---
type: "Topics"
locale: "en"
url: "https://longbridge.com/en/dolphin/post/44248232.md"
description: "$NVIDIA(NVDA.US) can raise memory capacity via two parallel paths: push process limits to make each stack denser, or avoid chasing per-stack density and win on unit count. While the industry races to stack HBM higher, Nvidia is taking the opposite route by cutting layer counts.First, the stacking route. HBM co-packages with the GPU under a shared lid, so total Z-height is capped (HBM4 relaxes it to 775 µm). To squeeze in more layers, dies must be thinned further and micro-bumps shrunk, or switched to Cu-Cu bonding, which raises process complexity, lowers yield, and complicates thermals.The upside is that each added layer benefits from first-rank bandwidth.Nvidia, however, is moving the other way. For the next-gen Rubin Ultra, HBM was initially specced at 16 layers, then cut to 12, and is now rumored to drop to 8.That would put per-stack capacity at roughly 192GB—below the current Rubin's 288GB—an unusual 'step back' for a flagship. These are not officially confirmed and only reflect options being weighed.Why cut layers? Capacity is the core constraint: the most advanced HBM4E is more complex, and suppliers cannot ramp to Nvidia's desired volume and cadence in the near term. At equal capacity, HBM already consumes ~3–4x the wafers of commodity DRAM, and higher stacks burn even more materials.More importantly, total system capacity can still go up via scale-up, linking more GPUs together. A B300 rack is around 20TB, and while Rubin Ultra's per-stack is smaller, interconnecting 576 GPUs can push system memory to ~110TB. In essence, cutting layers and shipping more stacks trades process difficulty and cost for volume under tight supply.This approach creates two new bottlenecks: first, the constraint shifts from memory makers' expansions to base-die and co-packaging capacity, i.e., how much capacity it can secure from TSMC. Second, it raises the bar for GPU interconnect and memory pooling.So whether to increase or decrease HBM layers hinges on two factors: whether bandwidth can keep up—stacking higher only helps if bandwidth scales—and whether packaging and base-die capacity is ample—if sufficient, cutting layers and using more stacks is viable."
datetime: "2026-10-09T02:02:01.000Z"
locales:
  - [en](https://longbridge.com/en/dolphin/post/44248232.md)
  - [zh-CN](https://longbridge.com/zh-CN/dolphin/post/44248232.md)
  - [zh-HK](https://longbridge.com/zh-HK/dolphin/post/44248232.md)
author: "[Dolphin Research](https://longbridge.com/en/dolphin.md)"
generator: "portal-rs"
---

# $NVIDIA(NVDA.US) can raise memory capacity via two…


### Related Stocks

- [TSM.US](https://longbridge.com/en/quote/TSM.US.md)
- [NVDA.US](https://longbridge.com/en/quote/NVDA.US.md)
- [TSMX.US](https://longbridge.com/en/quote/TSMX.US.md)
- [TSMY.US](https://longbridge.com/en/quote/TSMY.US.md)
- [TSMZ.US](https://longbridge.com/en/quote/TSMZ.US.md)
- [TSMG.US](https://longbridge.com/en/quote/TSMG.US.md)
- [TSMU.US](https://longbridge.com/en/quote/TSMU.US.md)
- [STSM.US](https://longbridge.com/en/quote/STSM.US.md)
- [NVDD.US](https://longbridge.com/en/quote/NVDD.US.md)
- [NVDL.US](https://longbridge.com/en/quote/NVDL.US.md)

---
> **Disclaimer: This article is for reference only and does not constitute any investment advice.**