Nvidia

Nvidia designs GPUs, networking and software (CUDA) widely used for AI training and inference (author’s description). It is headquartered in Santa Clara, California (Wikipedia) and has made reported investments in AI labs since 2025. (The filename is non-kebab-case; rename before setting share: true.)

Overview

  • Founded: 1993 by Jensen Huang, Chris Malachowsky and Curtis Priem (Wikipedia). CEO: Jensen Huang.
  • Products: GeForce / RTX, data-centre GPUs and DGX systems (Blackwell generation; successor details not verified), CUDA, networking (NVLink), Jetson edge modules (nvidia-jetson-nano), Omniverse, Drive.
  • Models: the Nemotron open model family, see nvidia-nemotron.
  • Scale: market-cap milestones and fiscal 2026 financials are reported by Wikipedia (secondary) and not verified here; figures omitted.

Investments, deals and acquisitions (2025-2026)

  • OpenAI, Anthropic, xAI, Intel, Nebius: investments are reported in press and Wikipedia (see openai, anthropic, xai); amounts and dates are not verified and omitted.
  • Groq: CNBC (2025-12-24, URL in Sources) reports Nvidia buying Groq for about $20B; the deal structure is not verified. See groq.
  • MediaTek: Bloomberg and Tom’s Hardware headlines (search results) report a $3.5B investment announced 2026-08-31 tied to NVLink Fusion; details not verified.
  • Hugging Face: a TechCrunch headline (2026-09-03) reports Nvidia confirming an agreement to buy Hugging Face for about $12.9B; closing timeline and openness commitments are not verified.

Context

  • Cloud and custom-chip rivals: amd (OpenAI MI450 deal), Google TPUs, Broadcom/OpenAI chips.
  • China export rules (H20, H200) were not researched here.

Open items

Sources (fetched 2026-10-05)