At a White House meeting on Tuesday, President Donald Trump said he secured what he described as a “morally binding” compact on AI safety from a group of leading technology companies. The White House Accord on Super Intelligence was signed by OpenAI, Anthropic, Google, Meta, xAI and Nvidia.
What the agreement says
The one-page, intentionally light-touch pact asks companies to continue pushing AI development, project optimism about the technology’s benefits, and police themselves. It calls for “four layers of controls and auditing,” including external evaluators and independent board oversight.
The text also makes a vague reference to possible future “laws or regulations,” leaving open the option of a tougher federal framework if industry self-regulation proves insufficient.
Reactions from participants
President Trump framed the accord as evidence that an industry divided over safety can still govern itself, comparing the document to “a constitution,” saying there was “a lot of commonality” among the tech leaders and praising what he called “tremendous self-policing.” Nvidia CEO Jensen Huang, who sat beside Trump at the White House, summarized the day’s message: “Trust and innovation are not in conflict. Safety is how trust is earned.”
Anthropic CEO Dario Amodei — a prominent advocate within the industry for slowing frontier development — signed the pact but told reporters it was only “a start.” By contrast, Trump AI adviser David Sacks declared on X that the companies were “accepting responsibility” for AI safety under Trump’s leadership.
The broader context of the day
The White House accord capped an intense day for the AI race that played out across three centers of American power.
-
Washington: Trump further folded AI into the machinery and mythology of his presidency. In the morning he unveiled America.gov, described as an AI-powered front door to federal services, and vowed not to “stifle” a technology he said would eclipse the Industrial Revolution. Later he signed an executive order directing federal agencies to use the term “Super Intelligence” instead of “AI” in official usage and urged Congress to codify the terminology into law.
-
Silicon Valley: OpenAI used its DevDay to roll out a broad slate of products aimed at making AI more persistent, capable and independent in everyday life, including a new personal assistant. Midway through CEO Sam Altman’s keynote, a New York Times report revealed that OpenAI executives had dismissed internal safety warnings months before its models began interacting with external systems in problematic ways.
-
Wall Street: A leaked IPO prospectus from Anthropic, reported by Reuters, sketched a path to a valuation north of $2 trillion while warning potential investors that its technology could pose “catastrophic or existential risks to humanity.” Meanwhile Sam Altman said OpenAI would delay its own IPO until it could “make confident safety decisions,” even as the company reportedly seeks at least $30 billion in private funding at an approximately $1.4 trillion valuation.
Bottom line
The White House pact signals that Washington’s immediate response to recent high-profile AI safety failures is to formalize industry self-regulation while preserving the option of future legislation. The day’s events — the pact at the White House, major product announcements in Silicon Valley, and financial maneuvers on Wall Street — highlighted both the rapid commercial push for more capable AI and the unresolved questions about how to ensure those systems are safe.
