[News] Microsoft Expands Local AI Push With NVIDIA RTX Spark; Surface Ultra Priced Up to $5,899 as Memory Costs Rise
Microsoft is stepping up its push to bring more powerful AI capabilities directly onto PCs as it seeks to make Windows a platform for AI agents. According to Reuters, the company unveiled an AI coding model that can run directly on PCs and demonstrated that models previously dependent on cloud access can now run locally on high-powered Windows machines. At the same event, the Windows maker also introduced the NVIDIA RTX Spark-powered Surface Laptop Ultra.
Microsoft is taking a similar approach with Copilot, keeping the most demanding tasks in the cloud while allowing local models to handle workloads where cost or privacy matter more, according to Microsoft Copilot chief Jacob Andreou. As AI agents become more autonomous, Microsoft also introduced Microsoft Execution Containers to prevent unauthorized data access and tasks on PCs. Anthropic, OpenAI, and NVIDIA will use the technology, according to Microsoft executive Pavan Davuluri, as noted by Reuters.
Rising Memory Costs Challenge Local AI Economics
The Surface Laptop Ultra starts at $2,599 and reaching $5,899 for a model with a 20-core processor, 128 GB of memory, and 1 TB of storage, Reuters notes. NVIDIA says RTX Spark supports up to 128 GB of unified memory and can run models with up to 120 billion parameters locally. The platform supports DirectX, OpenGL, and CUDA applications, with NVIDIA having tested more than 1,200 applications for performance and compatibility on RTX Spark. NVIDIA CEO Jensen Huang described it as “the world’s only computer” capable of natively running all applications on-device, according to ITHome.
The first wave of RTX Spark laptops is largely targeting high-end creators and local AI users. Models from Microsoft, Dell, Lenovo, MSI, HP, and ASUS start at around $2,599 and reach nearly $7,000, according to The Verge. By comparison, Reuters notes that Apple’s MacBook Pro with a 40-core graphics chip, 128 GB of memory, and 2 TB of storage costs $6,700, although the two machines are not directly comparable. Microsoft’s entry-level Surface Laptop Ultra is priced above Apple’s $1,999 base MacBook Pro, though it offers 24 GB of memory compared with 16 GB on Apple’s entry-level model, Reuters adds.
Reuters notes that the pricing marks a notable shift from when Microsoft pitched running AI workloads on PCs two years ago partly as a way to reduce costs, when such laptops were mostly priced below $2,000. A memory-chip crunch has since driven up hardware costs.
The impact is already visible in NVIDIA’s desktop AI systems. Its 128 GB DGX Spark now costs $6,950, about 75% above its launch price. Meanwhile, 64 GB versions with half the memory are available from NVIDIA hardware partners for around $4,999, according to The Register.
Read more
- [News] NVIDIA’s Supply Commitments Soar to $279B as Memory Costs Surge; New NVHBM Boosts Bandwidth 30%
- [News] Microsoft Eyes Maia 300 Launch in Sep., Reportedly Seeks 300K TSMC Capacity as Google, AWS Race Ahead
(Photo credit: Microsoft)
