
Sambanova introduces new AI accelerator, partners with Intel to deploy Xeon CPUs for inferencing and agentic workloads
Memory will consume 30% of hyperscaler data center spending this year, a 4X increase over 2023
If ZAM works, it could directly compete with HBM in a massive and fast-growing market, reduce power consumption in AI data centers for substantial cost savings, and ease supply constraints through a more scalable manufacturing approach.
However, the technology remains at an early prototype stage, with a projected path to mass production around 2029. Historically, many “next-generation” memory concepts have failed to progress beyond lab demonstrations, making execution the key uncertainty.
The NEDO-supported program is expected to run for approximately 3.5 years, with SAIMEMORY planning to invest around JPY 8 billion (USD 5 million) through fiscal 2027 to develop working prototypes. The longer-term goal is to establish mass production by around 2029.
That places ZAM firmly in the next-generation memory cycle, rather than as an immediate replacement for current HBM deployments. In the meantime, incumbent memory manufacturers are continuing to evolve HBM with higher stack counts and improved efficiency.
NEDO’s support signifies the Japanese government’s intention to re-enter the chip and semiconductor market — amid skyrocketing AI demand — which it had dominated before Taiwanese and South Korean manufacturers took over market share.
“We view the selection of this Project under the NEDO program as a significant milestone in demonstrating Japan-originated next-generation memory technology to the world. ZAM represents an innovative architecture that achieves both the performance and power efficiency required in the AI era and in anticipation of the accelerating AI supercycle. Through collaboration with Intel, RIKEN, and our investors, as well as other domestic and international partners, we aim to contribute to strengthening the global competitiveness of Japan’s semiconductor industry,” Hideya Yamaguchi, President & CEO of SAIMEMORY.
SAIMEMORY’s development program is backed by a consortium that includes SoftBank, Fujitsu, RIKEN, and the Development Bank of Japan, alongside government support via NEDO.
Follow Tom's Hardware on Google News , or add us as a preferred source , to get our latest news, analysis, & reviews in your feeds.
Etiido Uko is an engineer and technical writer with over nine years of experience in documentation and reporting. He is deeply passionate about all things gadgets, technology, and engineering. ","collapsible":{"enabled":true,"maxHeight":250,"readMoreText":"Read more","readLessText":"Read less"}}), "https://slice.vanilla.futurecdn.net/13-4-22/js/authorBio.js"); } else { console.error('%c FTE ','background: #9306F9; color: #ffffff','no lazy slice hydration function available'); } Etiido Uko Social Links Navigation News Contributor Etiido Uko is an engineer and technical writer with over nine years of experience in documentation and reporting. He is deeply passionate about all things gadgets, technology, and engineering.
Key considerations
- Investor positioning can change fast
- Volatility remains possible near catalysts
- Macro rates and liquidity can dominate flows
Reference reading
- https://www.tomshardware.com/tech-industry/artificial-intelligence/SPONSORED_LINK_URL
- https://www.tomshardware.com/tech-industry/artificial-intelligence/softbank-subsidiary-working-with-intel-to-develop-radical-new-zam-memory-is-now-receiving-japanese-govt-subsidies-new-memory-designed-as-a-lower-power-hbm-for-ai-workloads#main
- https://www.tomshardware.com
- The Future of AI Is Open and Proprietary
- Game On: Five New Titles Now Streaming on GeForce NOW
- Ransomware negotiator pleads guilty after leaking victims' insurance details to 'BlackCat' hackers — perp gave attackers a precise picture of exactly how much e
- Bolt Graphics tapes out its first Zeus GPU test chip on TSMC 12nm — firm touts 17x lower cost of compute
- Game On: Five New Titles Now Streaming on GeForce NOW
Informational only. No financial advice. Do your own research.