PulseAugur
EN
LIVE 08:12:27

VKORE software enables NPUs for AI training, challenging Nvidia dominance

By PulseAugur Editorial · [1 sources] ·

A new software acceleration layer called VKORE enables Neural Processing Units (NPUs) to be used for AI model training, not just inference. In performance tests, pretraining throughput on an NPU using VKORE surpassed a single Nvidia H100 by 170% and approached a Nvidia B200 at 75%. This development allows for "sovereign compute" by enabling full AI model development cycles on non-Nvidia hardware, addressing supply and policy concerns that limit access to GPUs in certain markets, particularly in China. AI

IMPACT Enables full AI model development cycles on non-Nvidia hardware, potentially reducing reliance on GPU supply chains and lowering training costs.

RANK_REASON New software enables a new class of hardware (NPUs) to perform a critical AI task (training) previously exclusive to a dominant vendor's hardware (Nvidia GPUs), impacting data center economics and sovereign compute initiatives. [lever_c_demoted from significant: ic=1 ai=1.0]

Read on dev.to — LLM tag →

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →

VKORE software enables NPUs for AI training, challenging Nvidia dominance

How we ranked this

Signal score
34 / 100
Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket
Research
New software enables a new class of hardware (NPUs) to perform a critical AI task (training) previously exclusive to a dominant vendor's hardware (Nvidia GPUs), impacting data center economics and …
Source corroboration
Single-source cluster
Only one publisher covered this so far. Single-source stories can still rank when the publisher is high-authority, but they lack cross-source corroboration.
Topics
infra, model release
Editorial topic classification. Feeds into how the story surfaces on /topic/<slug> hub pages and into the per-entity coverage mix.
AI-industry relevance
High
Clearly on-topic for AI-industry coverage.
Story freshness
Breaking (< 6h)
Fresh story with cross-source coverage still developing. Ranking may shift as more sources report.

Full methodology in our editorial standards.

COVERAGE [1]

  1. dev.to — LLM tag TIER_1 English(EN) · GINIGEN AI ·

    Sovereign AI Compute: Pretraining on an NPU at 170% of an H100, 75% of a B200

    <p><strong>TL;DR:</strong> NPUs are sold as inference-only chips. VKORE, a hardware-independent acceleration layer, turns them into training hardware too. In our measurements, pretraining throughput on an NPU reached 170% of a single Nvidia H100 and 75% of a single B200 per devic…

RELATED ENTITIES

RELATED TOPICS