This is AI Frontiers. Today, the AI law that won by doing less. Texas's Responsible AI Governance Act, Traga, has been enforced since the 1st of January 2026. The Attorney General's complaint portal is live. And as of this writing, no enforcement action under the law has been publicly announced. Seven months in, that absence tells you something important about what Texas actually built. Traga was reported as a clampdown. Read the enacted text and you find the opposite. A statute with broad jurisdiction and remarkably narrow obligations. It applies to anyone who does business in Texas or whose AI systems reach Texas residents. But what it actually prohibits is a short list of intentional harms, inciting self-harm, unlawful discrimination, infringing constitutional rights, and producing child sexual abuse material. Every prohibition hinges on intent, not outcome. Disparate impact alone cannot establish the required discrimination. The defense architecture is where the law's real character shows. There is no private right of action. Before any penalty, a firm gets written notice and a 60-day cure period. Substantial compliance with the NIEST AI risk management framework is a recognized offense. Penalties range from $10,000 to $200,000 per violation, real on paper, but the path to them is deliberately narrow. If you read that TRIGA requires risk assessments or impact documentation, you read a merger of two different laws. That machinery belongs to a separate statute, Senate Bill 1964, which governs the state's own use of AI, not the private sector. The significance of Traga is upstream of its text, an earlier proposal borrowed from Colorado and the European approach, high-risk classifications, impact assessments, duties to address foreseeable harms. The enacted version abandoned most of that architecture after months of negotiation and industry lobbying. Then came Executive Order 14365, signed in December 2025, directing the Attorney General to create an AI litigation task force to challenge state AI laws and contemplating federal preemption. Colorado, which had the first comprehensive state AI statute, never saw it take effect. XAI sued Colorado in April 2026, the Department of Justice intervened in support, and by May, Colorado had repealed and reenacted the law, dropping its three heaviest obligations. Texas did not defeat Colorado. The American incentive structure did. For European firms, the asymmetry is not that American companies are unregulated. It is about time. A U.S. first company defers European compliance costs until it chooses to enter that market. A European firm builds compliance in from the start and carries that cost into a US market that never asked for it. The advantage is not nationality, it is timing, that's AI frontiers.