On September 29, President Trump signed Executive Order 14434, “Inaugurating the Era of Super Intelligence,” ordering federal agencies to stop saying “artificial intelligence” and start saying “Super Intelligence” — SI, for short — in official communications. Three weeks later, on October 21 earnings calls, Elon Musk and Mark Zuckerberg said “SI” and “superintelligence” like it was already the house style. Sam Altman, on a Politico podcast, said OpenAI has “no plans” to follow.

That’s not a trivia item. It’s a live test of which CEOs treat a federal naming preference as something to perform for, and which ones think the performance costs them more than it buys.

Quick Summary: What Happened

Detail	Info
The order	EO 14434, signed Sept 29, 2026 — directs federal agencies to use “Super Intelligence”/“SI” instead of “artificial intelligence”/“AI”
Legislative deadline	Agencies have until roughly Nov 28, 2026 (60 days) to propose statutory language defining “SI”
Complying on earnings calls	Musk (Tesla/SpaceX, Oct 21): humanoid robot demos “aren’t really using any kind of advanced AI or SI.” Zuckerberg (Meta, Oct 21): Meta’s goal is “delivering personal superintelligence to everyone.” Huang (Nvidia) praised the “White House Accord on Super Intelligence”
Refusing	Altman: OpenAI has “no plans” to rename itself, per reporting on his Politico podcast appearance
The catch	”Superintelligence” already has an established technical meaning in AI research — a system far more capable than any that exists today. The order applies the label to current-generation AI

Bottom line: A government naming preference just became a loyalty test on earnings calls, and the CEOs failing it aren’t the ones you’d expect.

What Actually Happened

Start with the order itself, because the text matters more than the headlines about it. EO 14434 doesn’t ban the words “artificial intelligence” — it tells executive departments and agencies to use “Super Intelligence” and “SI” instead, “to the maximum extent permitted by law,” across official correspondence, websites, reports, and policy documents. For now, the order defines the two terms as identical, pointing to the existing statutory AI definition in 15 U.S.C. § 9401(3). It also directs the White House science office to deliver proposed legislative language for a standalone federal definition of “SI” within 60 days — a deadline that lands around November 28.

In other words: right now, “SI” is a find-and-replace. What it becomes after Thanksgiving is an open question with real regulatory weight behind it.

Then came earnings season. CNBC tallied who played along and who didn’t across the last week of October, and the split doesn’t sort along any obvious technical or business logic:

Who Said “Superintelligence,” and Who Didn’t?

1. Elon Musk (Tesla, Oct 21): Describing humanoid robot demos, Musk said some “look very impressive, but they’re not really using any kind of advanced AI or SI.” SpaceX had already renamed its AI division SpaceXSI earlier in October.
2. Mark Zuckerberg (Meta, Oct 21): Framed Meta’s mission as “delivering personal superintelligence to everyone” — a phrase that predates the executive order, since Meta launched its Superintelligence Labs unit in 2025.
3. Jensen Huang (Nvidia): Posted that the “White House Accord on Super Intelligence” offers “an extraordinary opportunity to advance discovery, productivity, security, health, and prosperity.”
4. Andy Jassy (Amazon, Oct 29): Used “AI,” repeatedly, with no apparent SI pivot.
5. Satya Nadella (Microsoft, Oct 28): Talked about “designing AI to empower people” on the call — though Microsoft’s annual report reportedly uses “super intelligence” eight times elsewhere.
6. Sundar Pichai (Alphabet, Oct 29): Stuck with “AI investments” driving growth.
7. Sam Altman (OpenAI): Told Politico’s Decoded podcast that OpenAI has “no plans” for a name change, adding the company has “been talking about super intelligence a lot, and we will continue to talk about it” — as a research goal, not a rebrand.

Read that list again and the pattern isn’t “AI labs vs. everyone else.” It’s CEOs with the most direct exposure to the administration — Musk runs two companies competing for federal contracts, Huang needs export-license goodwill, Zuckerberg was already sitting on a convenient pre-existing name — adopting the term fastest. The CEOs with the least to gain from the optics, or the most to lose on precision, didn’t bother.

What Does “Superintelligence” Actually Mean — And Why Does the Mismatch Matter?

Superintelligence, as AI researchers have used the term for over a decade, means a system whose cognitive abilities greatly exceed human performance across virtually every domain — not a chatbot that’s merely good at drafting emails. Philosopher Nick Bostrom’s 2014 book popularized the definition, and OpenAI itself used the word this way in its own 2023 Superalignment announcement, framing superintelligence as a future capability level requiring entirely new safety research — explicitly beyond AGI, not a synonym for it.

That’s the mismatch nobody in the executive order addressed. Current AI — the kind Musk’s robots aren’t running, the kind Meta ships in a chat app — doesn’t meet that bar by anyone’s definition, including OpenAI’s own. The order doesn’t claim otherwise; it just swaps the label on today’s technology for a word that researchers reserve for technology that doesn’t exist yet. Fast Company put it plainly: Trump has said “artificial” makes the term “sound fake,” which suggests this was never about precision. It was about the vibe.

Altman’s quote on Decoded cuts right to that: he asked, point-blank, whether “superintelligence” even accomplishes the thing it’s supposedly for — “Does superintelligence sound obviously less scary to people than artificial intelligence?” That’s not a safety objection. It’s a marketing objection, delivered by someone whose company’s research division already has a specific, narrower claim staked on the word.

Why This Matters

This isn’t the first time Washington has tried to rename its way out of an AI anxiety problem. It’s the first time Big Tech CEOs have had to decide, live, on a call where every word gets transcribed and quoted back to investors, whether they’re willing to use the administration’s preferred vocabulary for a product that doesn’t warrant it.

The split tells you more about who’s managing a government relationship than about anyone’s actual technology roadmap. Nvidia needs the Commerce Department’s continued cooperation on export licenses. Musk runs SpaceX and Tesla, both of which depend on federal contracts and goodwill that have nothing to do with chatbots. Meta got to adopt the term for free, because “Superintelligence Labs” already existed as a brand before the order was a gleam in anyone’s eye. Altman, by contrast, runs the company most publicly associated with the term’s actual technical meaning — and undercutting that meaning for a federal naming preference costs OpenAI more credibility than it buys goodwill.

OpenAI has already been managing Washington relationships carefully on antitrust and regulatory fronts this year, so this isn’t a company indifferent to politics. It’s a company calculating that this particular fight — over a word its own researchers coined with a specific meaning — isn’t one worth conceding.

The Bigger Picture

Zoom out, and the rebrand lands on an industry that’s already arguing about naming inflation. OpenAI called GPT-6 Astra evidence of an “AGI era,” a claim that got picked apart almost immediately by researchers who don’t think a benchmark score settles what AGI even means. Now there’s a second, government-issued label — “Super Intelligence” — sitting one rung up the hype ladder from a term the industry hasn’t agreed on yet. Stacking an undefined political label on top of an undefined technical one doesn’t resolve either argument. It just gives companies a reason to use whichever term tests better with whichever audience they’re talking to that week.

It also isn’t the administration’s first attempt to shape how the public talks about AI risk. Trump has previously dismissed AI-slowdown warnings as a “sick conspiracy,” and the broader push for lighter federal AI oversight has been a consistent theme all year. A rebrand that makes “artificial” — Trump’s own word for what sounds “fake” — disappear from federal paperwork fits a pattern: control the vocabulary, and the policy debate shifts with it. Even the government’s own AI safety coordination effort isn’t immune — Huang’s “White House Accord on Super Intelligence” reference suggests the earlier White House AI safety pact picked up the new branding too.

Whether the Nov. 28 legislative language actually redefines “SI” as something more than a synonym for “AI” is the detail worth tracking. If Congress writes a real technical threshold into that definition, every company currently using the word loosely on earnings calls will have said, on the record, that their current products already qualify. That’s a statement with real legal and competitive weight, and it’s being made casually, right now, for optics.

Our Take

I think the honest read here is that almost nobody on that list of CEOs who said “SI” actually believes their product is superintelligent by the term’s real definition. They said the word because saying it costs nothing and might buy goodwill with an administration that controls contracts, export licenses, and regulatory attention. That’s not a technology decision. It’s a government-relations decision wearing a technology costume.

What I find more interesting than who complied is why Altman didn’t. OpenAI isn’t allergic to playing nice with Washington — the antitrust lobbying alone proves that. But “superintelligence” is a word OpenAI’s own safety team staked a specific, narrower meaning on back in 2023, and diluting it to cover whatever ships this quarter undercuts the one claim OpenAI gets to make that its competitors can’t: that when OpenAI says “superintelligence” is coming, it’s talking about something that doesn’t exist yet, not something you can already demo on an earnings call.

My prediction: this split holds until the Nov. 28 deadline forces an actual legal definition onto the term. After that, every CEO who said “SI” loosely in October will have to decide whether to keep using a word that now carries statutory weight, or quietly walk it back. Altman’s refusal to play along now looks less like stubbornness and more like someone who read the calendar.

Frequently Asked Questions

What does Executive Order 14434 actually require?

It directs federal agencies to use “Super Intelligence” and “SI” instead of “artificial intelligence” and “AI” in official communications, websites, reports, and policy documents, “to the maximum extent permitted by law.” It currently defines the two terms as identical and gives the White House science office 60 days — until roughly November 28, 2026 — to propose legislative language for a standalone federal definition.

Did Trump actually ban the term “artificial intelligence”?

No. The order applies to federal agency communications, not private companies, and it doesn’t prohibit the term “AI” outright — it instructs agencies to prefer “Super Intelligence” where legally possible. Private companies choosing to adopt the term on earnings calls are doing so voluntarily.

Why did Musk and Zuckerberg start saying “superintelligence”?

Both used the term on their companies’ October 21 earnings calls, following Trump’s late-September order. CNBC reported that SpaceX had already renamed its AI division SpaceXSI, and Meta’s use of “superintelligence” partly predates the order, since Meta launched its Superintelligence Labs unit in 2025.

Why won’t Sam Altman use the term?

Altman told Politico’s Decoded podcast that OpenAI has “no plans” to rename itself, and questioned whether “superintelligence” actually sounds less alarming to the public than “artificial intelligence.” He also distinguishes it as a specific future capability level OpenAI’s research is aimed at, not a label for current products.

Is “superintelligence” a real technical term, or just marketing?

It’s a real, pre-existing technical term. Researchers, including Nick Bostrom, define it as a system whose capabilities greatly exceed human performance across nearly all domains — well beyond today’s AI. OpenAI’s own 2023 Superalignment announcement used the term the same way, as a future milestone distinct from AGI.

Which other tech leaders have adopted or avoided the term?

Per CNBC’s reporting, Nvidia’s Jensen Huang, Intel’s Lip-Bu Tan, IBM’s Arvind Krishna, and Palo Alto Networks’ Nikesh Arora have all used “super intelligence” or “SI” publicly. Amazon’s Andy Jassy, Microsoft’s Satya Nadella, and Alphabet’s Sundar Pichai stuck with “AI” on their respective October earnings calls.

Does this rebrand change any AI regulations right now?

Not yet. The order doesn’t alter existing regulations, contracts, or grants on its own. The thing to watch is the legislative language due around November 28 — if Congress adopts a federal definition of “SI” that’s substantively different from “AI,” that could carry real regulatory consequences for how AI systems get classified and governed.

---

Last updated: November 1, 2026. Sources: The White House — Executive Order 14434 · Federal Register — Inaugurating the Era of Super Intelligence · CNBC — Trump’s tech allies swap AI for ‘super intelligence’ in earnings calls · Yahoo News — Trump really wants to make ‘super intelligence’ happen · Fast Company — Trump’s AI rebrand raises an awkward question for OpenAI · OpenAI — Introducing Superalignment · Wikipedia — Superintelligence: Paths, Dangers, Strategies.

Related reading: Altman Wants to Slow AI Down. Antitrust Says Maybe Not · GPT-6 Astra Lands: Inside OpenAI’s ‘AGI Era’ Claim · Trump Calls AI Slowdown Warnings a ‘Sick Conspiracy’ · Great American AI Act: What It Means for Tool Buyers · 6 AI Giants Signed a Safety Pact. Then Came the Firings