TSMC's N3 logic wafer capacity has become one of the AI industry's biggest constraints, which could push customers to explore greater foundry diversification
AI demand had already exposed capacity pressure beyond wafer fabrication: TSMC described advanced-packaging capacity as very tight in 2023. This report identifies N3 logic wafers as another binding point in the same infrastructure buildout.
AI-chip customers dependent on TSMC N3 face a direct production-capacity constraint, limiting how quickly they can translate chip designs into available datacenter hardware.
The reported shortage gives those customers a stronger incentive to evaluate additional foundry options rather than concentrate future output at TSMC.
Second-order effects
Foundry diversification becomes a more concrete procurement and design priority: prospective alternatives can compete for AI-related programs, while TSMC customers must weigh supply resilience against the cost and complexity of shifting production.
The wafer bottleneck can compound other manufacturing constraints; later coverage of rapid CoWoS capacity growth and Nvidia's large reservations shows that added supply at one step does not automatically remove limits elsewhere in the AI hardware chain.
Third-order effects
If leading-edge demand continues to outrun available capacity, AI infrastructure deployment will be governed as much by manufacturing allocation as by chip-design roadmaps.
The industry could move toward more deliberate multi-foundry sourcing, though its practical extent will depend on whether alternatives can meet customers' required process and volume needs.
The trend: AI infrastructure is increasingly constrained by interconnected leading-edge manufacturing stages, making supply-chain resilience a strategic input to compute expansion.
If you've been wanting to understand AI inference and hardware economics... this is the best single place for a current take. It's long, but highly recommended for those interested!
Every AI bubble argument assumes the compute requirements keep going up forever. The actual trend is going the other way. GPT-4 required cutting-edge H100s to run at scale. Newer models at the same or better quality level run on hardware that is two to three generations older.
There's an economics theorem called Alchian-Allen. And it has the very interesting implication that AI labs will be able to charge *higher* margins on their best models as compute gets scarcer. As compute gets more expensive, the cost of running any model goes up. So you might [v…
.@dylan522p forecasts that iPhones could get $250 more expensive for consumers. Smartphone sales could drop from 1.1 billion a year to 500-600 million over the next couple of years, with the low end getting crushed hardest. Around a third of big tech's $600 billion in CapEx this …
Excellent listen. Key takeaways: 1) $ASML caps at 200GW by 2030 2) Memory eating 30% of Big 7 CapEx 3) H100s appreciate vs. depreciate 4) Neoclouds (and their agents such as $GLXY) control bottleneck 5) Early contracts ($CRWV 98% locked) print vs 50% spot markup @dylan522p 🐐
This podcast with @dylan522p is a terrific rebuttal to the Citrini doomer scenario by playing through the real world constraints of a fast-ish takeoff (I know it wasn't intended as such). The constraints to producing enough AI tokens to be disruptive to society will slow it down
This matches my current world model. An H100 GPU is worth more today than 3 years ago, not less. People viewing GPUs as rapidly depreciating assets are missing that an older GPUs can and will do economically valuable work. The total economic value of work that GPU can do for you
The logic squeeze that is happening (Dylan thinks it's going to get worse) ensures that Intel after its N2P orders are filled for Nova Lake will be forced back to internal manufacturing The idea of Intel flexing ~20% to TSMC doesn't work in this world. That's good news
this completely fucking breaks the AI Bubble narrative. a 3 year-old gpu is MORE valuable today because it serves higher-quality ai tokens FOR CHEAPER. translation: gpt 5.4 runs BETTER on an OLD GPU than gpt-fucking-FOUR read that again. a newer, better model runs more
if you're interested in the race to agi you have to watch this. much more in depth on the super cluster build out. how much compute / gw's can you get online quickly. sam (the dealmaker) altman's conviction and acceleration is paying off again. whilst anthropics sbf
The people I find most insightful/interesting on the economics of AI are non-economists. On hardware/macro topics, it is very hard to beat Dwarkesh and Dylan - I'm always enlightened by listening to them and humbled by their breadth and depth of knowledge. Definitely going to
.@dylan522p gives a deep dive on the 3 big bottlenecks to scaling AI compute: logic, memory, and power. And walks through the economics of labs, hyperscalers, foundries, and fab equipment manufacturers. Learned a ton about every single level of the stack. 0:00:00 - Why an H100 [v…
Narrative violation from Dylan on Dwarkesh: H100s are worth *more* today than they were 3 years ago. There's a sentiment that data center buildouts are priced into the risk of rapidly depreciating GPUs. But the models want to learn. Token prices are falling so fast that you can […
The AI supply chain has the craziest value cascade of any industry in the world. thinks that over the next five years, the biggest bottleneck to deploying AI will be EUV machines. ASML sells EUV machines for $300-400 million. You need about three and a half machines, so $1.2 [vid…
.@dylan522p lays out how we know the hard upper bound on how much compute can be produced annually by 2030: around 200 GW/year. That's a crazy number (there's about 20 GW of AI deployed in the world right now), but it's nowhere near enough to satisfy Sam/Elon/Dario/Demis's [video…