Trouble in Your Tank: Yield Improvement and Reliability
Dan's Biz Bookshelf: 'The Sale'
The Marketing Minute: Wired for Trust
HBM4 Raises the Bar on Manufacturing Complexity, Premium Expected to Exceed 30%
May 22, 2025 | TrendForceEstimated reading time: 1 minute
TrendForce's latest findings reveal that demand for AI servers continues to accelerate the development of HBM technologies, with the three major suppliers actively advancing their HBM4 product roadmaps.
HBM4 introduces a more complex chip design that expands die size due to a significant increase in I/O count. Additionally, some vendors are also transitioning to a logic-based base die architecture to enhance performance, with both factors contributing to higher production costs. For reference, HBM3e debuted with an estimated 20% price premium; HBM4’s elevated manufacturing difficulty is expected to drive premiums above 30%.
Leading AI chipmaker NVIDIA unveiled its next-generation Rubin GPU at this year’s GTC, while AMD is preparing its MI400 series as a direct competitor. Both are expected to feature HBM4.
TrendForce notes that HBM4 doubles the I/O count from 1,024 to 2,048 compared to previous generations, while maintaining a data transfer rate above 8.0 Gbps—on par with HBM3e. This implies that HBM4 can deliver twice the data throughput at the same speed, thanks to its increased channel count.
Moreover, current HBM3e base dies utilize a memory-only architecture that acts as simple signal pass-throughs. In contrast, SK hynix and Samsung are collaborating with foundries to adopt a logic-based base die design for HBM4. This new approach enables closer integration between HBM and SoC, offering lower latency, improved data path efficiency, and greater stability in high-speed transmission environments.
Amid robust demand, TrendForce forecasts total HBM shipments to surpass 30 billion gigabits in 2026. HBM4’s market share is expected to grow steadily as suppliers ramp up production, ultimately overtaking HBM3e as the mainstream solution by the second half of 2026. SK hynix is projected to retain its leadership with over 50% market share, while Samsung and Micron will need to further improve yields and capacity to narrow the gap in the HBM4 race.
Share on:
Suggested Items
Cadence Enables Next-Gen AI and HPC Systems with Industry’s Fastest HBM4 12.8Gbps IP Memory System Solution
04/21/2025 | Cadence Design SystemsCadence announced the industry’s fastest HBM4 12.8Gbps memory IP solution, which meets the increasingly higher memory bandwidth needs of SoCs targeted for the next generation of AI training and HPC hardware systems.
SK hynix Leads the Market with HBM3e 16hi Products, Boosting Capacity Limits
11/14/2024 | TrendForceSK hynix recently unveiled its development of HBM3e 16hi memory at the SK AI Summit 2024, featuring a 48 GB capacity per cube, with sampling scheduled for the first half of 2025.
JEDEC Approaches Finalization of HBM4 Standard, Eyes Future Innovations
07/15/2024 | JEDECJEDEC Solid State Technology Association, the global leader in the development of standards for the microelectronics industry, announced it is nearing completion of the next version of its highly anticipated High Bandwidth Memory (HBM) DRAM standard: HBM4.
Manufacturers Anticipate Completion of NVIDIA’s HBM3e Verification by 1Q24; HBM4 Expected to Launch in 2026
11/28/2023 | TrendForceTrendForce’s latest research into the HBM market indicates that NVIDIA plans to diversify its HBM suppliers for more robust and efficient supply chain management. Samsung’s HBM3 (24GB) is anticipated to complete verification with NVIDIA by December this year.