SemiconductorsFEATURE

Why Does AI Need HBM? How High Bandwidth Memory Breaks the Memory Wall

E
EffectStory 編輯部Editorial Team
Published · Updated
HBM solves AI's core bottleneck—the memory wall—by stacking DRAM on a wide silicon-interposer bus beside the GPU, lifting bandwidth from 256 GB/s (HBM2) to 2 TB/s (HBM4) and capacity from 8GB to 64GB, though supply stays concentrated in three makers led by SK Hynix's 58% share.

The Memory Wall: Why Does AI Computing Need HBM?

The memory wall describes a mismatch in which GPU compute units process data far faster than memory can supply it, and since large language model inference repeatedly reads weights back from memory during each pass, memory bandwidth—not raw compute—becomes the practical limit on inference speedCITE:E1.

HBM's 3D Stacked Architecture: Breaking Traditional Memory Design Limits

High Bandwidth Memory (HBM, 高頻寬記憶體) stacks DRAM dies vertically on top of a silicon interposer wired with an extremely wide data bus, a 3D packaging approach that trades chip footprint for bandwidth far beyond what traditional DDR or GDDR memory can deliverCITE:E2. Because this stacked package sits directly next to the GPU, the physical distance data must travel is shortened, compounding the bandwidth gain already achieved by the wide-bus architectureCITE:E2.

HBM vs. DDR/GDDR: How Bandwidth Differences Reshape AI Accelerators

A single DDR5 DIMM channel used in conventional servers delivers roughly tens of gigabytes per second, and consumer GPUs running GDDR6/7 reach roughly hundreds of gigabytes per second, while a single AI accelerator carrying multiple HBM stacks reaches a combined bandwidth in the terabytes-per-second rangeCITE:E7. This gap is the reason HBM exists as a distinct standard purpose-built for AI and high-performance computing workloads rather than as a general-purpose graphics or server memory formatCITE:E7.

A Decade, Eightfold: HBM's Leap From HBM2 to HBM4 in Performance and Capacity

Per-stack bandwidth rose from 256 GB/s under the 2016 HBM2 standard to 2 TB/s under the 2025 HBM4 standard—a nearly eightfold increase over a decade—while maximum stack capacity climbed in parallel from 8-Hi/8GB to 16-Hi/64GBCITE:E3CITE:E4.

GenerationStandard YearBandwidth per StackMax Stack Height / Capacity
HBM22016256 GB/s8-Hi / 8GB
HBM2E307 GB/s
HBM32022819 GB/s16-Hi / 64GB
HBM3E2024 (mass production)~1.2 TB/s
HBM420252 TB/s16-Hi / 64GB

The combination matters because bandwidth and capacity scaled togetherCITE:E3CITE:E4: a wider pipe alone would not help if an accelerator still could not hold a large model's weights close to the compute die, and the jump from 8GB to 64GB per stack lets a single AI accelerator keep far more of a model's weights in nearby memoryCITE:E4.

The HBM Supply Bottleneck: Manufacturing Constraints and Market Concentration Risk

HBM production is manufacturing-constrained because vertical stacking requires through-silicon via (TSV) technology to connect each DRAM layer, a process with yield and packaging complexity far higher than planar DRAM, and this difficulty is why capacity expansion has not kept pace with the surge in AI demandCITE:E6. Supply is also concentrated: in 2026 Q1, SK Hynix held 58% of the HBM market, down from 69% in the same period a year earlier, while Samsung Electronics and Micron each held roughly 21% and tied for second placeCITE:E5.

Supplier2026 Q1 ShareYear-Earlier Share
SK Hynix58%69%
Samsung Electronics~21%
Micron~21%

All three suppliers' HBM4 products have already been certified by NvidiaCITE:E5.

What This Means

The same reason HBM exists—closing the gap between GPU compute speed and memory bandwidth described by the memory wallCITE:E1—is what makes its stacked, wide-bus design necessary rather than optionalCITE:E2, and the resulting terabyte-per-second scale is precisely what separates it from DDR and GDDR memoryCITE:E7. That same stacking, however, is also what makes HBM hard to produce: the TSV process behind a decade of nearly eightfold bandwidth and capacity gainsCITE:E3CITE:E4 is the identical constraint now limiting how fast three suppliers—one of them losing share—can expand output to match AI demandCITE:E5CITE:E6.

📊 Evidence

FAQ

The Memory Wall: Why Does AI Computing Need HBM?

The memory wall describes a mismatch in which GPU compute units process data far faster than memory can supply it, and since large language model inference repe…

📎 Sources

  1. jedec.org
  2. effectstory.com
  3. effectstory.com
  4. jedec.org

Related data

Author's TakeEffectStory 編輯部

The generational data show HBM scaling on two axes at once—bandwidth grew nearly eightfold from 256 GB/s to 2 TB/s while stack capacity rose from 8GB to 64GB—which matters because AI accelerators need both faster data movement and more local storage for growing model weights at the same time, not one or the other. The supply picture is the more interesting tension: SK Hynix's share slipped from 69% to 58% in a single year even as Samsung and Micron each hold roughly 21%, and all three now have HBM4 certified by Nvidia, which suggests buyers are actively diversifying their supplier base rather than staying locked to one vendor. The metric worth watching next is whether SK Hynix's share keeps eroding as Samsung's and Micron's HBM4 output ramps into shipped volume—that would confirm the diversification is structural rather than a one-quarter blip.

E
EffectStory 編輯部Editorial Team

Related

BRIEF

SpaceX and NVIDIA Plan to Launch Orbital AI Data Centers Starting Next Year, Musk Says

SpaceX plans to launch its first NVIDIA-chip-powered AI satellites, Starmind AI1, in the fourth quarter of 2027, reaching large-scale deployment by 2028 — at least a year ahead of the original schedule. The company has also filed with the FCC for a network of up to 1 million satellites, while Taiwanese suppliers Unitech and Sesoda report rising satellite-related revenue and order shares tied to the buildout.

林紀旭 James Lin ·
BRIEF

NVIDIA Launches Jetson Orin Nano 2 Robotics Computer, Targets Entry-Level Edge AI

NVIDIA announced the Jetson Orin Nano 2 robotics computer on August 25, 2026, packing 78 TOPS of AI compute, 8GB of memory, and an 8-core Arm CPU while doubling inference performance and cutting power draw 40% versus its predecessor. Modules and developer kits ship in the first half of 2027, targeting a robotics developer base NVIDIA says already tops 3 million.

Nathan ·