Strategic PartnersIssue 646 records
NVIDIA
A governance-and-constraints edition. NVIDIA's board opened an independent investigation on Nov 8 after reports its CEO was the target of a foreign intelligence effort; CNBC links the related reporting to Megaspeed, an NVIDIA customer accused of helping restricted Chinese companies obtain chips, though the allegations are unproven and no SLED effect is shown. A Nov 5 report says NVIDIA is exploring small AI sites in telecom buildings to sidestep data-center opposition. A Nov 4 independent preprint finds disaggregated serving gains of at most 1.17x on GB200 NVL72, while NVIDIA's FLUX 3 tutorial and Fermaca's Mexico AI factory plan remain vendor claims. No new measured U.S. state, local or K-12 NVIDIA deployment was found.
- Evidence records
- 6
- Cross-source patterns
- 3
- Evidence classes
- 3 vendor claim2 independent reporting1 academic research
- Outcomes
- 4 emerging1 cautionary1 mixed
- Source freshness
- 6 new this fortnight
- Research completed
- 2026-11-09
Choose a role to see its takeaway beside every record in the ledger.
Patterns across the evidence
Synthesis · Lighthouse Advisory interpretation
NVIDIA ecosystem capacity plans are phased and depend on siting, power and permits, not just chip supply
Two sources show infrastructure growth shaped by siting and phasing. Fermaca's press release targets 60 MW by end-2027 with an initial Blackwell Ultra cluster before Vera Rubin, and a separate 1 GW IT ambition. The Yahoo/The Information report describes NVIDIA exploring retrofits of existing telecom buildings partly to avoid new-build opposition, while noting retrofits still need local permits. Both are unconfirmed plans, not delivered capacity.
Operating questionFor any GPU capacity proposal, what is live today versus planned, and who owns power, cooling and permit approvals?
Supporting evidenceFermacaYahoo Finance (relaying The Information)
Reported NVIDIA-stack speedups are workload- and configuration-conditional
NVIDIA's FLUX 3 tutorial reports up to 3.46x from combined optimizations on one B200, with fidelity trade-offs per option. The independent FastAFD preprint finds an attention-FFN disaggregation gain of at most 1.17x over a tuned baseline and parity on one model. Together they show that headline speedups depend on baseline, model and quality tolerance, so customer-specific profiling should precede commitments.
Operating questionWhat baseline, model and quality gate will we use to validate any claimed speedup on our own workloads?
Supporting evidenceNVIDIA Technical BlogFu et al. (arXiv cs.DC)
Supplier governance and export-control questions are now part of NVIDIA platform diligence
NVIDIA's board announced an independent investigation on Nov 8 and states no reason to believe secrets were compromised, while CNBC reports separate concern about Megaspeed, a customer accused of helping restricted Chinese companies obtain chips. The allegations are unproven and no SLED effect is shown; the common operating implication is to add provider provenance and supplier-risk checks to GPU procurement.
Operating questionWhich of our GPU capacity sources require documented provenance, compliance attestations and a named risk owner?
Evidence ledger
Full record · every source keeps its link and limitations
NVIDIA board opens independent investigation after reports its CEO was the target of a foreign intelligence effort
On Nov 8 NVIDIA's board said a committee of independent directors, with outside counsel, will investigate recent reporting that CEO Jensen Huang was the target of a foreign intelligence effort. The board says it has no reason to believe national security or NVIDIA secrets were compromised and expresses confidence in Huang. This is NVIDIA's own statement; it contains no findings and no timeline.
Why it matters, evidence and limitations
- Why it matters
- Government and education buyers of NVIDIA-based platforms may be asked about supplier governance and national-security posture. Relevance is indirect: no product, service or contract is affected in this statement, and no SLED customer is mentioned.
- Evidence and measured results
- NVIDIA board statement dated Nov 8, 2026; companion CEO message to employees listed on the same newsroom page. Single vendor source, unverified by investigators.
- Limitations and uncertainty
- Vendor statement only; outcome of the inquiry is unknown, and independent confirmation of facts comes from separate reporting (see CNBC item). No claim here about product or supply-chain impact.
CNBC: Nvidia investigates CEO's relationship with a woman affiliated with Megaspeed, a customer accused of helping restricted Chinese companies obtain chips
CNBC reports, citing New York Times and Financial Times reporting, that U.S. intelligence agencies have expressed concern about a businesswoman affiliated with Megaspeed, an NVIDIA customer accused of helping restricted Chinese companies acquire NVIDIA AI chips; NVIDIA says Megaspeed was previously investigated for evading export controls. CNBC notes the U.S. restricts exports of NVIDIA's most advanced chips to China, that NVIDIA reports Q3 earnings Nov 17 and holds a Washington conference at month end.
Why it matters, evidence and limitations
- Why it matters
- For SLED buyers the relevance is export-control and supply-chain assurance: public institutions sourcing GPU capacity through neoclouds or resellers may need provenance and customer-screening questions. No SLED deployment is described.
- Evidence and measured results
- CNBC article published Nov 8, 2026 3:12 PM EST; allegations are attributed to other outlets' reporting and unnamed U.S. intelligence concerns; NVIDIA says it has no reason to believe secrets were compromised.
- Limitations and uncertainty
- Allegations are attributed and unproven; CNBC relays NYT/FT reporting rather than primary evidence; investigation outcome unknown. Personal details are deliberately omitted here as not relevant to SLED applicability.
Report: Nvidia explores retrofitting Apollo-backed Brightspeed telecom buildings into small AI data centers to avoid new-build opposition
Relaying The Information, Yahoo Finance reports NVIDIA approached Apollo about converting existing telecom buildings, centered on Brightspeed's 130+ sites with about 1 GW of available power, into a network of small 'inference' sites across as many as 20 states, with possible retrofits using Vera Rubin racks from Q1 2027. Retrofits would still need local permits for specialized cooling, and buyers of the capacity are unclear.
Why it matters, evidence and limitations
- Why it matters
- Directly relevant to local and state governments: siting, permitting, grid and moratorium decisions on data centers. The report implies small distributed AI sites could appear in existing telecom buildings, which would involve local zoning and utility review.
- Evidence and measured results
- Secondary, unconfirmed report; Yahoo Finance article dated Nov 5, 2026, citing The Information and a Datum CEO comment that older telecom buildings can be costlier to run.
- Limitations and uncertainty
- Unconfirmed, relayed reporting; NVIDIA has not announced the plan; Brightspeed is described as nearing bankruptcy; purchasers and counts of sites unknown; stock-market commentary in the article is not evidence of technical outcomes.
Preprint FastAFD: attention-FFN disaggregation gains up to 1.17x over the best wide-EP baseline on GB200 NVL72, and not on every model
The preprint builds FastAFD for GB200 NVL72 and compares attention-FFN disaggregation with the best vLLM wide expert-parallel setting. AFD gains up to 1.17x on Qwen3 and 1.16x on Qwen3.5 but only matches wide EP on MiniMax-M2.5. An analytical model predicts AFD throughput within 3% from colocated profiles, so operators can estimate payoff before deploying.
Why it matters, evidence and limitations
- Why it matters
- Mostly relevant to research universities and state research computing running MoE inference on rack-scale NVLink; a modest, workload-dependent gain rather than a general uplift. No SLED deployment is evaluated.
- Evidence and measured results
- arXiv:2611.07149v1 submitted Nov 4, 2026; abstract opened; experiments on GB200 NVL72 with Qwen3, Qwen3.5 and MiniMax-M2.5; code reported open source.
- Limitations and uncertainty
- Preprint, not peer reviewed. Authors' systems and models; only abstract inspected here, full methods not independently reproduced; the 2.48x Groq LPX/Vera Rubin figure is a projection, not a measurement.
NVIDIA tutorial: FLUX 3 Image Klein on Blackwell with TensorRT LLM VisualGen reports up to 3.46x faster text-to-image on one B200, with quality trade-offs
NVIDIA reports combined optimizations (NVFP4, fused kernels, torch.compile, FP8 attention, sparse attention, step caching) deliver up to 3.46x faster text-to-image and 4.45x faster image-to-image for FLUX 3 Image Klein Base versus uncompiled BF16 at 2048x2048 on one B200. Individual options trade fidelity: LPIPS scores are reported for FP8 attention and caching. Eight B200s cut median T2I latency from 14.90 to 3.00 seconds for Base.
Why it matters, evidence and limitations
- Why it matters
- Limited and indirect relevance: image-generation serving for design, communications or media programs in universities and agencies. No SLED deployment or independent benchmark is included.
- Evidence and measured results
- NVIDIA Technical Blog post dated Nov 6, 2026 with YAML configurations; numbers are NVIDIA-run on B200 hardware; model is from Black Forest Labs.
- Limitations and uncertainty
- Vendor-run measurements on B200 only; quality measured with LPIPS on selected settings; no independent replication; speedups vary by option and not all combinations preserve fidelity.
Fermaca becomes Mexico's first NVIDIA Cloud Partner; phased AI factory targets 60 MW by end-2027
Fermaca announced it is the first NVIDIA Cloud Partner based in Mexico and unveiled an AI factory platform, starting with Fermaca Digital Origins in Mexico City with an initial NVIDIA Blackwell Ultra (B300 NVL72) cluster and a Vera Rubin roadmap. Inventec supplies compute and DDN storage. It targets 60 MW of total facility power for operational AI facilities by Dec 31, 2027, and a longer-term ambition of up to 1 GW of IT capacity at Durango.
Why it matters, evidence and limitations
- Why it matters
- Limited SLED applicability: international, early-stage and company-announced. It illustrates the NVIDIA Cloud Partner ecosystem expanding by region, which may matter for institutions comparing sovereign or regional GPU capacity abroad.
- Evidence and measured results
- Fermaca press release dated Nov 4, 2026 on PR Newswire; includes an NVIDIA enterprise director quote; no customer, benchmark or delivery evidence.
- Limitations and uncertainty
- Company press release; plans, not delivered capacity; no named customers, timeline for first service or pricing; Mexico-specific and not transferable to U.S. SLED without caveats.
How to read this edition
Source findings, measured results and limitations come from the cited publications. Patterns, operating questions, role takeaways and implementation considerations are Lighthouse Advisory interpretation, stated as questions to validate locally rather than guaranteed outcomes. Vendor and operator claims are labeled as claims. Full research method.
- Vendor claim
- A supplier-provided assertion that has not been upgraded to independent evidence.
- Independent reporting
- Independent reporting with attributable sources but without a formal evaluation design.
- Academic research
- Research produced through an academic institution or peer-reviewed venue.