What Is Super Intelligence (SI)?

“Super Intelligence” is a phrase with two distinct meanings in late 2026, and they are easy to confuse.

The first is new and administrative. Since September 29, 2026, “Super Intelligence” and “SI” have been the U.S. federal executive branch’s preferred terms for what the rest of the world still calls artificial intelligence. The second is older and technical: a hypothetical level of machine intelligence that would exceed human cognitive performance across essentially every domain. The first meaning is a label. The second is a goal that no existing system has reached.

This guide explains both, shows how companies use the word, and sets out how SI Builders uses it.

The federal usage: “SI” as the new name for AI

On September 29, 2026, President Trump signed Executive Order 14434, “Inaugurating the Era of Super Intelligence”. It was published in the Federal Register on October 2, 2026 as document 2026-20321.

The order’s rationale, from Section 1, is that “The terminology used by the Federal Government should reflect the transformative capabilities of these technologies.” Section 2 then directs that executive departments and agencies “shall use ‘Super Intelligence’ and ‘SI’ in place of ‘Artificial Intelligence’ and ‘AI’” in “official correspondence, public communications, websites, reports, policy documents, and other non-statutory documents,” and it qualifies that instruction with the phrase “To the maximum extent permitted by law.”

What the order does not do matters as much as what it does:

• It applies to federal executive departments and agencies. It does not bind Congress, the courts, states, or private companies.
• It is prospective. Section 2(b) states that “Nothing in this section requires the alteration of previously issued regulations, Presidential actions, contracts, grants” or other historical documents.
• It does not change any statute. Section 3(a) provides that the terms “Super Intelligence” and “SI” “mean the technologies and systems encompassed by the term ‘artificial intelligence’” as defined in 15 U.S.C. § 9401(3), “unless and until superseded by subsequent Presidential action” or an Act of Congress.

That statutory definition, enacted in the National Artificial Intelligence Initiative Act of 2020, describes “a machine-based system that can, for a given set of human-defined objectives, make predictions, recommendations or decisions influencing real or virtual environments.” In other words, the federal “SI” currently means exactly what federal “AI” meant: the same broad class of systems, under a new name.

The 60-day definition process

Section 3(b) sets up a possible future change. “Within 60 days of the date of this order,” the “Assistant to the President for Science and Technology (APST),” consulting other agencies as appropriate, “shall submit to the President proposed legislative language” to establish a federal definition of Super Intelligence. The White House fact sheet describes this as a task “to propose a Federal definition of ‘Super Intelligence’ and ‘SI’.”

Two points are worth stressing. First, the deliverable is a proposal of statutory language, which Freshfields calculates is due by November 28, 2026. Second, a proposal is not a law. Only Congress can amend 15 U.S.C. § 9401. Until that happens, the legal meaning of “SI” in federal documents remains the existing AI definition.

si.gov and NIST

Agencies have begun implementing the order. NIST’s Super Intelligence page now states: “Per the Sept. 29, 2026, Executive Order 14434: Inaugurating the Era of Super Intelligence, NIST is working to update its communications to incorporate the term ‘super intelligence’ as directed.” The page does not define the term or change the substance of NIST’s work, which still covers measurement, testing and evaluation, and standards.

The White House also relaunched its AI policy site at si.gov, which describes itself as “America’s home for Super Intelligence policy.” According to NewsBytes, the site was “originally created during the Biden administration” and moved to the new address, which was announced on X by Michael Kratsios, who heads the Office of Science and Technology Policy.

The research meaning: a level of capability, not a label

Long before 2026, “superintelligence” had a specific meaning in AI research and philosophy. Nick Bostrom’s 1998 paper “How Long Before Superintelligence?” defined it as “an intellect that is much smarter than the best human brains in practically every field.” His 2014 book Superintelligence: Paths, Dangers, Strategies refined this to “any intellect that greatly exceeds the cognitive performance of humans in virtually all domains of interest”.

In this tradition, the field distinguishes between artificial general intelligence (AGI), usually meaning roughly human-level general capability, and the stronger concept often written as “artificial superintelligence” (ASI). Google DeepMind’s June 2026 paper “From AGI to ASI” describes the transition from human-level AGI to “a system that is more intelligent and cognitively capable than large organizations of humans,” and discusses four pathways: “scaling AGI, AI paradigm shifts, recursive improvement,” and emergence from large multi-agent collectives.

The research meaning is a threshold. No company, and no government, has claimed to have crossed it. When the federal government calls a chatbot or a fraud-detection model “SI,” it is using the label, not making a capability claim.

How companies use the word

Several companies adopted the research sense of the word before the federal rename, usually to describe an ambition.

• Safe Superintelligence Inc. was announced on June 19, 2024 by Ilya Sutskever, Daniel Gross and Daniel Levy with the opening line “Superintelligence is within reach.” Its website still states: “Building safe superintelligence (SSI) is the most important technical problem of our time.”
• Meta Superintelligence Labs was created on June 30, 2025. Mark Zuckerberg’s memo said, “As the pace of AI progress accelerates, developing superintelligence is coming into sight,” and “We’re going to call our overall organization Meta Superintelligence Labs (MSL).” A July 30, 2025 letter added: “Meta’s vision is to bring personal superintelligence to everyone.”
• Microsoft AI announced on November 6, 2025 that it had “formed the MAI Superintelligence Team,” led by Mustafa Suleyman, to pursue what it calls humanist superintelligence: “incredibly advanced AI capabilities that always work for, in service of, people and humanity more generally.”
• Lila Sciences states on its website that “LILA is building Scientific Superintelligence,” positioned as a system “to solve humankind’s greatest challenges” through AI-directed laboratories.
• OpenAI has not renamed anything, but its leadership uses the research sense. In January 2025 Sam Altman wrote, “We are beginning to turn our aim beyond that, to superintelligence in the true sense of the word.” In June 2025 he wrote that “before anything else, we are a superintelligence research company” and that “Humanity is close to building digital superintelligence.” After the executive order, Altman said OpenAI had “no plans” to rename, adding, “we have been talking about super intelligence a lot, and we will continue to talk about it.”

Each of these is a statement of intent or branding. None is a claim that the research threshold has been reached.

How SI Builders uses the term

SI Builders writes Super Intelligence as two capitalized words in its own editorial voice, including when describing the research concept. We keep official spellings in proper names (Safe Superintelligence Inc., Meta Superintelligence Labs) and in direct quotations.

We distinguish the two meanings explicitly. When a page refers to the federal usage, it means the labeling convention established by Executive Order 14434, which currently maps to the statutory AI definition. When a page refers to the research concept, it means a hypothetical system exceeding human cognitive performance across domains. The directory tracks who has publicly set out to build the latter. It never implies that anyone has.

Corrections: corrections@sibuilders.io.