Trump Launches Super Intelligence Force for AI Race

Trump has formally placed America's intelligence apparatus at the center of its AI strategy. The "Super Intelligence Force," signed into existence by executive order on October 4, puts the Director of National Intelligence at the helm of a sweeping cross-agency task force that will define what superintelligence means under American law, who gets to build it, and under what conditions. That single personnel decision, appointing Jay Clayton, a former Goldman Sachs partner and SEC chairman who now runs US intelligence, rather than a computer scientist, an AI safety researcher, or the Commerce Secretary, is the clearest signal yet of how Washington views artificial intelligence: not as a product to be regulated, not as a public good to be supported, but as a strategic weapon to be controlled, classified when necessary, and wielded as a national power instrument. The Super Intelligence Force treats AI the way America has historically treated nuclear capabilities: as a sovereign asset whose incidents, development timelines, and competitive advantages are matters of national security, not public discourse.

What Actually Happened

President Trump signed an executive order on October 4 establishing the Super Intelligence Force, a new cross-agency structure tasked with coordinating the entire federal government around AI dominance. According to Semafor, the SIF is formally led by Director of National Intelligence Jay Clayton, with FTC Chairman Andrew Ferguson, OPM Director Scott Kupor, and Emil Michael, the Undersecretary of War for Research and Engineering, serving alongside him. Vice President JD Vance, Defense Secretary Pete Hegseth, and Treasury Secretary Scott Bessent round out the full membership, and Deputy Chief of Staff Richard Walters is listed as a participant. The outside advisory layer, which carries no formal power but outsized influence, includes former AI czar David Sacks, who built the Trump administration's original AI executive order framework, and former Secretary of State Condoleezza Rice, whose foreign policy expertise signals that the SIF's scope extends well beyond domestic regulation. This is not a technology committee. It is a national security structure with technology as its brief.

The SIF's stated mission, as reported by CBS News, is to "coordinate the effort of the Federal Government to ensure that America continues to lead the World in Super Intelligence, which many say is bigger than the Industrial Revolution, and the Internet, and will protect the interests, and improve the lives, of all Americans." The order requires the SIF to deliver a comprehensive report within 120 days, covering the full range of AI risks and opportunities and the federal government's role in each. That deadline falls in early February 2027, and the report is expected to include a working definition of what the order calls "Super Intelligence Companies," a legal category that does not currently exist in US statute. Every AI lab in America will spend the next four months trying to influence how that category is drawn, because the companies that qualify will face new oversight obligations while also gaining access to federal protection, contracts, and strategic support.

The timing of the announcement was deliberate. ABC7 reported that the executive order came directly after a coordinated push from AI safety researchers and several Democratic legislators for a voluntary pause on frontier AI development. Trump's framing in the order is an explicit rejection of that position: the document describes any slowdown as a threat to American leadership and frames the SIF as the mechanism for ensuring that threat is never acted on by any US government agency. The president also referenced China's accelerating AI infrastructure investments and a wave of European regulations in the same remarks, framing the SIF as America's answer to geopolitical pressure rather than a response to domestic safety concerns. The message was unambiguous: American AI moves faster, not slower, and the intelligence community now holds the accelerator.

Why This Matters More Than People Think

The architecture of the SIF reveals a fundamental reclassification of AI in US policy. For the previous four years, AI governance in Washington ran primarily through NIST, the National Science Foundation, and the FTC, all civilian agencies with mandates rooted in commerce, science, and consumer protection. The shift to a DNI-led structure is not a superficial change in reporting lines. It means that AI capability assessments, model safety reviews, and incident reports will increasingly flow through intelligence channels that operate under classification rules, need-to-know access restrictions, and executive privilege protections. A safety incident at a frontier AI lab that under the old regime might have triggered a public FTC investigation could under the SIF regime be classified as a national security matter and handled through closed-door intelligence community proceedings with no public disclosure requirement. The legal architecture of the intelligence community was not designed for consumer product safety. It was designed for state secrets.

Jay Clayton's career arc is the most precise available signal of what the SIF is actually for. As SEC chairman, Clayton did not focus on breaking up large technology companies or imposing new disclosure requirements on AI systems. His signature achievement was updating stock exchange listing standards to make it easier for large technology companies to go public while reducing the information they were required to disclose to ordinary investors. His background is the structuring of capital markets to support the growth of large private technology enterprises and their eventual public monetization. As DNI, he has spent the past year building relationships with the defense technology sector and overseeing intelligence assessments of adversary AI capabilities. The SIF hands him jurisdiction over how US AI companies are structured, financed, and permitted to scale. That is not regulation of AI. That is strategic investment banking for AI, with the intelligence community as the principal.

The inclusion of Emil Michael, the Pentagon's Undersecretary of War for Research and Engineering, adds a specific operational dimension that has received almost no coverage. Michael spent four years at Uber as head of business development during the company's global expansion phase. That period is best remembered for Uber's strategy of entering markets where its business model was legally ambiguous, building user and driver bases rapidly, and then using the resulting political difficulty of banning a popular service to force regulatory acceptance after the fact. Michael's playbook is known in regulatory circles: move fast, create facts on the ground, and use public support and government relationships to convert illegal expansion into legal operation. His appointment to the SIF suggests that at least part of the task force's mandate involves securing regulatory permission for AI capabilities before formal governance structures are in place to evaluate them, which is precisely how Uber won in city after city.

The Competitive Landscape

China's AI governance operates through two primary agencies: the Cyberspace Administration, which handles content and platform regulation, and the Ministry of Science and Technology, which coordinates national AI research priorities and funding. Both are civilian regulatory bodies with economic development mandates. The US SIF routes through the intelligence and defense apparatus instead, and the competitive implications are substantive. China's AI governance optimizes for commercial deployment at scale, as demonstrated by the rapid rollout of AI services inside China's consumer technology ecosystem. America's SIF structure optimizes for strategic advantage, with defense applications and intelligence exploitation of AI capabilities as the primary lens. This creates an asymmetry: China gets a consumer AI ecosystem that produces massive training data and real-world deployment feedback; America gets a defense AI ecosystem with controlled access and strategic superiority in classified applications. Whether that tradeoff is favorable depends on which dimension of AI competition proves more consequential over the next decade.

The European Union now occupies an increasingly isolated position in the global AI governance landscape. The EU AI Act, which entered full enforcement in August 2026, regulates AI as a product with consumer rights attached. Its risk-tiered framework requires transparency disclosures, bias audits, human oversight mechanisms, and restrictions on high-risk applications including many of the agentic AI capabilities that American and Chinese labs are racing to deploy. Several frontier AI labs have already disclosed that their most capable models may not be available simultaneously in European and non-European markets, citing compliance costs. The SIF's launch crystallizes a structural divergence: America now has an intelligence-backed mandate to deploy the most capable AI systems as fast as possible, while Europe has a regulatory mandate to slow deployment until safety requirements are met. That gap does not close. It widens every quarter that both regimes remain in force, and the AI companies doing business in both jurisdictions will increasingly face irreconcilable compliance obligations.

The historical parallel that most accurately captures the SIF's architecture is not the internet's early deregulation. It is the Atomic Energy Act of 1946, which placed America's nuclear program under civilian oversight at the newly created Atomic Energy Commission while maintaining the classified operational structure that had been built during the Manhattan Project. The AEC's civilian mandate lasted less than a decade before the National Security Act amendments of the 1950s began progressively reclassifying nuclear information and subordinating civilian oversight to defense requirements. By the time the Cold War reached its peak, nuclear governance in the United States was so thoroughly entangled with national security classification that public accountability for nuclear decisions was functionally impossible. The SIF begins with a civilian advisory layer, Condoleezza Rice and David Sacks, and a mixed civilian-intelligence leadership structure. Students of nuclear governance history will recognize the template.

Hidden Insight: The SIF Is Building AI's CFIUS

The most consequential word in the entire executive order is not "super" or "intelligence" or even "force." It is "coordinate." CFIUS, the Committee on Foreign Investment in the United States, does not build anything, regulate anything directly, or enforce any statute. It coordinates government review of transactions involving foreign access to US assets and technology. In its first two decades, CFIUS was a largely ceremonial body that cleared most transactions with minimal scrutiny. Beginning in 2018, it became the single most powerful tool in US industrial policy, blocking Chinese acquisitions of US semiconductor firms, denying export licenses, forcing divestiture of American technology assets from Chinese investors, and effectively rewriting the global semiconductor supply chain without passing a single new law. CFIUS accumulated that power not through legislation but through the gradual expansion of its coordination mandate across more agencies and more transaction types. The SIF is structured to accumulate power the same way: through coordination authority that expands incrementally and becomes, over years, the de facto gatekeeper for American AI development.

The bear case for effective governance from the SIF is straightforward. Critics argue, and the evidence supports the argument, that a 120-day report requirement is a well-established Washington mechanism for converting urgent issues into procedural delay. One hundred and twenty days past October 4 is February 1, 2027, well into the post-midterm political calendar and past the window where most executive branch initiatives achieve their maximum momentum. The absence of any AI safety officer, AI ethics expert, or consumer protection specialist from the formal membership of the task force suggests that the 120-day report will be oriented toward strategic competitive positioning rather than genuine risk assessment. Andrew Ferguson's FTC is the only consumer-facing agency represented in the leadership, and Ferguson has been focused primarily on antitrust cases against technology platforms rather than AI safety regulation. Skeptics point out that a task force with no technical AI expertise in its leadership structure, a classified operational channel, and a 120-day timeline is more likely to produce a document that protects industry from liability than one that protects the public from AI risk.

The deepest insight here is about what the executive order doesn't define. It uses the phrase "Super Intelligence Companies" multiple times without defining it, and the 120-day report is explicitly tasked with working through what the federal government's responsibilities are toward these entities. That ambiguity is not an oversight. Undefined regulatory categories are the most powerful regulatory tools available to executive agencies, because they create uncertainty that forces companies to preemptively cooperate with the agency that will eventually define them. Every major AI lab in America is now calculating what threshold of capability would qualify them as a Super Intelligence Company, what new obligations that status would carry, and whether those obligations are preferable to being excluded from the category and losing access to the government contracts, infrastructure support, and strategic protection that the order implies the category will provide. The definitional fight that plays out over the next four months will set the terms of American AI governance for longer than any explicit regulation could.

What Emil Michael brings to this process is not just a regulatory playbook. It is a specific understanding of how to use the gap between what a government says it will do and what it can actually enforce. Uber's global expansion worked because the company correctly calculated that enforcement of local transportation regulations against a service with millions of users was politically costly enough that regulators would eventually accommodate the service rather than eliminate it. The analogous insight for AI is that frontier AI capabilities deployed at sufficient scale, with sufficient user adoption, and embedded in sufficient critical infrastructure, become politically and practically difficult to restrict retroactively. The SIF's coordination mandate gives the executive branch the ability to support AI deployment at speed while formal regulatory frameworks are still being developed, creating the same facts-on-the-ground dynamic that made Uber's strategy so effective. The question is whether the downstream consequences of AI deployed faster than governance can evaluate it are as benign as Uber's ride-hailing business proved to be.

What to Watch Next

The 120-day deadline runs out in early February 2027. Before that report lands, two leading indicators will tell you what's in it. The first is any formal FTC action targeting AI market structure. Andrew Ferguson has been explicitly skeptical of large AI labs' claims that voluntary safety commitments are adequate governance substitutes, and his formal inclusion in SIF leadership gives him both institutional cover and cross-agency support for antitrust moves that have stalled under previous leadership. If Ferguson announces investigations or consent decrees against major AI companies for market concentration before February 2027, the SIF's report will reflect a more interventionist stance than its intelligence-community leadership suggests. The second indicator is whether any AI safety incident is handled under classification rather than public disclosure: the first time a frontier AI lab's model escape or dangerous capability evaluation is officially classified as a national security matter, the SIF's transformation into an intelligence oversight body rather than an industrial policy body will be confirmed. Watch for either event in the next 60 to 90 days.

Over the 90-day horizon, the definitional fight over "Super Intelligence Companies" will be the most consequential lobbying battle in Washington. The threshold matters enormously. If the definition applies only to systems that have achieved AGI by some formal benchmark, most current AI labs escape formal SIF oversight and the task force becomes a theoretical exercise. If the definition applies to any system capable of autonomous goal pursuit in open-ended environments, every major AI company including those building agentic AI products currently in market becomes subject to new reporting and oversight obligations. OpenAI, Anthropic, Google DeepMind, Meta AI, and xAI will all lobby for different threshold definitions, each designed to capture their competitors while excluding themselves. The outcome will be visible in the February 2027 report, but the real negotiation happens in the 90 days of private briefings, legislative outreach, and closed-door SIF consultations that precede that report's publication.

At the 180-day horizon and beyond, the most consequential question is whether the SIF model spreads through America's intelligence-sharing relationships. The Five Eyes network, which links the intelligence communities of the United States, United Kingdom, Canada, Australia, and New Zealand, has been used to coordinate export controls, financial sanctions, and technology supply chain restrictions over the past five years. A coordinated SIF-equivalent framework across Five Eyes would give the Western alliance a mechanism for setting unified AI governance standards, denying market access to AI systems that don't comply, and coordinating frontier AI capability disclosures as classified intelligence rather than public safety information. The UK's own AI Safety Institute already shares research data with its US counterpart and has a formal liaison relationship with the intelligence community. Whether those channels expand from safety research coordination to full strategic AI industrial policy coordination is the most important open question in global AI governance today, and the SIF's February 2027 report will contain at least a preliminary answer.

When America puts its spy chief in charge of AI strategy, the message is clear: superintelligence is a weapons program with a civilian veneer, and the intelligence community will decide who builds it and who doesn't.

---