Surface Laptop Ultra With NVIDIA RTX Spark Runs 94GB Model, Finishes 4+ Enterprise Agent Workflows Per Hour
ryanshrout · x · 2026-11-19
Ryan Shrout cites Signal65 testing: the Microsoft Surface Laptop Ultra with NVIDIA RTX Spark packs 128GB of unified memory and the CUDA stack, enough to run complete enterprise AI agent workflows locally.
Key points:
- Running the 94GB Qwen3.8-Flash-Next model, the laptop correctly completes more than four enterprise workflows per hour, each up to 69 agent steps with retrieval over a 128K-context corpus.
- Signal65 built the PINNACLE Client benchmark around 'correct work finished per working day' rather than tokens per second for this class of PC.
- With 128GB unified-memory Windows machines shipping from multiple vendors (NVIDIA RTX Spark, AMD Ryzen AI Max, Intel next), models that were data center deployments 18 months ago now fit in a bag — local AI compute competition is set to intensify.
More from Infra
- Oracle co-CEO admits 'classic mistake' selling data centers to towns — jeremyakahn · 2026-11-19
- Nace ships 71.8GB DeepSeek V4.1 Flash coding quant with 39% lower KL than rivals — m4rkGS · 2026-11-19
- Dev open-sources self-hosted model registry so inference workers boot when HF is down — cmm324 · 2026-11-19
- Architect Labs unveils Frostbite, an agent-co-designed CPU claiming 4.6x perf per watt over Arm — IgorCarron · 2026-11-19
- OpenAI models boost AMD flagship GPU PPA 14% in 5 months, matching senior architects — 智东西 · 2026-11-19
- Nvidia's unusual delay in filing its 10-Q sparks Wall Street intrigue — analyst explains why — TechCrunch AI · 2026-11-19