Supply Chain

From compute bottlenecks to industrial consequences — where value may actually concentrate through 2030 Series: AI Compute Supply Chain | Part 5 of 5 Author:* Po-Sung(Sinclair) Huang |For the past four articles in this series, I have written about CoWoS, HBM, ABF substrates, SEC filings, and the fault lines that could eventually crack today’s moats. On the surface, that may look like a semiconductor series. It is not. What these articles really reveal is something larger: AI is no longer just a software story, and no longer just a model race. It is becoming an industrial system — one that depends on power, cooling, capital expenditure, advanced packaging, memory bandwidth, substrate materials, qualification cycles, and the physical discipline of manufacturing scale. ...

TurboQuant, HBM demand reversal, geopolitics, and glass substrates — not a doomsday scenario, but a disciplined analysis. Series: AI Compute Supply Chain | Part 4 of 5 Author: Sinclair Huang Four pressure vectors. Four clocks are already ticking. The question is not whether these moats will last forever — it’s whether you know which one breaks first. I developed a habit during my years in the electronics industry that I haven’t been able to shake in research and writing: every time I become confident about something, I force myself to find the strongest argument against it. ...

Customer prepayments, HBM margin structure, capital expenditure intensity — the numbers say more than the narratives do.Series: AI Compute Supply Chain | Part 3 of 5 By Po-Sung(Sinclair) Huang The previous article built a framework. This one tests it with documents. Before writing this piece, I set aside the analyst summaries and went to the source: TSMC’s 2024 20-F filed with the SEC on April 17, 2025; Micron’s most recent 10-K and two 10-Q filings; and SK Hynix earnings call transcripts for the past three quarters. ...

From TSMC to SK Hynix to Ajinomoto — who holds pricing power, and who is just riding the narrative? Series: AI Compute Supply Chain | Part 2 of 5 Author: Po-Sung(Sinclair) Huang The first article in this series explained what CoWoS, HBM, and ABF are. This one answers the harder question underneath: who controls them? Knowing that these three layers matter is the entry ticket. The real analytical edge comes from understanding who in each layer has the power to say no — and what happens to the whole chain when they do. ...

Why is everyone suddenly talking about CoWoS, HBM, and ABF whenever AI, NVIDIA, or AI servers come up? Many people know they are important, but still get stuck the first time they run into these terms. This essay is a plain‑language walkthrough of what they actually are, why they are always mentioned together, and how they map onto Taiwan’s role in the global AI supply chain. When Google, Amazon, and Tesla are all designing their own chips, is Taiwan’s manufacturing ecosystem still structurally important — or just enjoying a temporary window? ...

Every layer that looks solved hides another constraint beneath it. §1 The Illusion of Infinite Compute The headlines say NVIDIA is winning. The hyperscalers are spending. The models are getting bigger. But the real question is not where demand is going. It is where compute, physically, can still be built fast enough to meet it. The harder answer requires tracing the full physical stack — from silicon wafer, through memory stack, through packaging interposer, through substrate material — and asking at each layer: can this actually scale at the speed the demand curve requires? ...