PulseAugur
EN
LIVE 01:51:59

Preferred Networks trains finance LLM on custom chips, bypassing Nvidia

By PulseAugur Editorial · [1 sources] ·

Preferred Networks has developed an 8-billion parameter finance-specific large language model named PLaMo 3.0 Fin. This model was trained using their in-house MN-Core 2 chips, eschewing traditional Nvidia GPUs. This approach aligns with Japan's focus on data sovereignty, fostering a unique market for custom-built, on-premise architectures tailored for the banking sector. AI

IMPACT This development highlights a growing trend of specialized hardware and models for specific industries, potentially reducing reliance on general-purpose cloud AI infrastructure.

RANK_REASON Frontier-lab model release with system card [lever_c_demoted from frontier_release: ic=1 ai=1.0]

Read on Mastodon — mastodon.social →

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

Preferred Networks trains finance LLM on custom chips, bypassing Nvidia

How we ranked this

Signal score
3 / 100
Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket
Significant
Frontier-lab model release with system card [lever_c_demoted from frontier_release: ic=1 ai=1.0]
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
model release, infra
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. Mastodon — mastodon.social TIER_1 English(EN) · asiaai ·

    Preferred Networks just trained its 8B finance model, PLaMo 3.0 Fin, on its proprietary MN-Core 2 chips rather than Nvidia GPUs. While global labs push cloud AP

    Preferred Networks just trained its 8B finance model, PLaMo 3.0 Fin, on its proprietary MN-Core 2 chips rather than Nvidia GPUs. While global labs push cloud APIs, Japan's push for data sovereignty creates a real moat for domestic, custom-silicon architectures built specifically …

    LINKS asiaai.fyi/east-asia-ai-chip-news-2026-11…

RELATED ENTITIES

RELATED TOPICS