Wednesday, August 26, 2026
Source trace. Via News points to the documents behind its reporting and shows what we drew from each — so you can check any claim. How we source
Press releaseGlobeNewswire· March 24, 2026

0G Labs Trained World's Largest Decentralized AI Model at 107B Parameters in 2025 - Eight Months Before This Week's Industry Headlines

View original at globenewswire.com
0G Labs Trained World's Largest Decentralized AI Model at 107B Parameters in 2025 - Eight Months Before This Week's Industry Headlines San Francisco, CA, March 24, 2026 (GLOBE NEWSWIRE) -- While the crypto industry celebrated Bittensor's Covenant-72B this week as a breakthrough in decentralized AI training, 0G Labs had…
Opening lines of the source · GlobeNewswire · short snapshot — read the full document at the original

What we drew from this source

The claims Via News extracted from this document. We point to the source; we don't replace it.

  • While the industry celebrated Bittensor's 72B model this week, 0G had already trained 107B parameters in July 2025 - 48% larger, 8 months earlier

    60% confidence
  • This isn't about breaking records, it's about building AI as a public good

    60% confidence
  • We proved decentralized infrastructure can train a 107 billion parameter model in 2025, before anyone else

    60% confidence
  • 0G's approach achieves approximately 95% cost reduction compared to centralized GPU cluster training

    60% confidence
  • This week's headlines celebrating 72 billion parameters as a milestone missed that 0G had already operated at significantly larger scale

    60% confidence
  • 0G set the benchmark for decentralized AI training on standard consumer bandwidth

    60% confidence
  • The industry is finally paying attention to decentralized AI

    60% confidence
  • Distributed, open-source AI training is complementary to centralized approaches and will play a growing role in frontier model development

    60% confidence

Cited in these Via News reports

What we know · the intelligence behind this page
Live from the substrate
What we're seeing
AI Leadership Exodus Rattles Investor Confidence Amid Capex Boom
High-profile departures at top AI labs — Brad Lightcap's exit from OpenAI and an unnamed researcher's departure from Alphabet/Google that triggered a share-price drop — are surfacing talent retention as a market risk factor even as hyperscalers pour record capital into AI infrastructure. The reaction shows investors treating key-person risk at frontier AI labs as material to valuation, a new fragility layered onto an otherwise bullish AI-driven capex cycle.
Our read on the data ›
Signals we're tracking
EPKINLY Regulatory-Clinical Success Cascade
High probability of expanded label indications, additional combination approvals, and competitive positioning strength in follicular lymphoma market. Predicts positive commercial uptake and potential accelerated review for related indications.
Patterns we're watching ›
Where sources disagree
Broadcom Inc.
Both facts report EPS for Broadcom Inc. for the same fiscal period (Q1 2026) observed on the same date (2026-02-01). However, they report conflicting values: 1.5 USD per share vs 2.05 USD per share. This is a 37% difference for the identical metric and time period, not a value change over time.
We flag conflicts openly ›
Recently verified
Checked against the original source
4,978
facts traced to their source — and we flag the ones that don't hold up.
101 entities tracked4,978 facts checked against source5,251 source documents archived
Query this data → isubstrate.com