# HBM demand — X 热门讨论 (2026-09-23 16:40 UTC)

## @Abbycadabby87 (Abby) · 09-23 15:29 · ♥62 ↻4 💬1 AI has just surfaced another materials theme worth watching: CCL + high-end copper foil.

Morgan Stanley expects the global CCL market to grow from around $19 billion to $47 billion by 2030. That number is even bigger than I originally expected.

The key isn’t the market size itself, but that each new generation of AI racks is using more materials, and more expensive materials. That keeps pushing value further upstream. I’m paying more attention to HVLP4 copper foil now, because the logic is very similar to what we previously saw with HBM and Lasers. Demand is accelerating, but high-end supply isn’t that easy to catch up. Qualification is slow, yields are difficult to ramp, and nameplate capacity doesn’t necessarily equal truly usable capacity.

For now, I’m watching two companies on this theme: 1 TUC $EMC, high-end CCL 2 Co-Tech, HVLP4 copper foil

I’m not in a rush to treat this as a long-term core theme yet. The key is whether the material value per rack can keep increasing on next-generation AI platforms. If that continues to play out, then this theme could have a chance to move from a short-term trade into a longer-term trend. $TSMC #AIInfrastructure #CCL https://x.com/Abbycadabby87/status/2102782390280032463

## @Abbycadabby87 (Abby) · 09-23 14:13 · ♥50 ↻3 💬3 The last time $MU stayed around 1000 for a long time, a lot of people thought it couldn’t move any higher.

Now looking back, the stock has firmly held above 1000.

In my view, HBM is still consuming DRAM wafers, while Server DRAM capacity keeps moving higher. That’s one of the strongest supports for the entire industry.

As long as upstream supply can’t keep up, pricing and margins will still have support.

Right now, $LITE looks very similar to MU back then. It’s sitting in the 900s, and my view is very similar.

Optical communication is now moving from between racks, further toward the Board, and even closer to the Chip.

The number of demand scenarios keeps expanding. Growth just needs some time.

The biggest thing Memory and Optics have in common this year is that demand is expanding. Next, I’ll just be watching how strongly institutional capital continues to support the move.

$COHR $HBM $CPO https://x.com/Abbycadabby87/status/2102763392054288780

## @SemiconductorsX (Semiconductor Insider) · 09-23 11:01 · ♥35 ↻2 💬3 Hyperscaler capex is exploding. Memory is taking an even bigger slice.

UBS now puts memory at 14% of AI capex in 2025, 37% in 2026, and 64% by 2027. That is $71 billion, then $367 billion, then $923 billion, while total AI spend heads toward $1.45 trillion.

This is not just more servers. Next-gen GPUs and TPUs need stacked HBM that eats about 3x the wafer area of regular DRAM. Capacity is being pulled from DDR and NAND at the same time inference, agent workloads, and storage demand rise.

Result: shortages and higher ASPs through 2027. Hyperscalers already capture most of the relevant HBM, DDR, and NAND supply. Memory has gone from a line item to the line item.

The boom is real. The mix is shifting even faster. Watch the memory names as closely as the GPU names.

If memory is 64% of 2027 AI capex, what part of that $923 billion do you think is HBM versus DDR and NAND? $MU $SKHY $NVDA $GOOGL https://t.co/G0zE3FnjT2 https://x.com/SemiconductorsX/status/2102714874120081698