big deal. Or did I miss new information about TSMC’s roadmap?Really doesn’t have much to do with competition, see what I wrote above. The 3rd generation of the N2 node should be better optimized and provide significant better yields and lower costs (plus significant better performance). Because they are becoming so expensive to produce and manufacture, the N2 will likely be with us until 2030 (not without significant advancements) and 1.4 nm / 14A possibly even longer with those platforms costing significantly more.
However, the products those future platforms will allow are the thing of dreams. We will see mind blowing tech by the early 2030s with a bigger technical leap than we have seen in the last 15 years.
My guess is the M7 Pro/Max/Ultra tier will be on A16, while M7 itself will be on N2P. Both are currently expected to begin volume production in 2H 2026.