Entity trail

IREN

Source-backed findings, relationship evidence, citations, and briefing history from the public MindPattern archive.

Briefing refs

Findings

Edges

Sources

Corpus findings

2026-08-16 / thought-leaders-researcherHugging Face's 'State of Open Models: Summer 2026' Puts Local Inference at Trillion-Parameter Scale — Kimi-K3 GGUF at ~2.8T Params, DeepSeek-V4-Flash at ~284BIrene Solaiman and Hugging Face's policy team published their quarterly open-model assessment on August 14, 2026, covering January through August. Public model repos grew from 2.43M to 2.96M, datasets from 711K to 1M, and Spaces from 1.00M to 1.44M over the period. The most consequential observation for builders: the July snapshot includes GGUF builds of DeepSeek-V4-Flash (~284B params) and Kimi-K3 (~2.8T params), meaning 'local inference' now routinely means a trillion-parameter MoE sharded across a few consumer machines — while the report flags that safety and evaluation gaps between open and closed models remain the recurring blocker for enterprise deployment.
2026-08-07 / rss-researcherMirendil, founded by ex-Anthropic and Google researchers, commits half its seed round to a $100M+ Google Cloud dealTechCrunch exclusively reported on August 6 that Mirendil — led by CEO Behnam Neyshabur with Harsh Mehta and co-founders from Anthropic — signed a multiyear Google Cloud partnership worth over $100 million for TPUs, Nvidia GPUs, and managed training clusters. The startup raised seed funding in late June 2026 at a $1 billion valuation, making the compute commitment roughly half the round. Its thesis is recursive self-improvement: "you can have a self-improving AI where you can point a problem at it and it keeps getting better with time," targeted at medicine, biology, and materials science.
2026-07-23 / arxiv-researcherTRAVEL: Rule-Guided MCTS Plus RL Lifts C-to-Rust Translation Accuracy 26% and Cuts Unsafe Code 13% on Linux Kernel and Huawei CodeLLMs translating C to Rust reliably violate Rust-specific rules and lose semantics on complex code. TRAVEL steers Monte Carlo Tree Search along Rust rule constraints during reasoning-path construction, then applies RL coupling execution feedback with reasoning-quality signals. Against the strongest prompting baseline IRENE on xCodeEval it improves computational accuracy 26.22% and compilation success 18.77%; on Linux-kernel functions (OS-Bench) and an industrial Huawei dataset it adds 16.51% and 18.28% compilation success while cutting unsafe-block rate by roughly 13% on each — the unsafe-rate number is the one that matters for a memory-safety migration.
2026-07-22 / saas-disruption-researcher34 Companies Joined the Unicorn Board in June — 10 Were AI Labs Worth $65B, Led by DeepSeek at $50BCrunchbase's June tally added $110B+ in new unicorn value across 34 companies, with 10 frontier AI labs accounting for $65B of it. DeepSeek alone is $50B; the remaining nine — Flourish ($2.5B), PhysicsX ($2.4B), General Intuition ($2.3B), Generalist AI ($2B), Bulage ($2B), Odyssey ($1.5B), Sarvam AI ($1.5B), Mirendil ($1B), Vast ($1B) — split $15B. The concentration is the story: one lab is 77% of the AI-lab value created that month, and the exit column included Anysphere (Cursor).
2026-05-10 / hn-researcherNvidia Commits $40B+ in AI Equity Investments in 2026, Financing the Entire AI Supply Chain — CNBCNvidia has committed over $40 billion in AI equity investments in 2026 alone, led by a $30 billion stake in OpenAI, plus new deals this week with data center operator IREN (up to $2.1B) and Corning (up to $3.2B for three new fiber-optic facilities). The strategy goes beyond chips — Nvidia is financing companies across the entire AI infrastructure stack while securing commercial deals that ensure deployments run on Nvidia hardware. Critics flag circular investment concerns, as Nvidia is effectively investing in its own customers, but the scale represents an unprecedented vertical integration play in AI infrastructure.

Graph relationships

INVESTED IN
NVIDIA
Nvidia committed up to $2.1B to data center operator IREN.
Source finding
INVESTED IN
NVIDIA -> IREN
Nvidia committed up to $2.1B to data center operator IREN.
Source finding

Source trail

CNBC reports Hugging Face TechCrunch arXiv 2607.19966 Crunchbase News

Graph sources

entity graphfindings textkg edgeskg entitiesnewsletter issues

Fetching from the wire…