NuAxiSemiconductor trade intelligence
DeepSeek's new models are so efficient they'll run on a toaster ... by which we mean Huawei's NPUs
What happened
DeepSeek released its V4 and V4-Pro open-weight LLMs, claiming benchmark performance rivaling top proprietary models and introducing architectural changes for improved inference efficiency. The models extend support for Huawei's Ascend AI accelerators. Neutral: This is a software/algorithm development with no direct, immediate impact on semiconductor component supply, demand, or pricing.
Impact and signal
Low impactImpact score: 10/100
Affected components and suppliers
AI Accelerator
ai-modelinference-efficiencyhuawei-ascend
Source and time
Source: The Register · 4/24/2026 · View original source
Open the NuAxi live feed →
Real-time pricing, supply signals and AI analysis for semiconductor traders