Nvidia Launches 64GB DGX Spark at $4,999 to Counter Memory Shortage

Key points

• Nvidia introduced a 64GB unified memory version of the DGX Spark, available from partners like Acer, Dell, and HP starting October 23, 2026.
• The new 64GB model is priced at $4,999, while the original 128GB version has increased to approximately $6,950, a 75% rise from its launch price.
• The 64GB unit supports models up to 100 billion parameters; clustering two units via NVIDIA Sync Cluster Assistant provides 128GB of memory and up to 1.7x performance improvement.
• The system includes a built-in ConnectX-7 NIC for direct clustering and supports NVIDIA's full AI software stack, including CUDA and Ollama.
• The release is a response to the 'memory crunch,' allowing Nvidia to maintain a lower price point for entry-level local AI workstations despite component cost hikes.

Background

Nvidia's DGX Spark has been a focal point in the local AI hardware market, previously compared to Apple's Mac Studio for its superior CUDA compatibility and multi-node scalability. Recent supply chain reports indicated tightening memory markets, with Micron's earnings highlighting the strength of the AI memory cycle. Additionally, early RTX Spark laptop launches faced allocation issues, suggesting broader component scarcity. The new 64GB Spark is a strategic adjustment to these market pressures, aiming to keep local AI accessible despite rising component costs.

How outlets are covering it

The Register and Tom's Hardware emphasize the 'memory crunch' and the price hike of the 128GB model as the primary drivers for the new 64GB release. Wccftech highlights the competitive disadvantage of the 64GB Spark against AMD's Ryzen AI systems, which offer higher memory configurations at lower price points. Nvidia's official blog focuses on the technical capabilities, such as clustering and software support, framing the 64GB model as an accessible entry point for developers rather than a compromise. The outlets generally agree on the pricing and specs but differ on the competitive landscape and the impact of the memory shortage on consumer choice.

Why it matters

The launch of the 64GB DGX Spark signals a shift in Nvidia's strategy to adapt to supply chain constraints while maintaining market presence in the local AI segment. It provides a more affordable entry point for developers and researchers, potentially expanding the user base for local AI workloads. The price increase of the 128GB model reflects the broader impact of the memory crunch on hardware pricing, which could influence purchasing decisions across the AI hardware market. This move also highlights the importance of clustering and software optimization in maximizing the utility of lower-memory configurations.

What to watch

Nvidia is expected to continue monitoring memory supply and pricing, potentially introducing further configurations or price adjustments. The availability of the 64GB model from major partners will be a key indicator of market demand. Additionally, the performance of clustered 64GB units in real-world workloads will determine their viability for larger AI tasks. The competitive response from AMD and other manufacturers will also shape the local AI hardware landscape in the coming months.