Tech News Assistant Editor
Tez Mukamba
Samsung Electronics has made a major leap in the artificial intelligence hardware race, as its PM9E1 solid-state drive (SSD) is now confirmed to be installed in NVIDIA’s latest ultra-compact AI supercomputer, the DGX Spark.
The development signals a renewed partnership between Samsung and NVIDIA, as both companies deepen their collaboration in next-generation AI computing devices.
A Compact Powerhouse for the AI Era
According to semiconductor industry reports on Sunday, NVIDIA’s DGX Spark — introduced by CEO Jensen Huang at CES 2025 — is equipped with Samsung’s PM9E1 SSD, unveiled in October 2024.
The DGX Spark, roughly the size of a human palm (15 cm × 15 cm × 5 cm), is powered by NVIDIA’s GB10 Grace Blackwell superchip, which is capable of performing up to one quadrillion (1,000 trillion) AI operations per second.
Unlike traditional enterprise servers, the device is designed for AI developers and researchers, offering powerful inference capabilities in a compact, energy-efficient form. Huang famously introduced it as “Project Digits” earlier this year and even personally delivered units to tech leaders like Elon Musk and Sam Altman following its global release in September.
Samsung’s PM9E1: Built for High-Speed AI Processing
The Samsung PM9E1 SSD is engineered specifically for AI-enabled PCs and compact computing systems. Samsung explained that the device can load a 14GB large language model (LLM) into DRAM in just one second, drastically improving AI model responsiveness and energy efficiency.
Available in 1TB and 4TB configurations, the PM9E1 offers industry-leading read/write speeds, according to tech analysts. Its performance makes it ideal for real-time data access and local model training — critical requirements for modern AI applications.
A semiconductor expert noted:
“The PM9E1 ranks among the fastest SSDs in its class and is especially suited for handling AI model workloads. NVIDIA’s choice highlights Samsung’s technical strength in this field.”
Samsung’s Strategic Comeback in AI Components
While Samsung has struggled in the high-bandwidth memory (HBM) sector — a market dominated by rival SK Hynix — it has steadily regained ground by supplying general-purpose DRAMs such as LPDDR and GDDR for NVIDIA’s new AI products.
Industry insiders say that Samsung’s inclusion in NVIDIA’s DGX Spark represents a rebuilt trust relationship and could pave the way for broader collaboration in advanced memory and storage solutions.
Samsung recently passed NVIDIA’s quality tests for HBM3E memory, signaling its readiness to participate in future HBM4 supply agreements, a development that could significantly boost its position in the AI chip ecosystem.
The Bigger Picture: SSDs Driving the AI Storage Boom
Analysts predict that demand for high-performance SSDs will surge alongside the expansion of AI computing, especially as companies seek faster, smaller, and more energy-efficient systems.
An industry source added:
“NVIDIA’s focus on compact AI PCs like the DGX Spark shows a clear trend toward edge-based AI development. Once a supplier is integrated into a flagship model, there’s a high chance of continuity in future versions — which is excellent news for Samsung.”
With this move, Samsung Electronics is positioning itself not just as a memory leader but as a core enabler of AI innovation, leveraging its advanced SSD technology to fuel the next generation of intelligent computing.
