Let's talk about Nvidia earnings.
In my note yesterday, we talked about the 2026 revenue pipeline — half-a-trillion-dollars worth (that's 2026, calendar year).
Jensen revealed last October that the demand was on the books already. But fulfilling it was/is a different story. It would take significant expansion of global manufacturing capacity.
With that, we had evidence in Q3 that Nvidia's growth was back (after seven quarters of stagnant growth) — that supply was opening up.
And we got more confirmation of that today, from Q4.
Here's what the quarterly revenue change looks like now.
That said, to fulfill on Jensen's half-a-trillion-dollar (datacenter revenue) target by the end of the year, manufacturing capacity will have to have another step-change — and soon.
But, what's rapidly carrying more of the growth load for Nvidia's data center revenue is networking.
Compute (Nvidia's GPUs) gets all of the attention. But networking growth is exploding. It did $11 billion last quarter — up 34% quarter-over-quarter, and 3.5x year-over-year. Nvidia is now (maybe) the largest ethernet networking company in the world.
Keep in mind, this is all Q4, prior to the game changer on February 5th.
What happened on Feb. 5th?
On that single day, two of the most important AI companies in the world released models that represented a big leap in model capabilities.
Anthropic released Claude Opus 4.6, with autonomous agent teams that can manage entire workflows in parallel. Not answering questions. Doing the work.
Twenty minutes later, OpenAI fired back with GPT-5.3-Codex — a model that helped build itself.
And then, almost overnight, an open-source personal AI agent called OpenClaw went viral. People began running autonomous AI assistants on their laptops, around the clock, with access to their files, their browsers, their messaging platforms.
This is the moment AI crossed from "AI can think" to "AI can act."
On that note, on the Nvidia call, Jensen proclaimed that the ChatGPT moment of Agentic AI has arrived.
What else did he say?
He said this line over and over: "compute equals revenue."
He's saying the capex boom isn't stopping. Because with each unit of compute added into the world, it's generating tokens, and tokens are generating revenue. And revenue equals more capex. It's proving productive for the consumer, and profitable for the datacenters.
It's a self-reinforcing loop.
And this new phase of "always-on" AI agents, means more and more economic output, higher productivity, and (related) higher potential economic growth.
That brings us back to Elon's thesis — his view that humanoid robots and autonomous AI will ultimately remove any meaningful limit on the size of the economy.
The logic is simple: if the cost of labor approaches zero, economic output has no ceiling.
