Nvidia

$96.22B in Q2 2026 revenue, up 106% year over year, underscored Nvidia’s shift from AI-chip bellwether to a focal point for AI infrastructure, policy and competition.

Who they are

Nvidia is the chipmaker at the center of coverage of the AI-compute buildout, led publicly by CEO Jensen Huang and best known in these stories for its flagship data-center GPUs. Its role is widening beyond hardware: coverage tracks consumer PC chips, open AI models and datasets, investments in AI companies, and agreements for inference technology and talent.

The recent arc

Coverage accelerated through 2025 and reached its high point in 2026Q3, as Nvidia’s AI-chip demand, product expansion and corporate reach converged. Huang’s March projection of more than $1 trillion in flagship AI-chip sales through 2027 was followed by Nvidia’s reported Q2 results: $96.22 billion in revenue, with $89 billion from Data Center. In June, the company also announced RTX Spark, an Arm-based consumer chip family developed with MediaTek and manufactured on TSMC’s 3nm process, extending the story beyond data centers.

The late-2026 phase has shifted toward the consequences of that scale. Nvidia’s non-exclusive agreement with Groq gave it rights to the startup’s inference technology and brought Groq CEO Jonathan Ross and other executives into Nvidia, while reporting said it was considering a $12.9 billion Hugging Face acquisition. The resulting scrutiny is now explicit: Reuters reported that the US DOJ asked about the Hugging Face deal’s effect on AI competition. At the same time, stories tied Nvidia to OpenAI financing and prospective infrastructure support for Nscale’s Anthropic data center.

The tension

The coverage repeatedly pits Nvidia’s AI-computing lead against two constraints: geopolitical access and emerging alternatives. China policy has directly shaped sales, from Trump’s conditional approval for H200 shipments to approved Chinese customers to reporting that Nvidia and AMD would share China-chip revenue with the US government as an export-license condition. Meanwhile, AMD is building its position with Meta, and reporting that OpenAI may buy Google TPUs highlights customers’ incentives to diversify away from Nvidia; the Groq deal also signals competition for inference technology and specialized talent.

Why it matters

If Nvidia sustains its data-center momentum while extending into inference software, consumer chips and AI-platform investments, it could exert influence across more layers of the AI stack than GPU supply alone. But that trajectory also makes China restrictions, customer diversification and antitrust oversight more consequential: the corpus shows these are not peripheral risks but forces already shaping where Nvidia can sell, whom it can partner with and how aggressively it can expand.

Related: AI · China · AMD · Jensen Huang · Google · Jensen Huang says Nvidia expects to generate $1T+ in sales from its fl