Frontier AI Executives Now Lobby for Federal Standards

The AI executives who funded Donald Trump's deregulation push are now lobbying Washington for formal rules — because June's export controls and government-vetted model launches delivered the licensing regime they spent two years trying to kill.

What's new: On June 27, POLITICO reported that frontier labs and their lobbyists are developing a "coordinated push for an actual framework" on advanced AI — a reversal few predicted when Silicon Valley billionaires warned Biden-era safety policy would crush U.S. innovation.

• Paul Lekas, head of global public policy at the Software and Information Industry Association, told POLITICO there is "a real need for a formal process" and that the industry wants to avoid releases governed by "an ad hoc process and a one-off license."
• One senior AI executive, granted anonymity, called the result "a de facto European-style licensing regime."
• An OpenAI executive told POLITICO the industry expects the administration to finalize its June 2 executive order and replace the current crackdown with the voluntary vetting framework it originally outlined.

Why it matters: The industry got the president it paid for — and the governance it feared.

• Trump spent his first year blocking state AI laws and signed Executive Order 14409 on June 2, calling for a voluntary 30-day pre-release review and explicitly disclaiming "mandatory governmental licensing, preclearance, or permitting."
• Before agencies could implement it, Washington went further: Commerce imposed export controls on Anthropic's Fable 5 and Mythos 5 on June 12, and the White House limited OpenAI's GPT-5.6 Sol debut on June 26 to roughly 20 government-vetted partners.
• On June 26, Commerce partially restored Mythos for roughly 100 approved organizations. Fable 5 stays dark. No public explanation for the split.

The irony: Many executives now say Trump's oversight shift is necessary — they just hate how it runs.

• Saif Khan, who served as senior adviser on critical and emerging technology at Commerce under Biden, told POLITICO the Trump approach is an overreaction born of earlier dismissiveness: "Because there has been some dismissiveness of the risks, there's been no preparatory work, no hiring of experts." He called the result "opaque, almost vibes-based" and warned it amounts to "an almost complete moratorium on new releases" that will "start seriously impacting companies' bottom lines."
• Dean Ball — former Trump White House AI adviser, author of the administration's AI Action Plan, and incoming OpenAI head of strategic futures — told POLITICO the administration's concerns are "100 percent legitimate" but that "they are likely overreacting to these legitimate concerns." He separately told TechCrunch the June 2 order has created a "de facto involuntary licensing regime" when paired with undefined safety standards.
• Ball added he is glad Washington takes AI safety seriously — even if the execution is flawed.

The fear factor: The lobby wants rules — but is afraid to ask for them out loud.

• One AI policy adviser who works with major frontier labs told POLITICO: "It feels like they're walking on eggshells a little bit." Companies worry that pushing too hard for clarity could invite export controls or other retaliation.
• That paralysis matters weeks before OpenAI and Anthropic price trillion-dollar IPOs against a regulatory regime that can kill flagship products overnight.

What triggered the whiplash:

• Fortune and other outlets reported that Amazon CEO Andy Jassy raised jailbreak concerns about Anthropic's models with Treasury Secretary Scott Bessent and other officials — reportedly after Amazon researchers stress-tested Fable 5 days after its June 9 launch. Amazon declined to confirm specifics, saying only it is "not uncommon for governments to seek our counsel on potential security risks."
• Anthropic said the flagged jailbreak surfaced "a small number of previously known, minor vulnerabilities" and that applying the same standard industry-wide would "essentially halt all new model deployments for all frontier model providers."
• Trump AI adviser David Sacks told CNN: "It's difficult to fathom how they could claim a jailbreak allowing operability of a cyber weapon could be defined as not 'serious.'" — the same dispute that left Fable offline while Mythos returned to a guest list.

What the White House said:

• Spokesperson Liz Huston defended the record, citing fast-tracked AI infrastructure permits and the June 2 order aimed at stopping state-level regulation: "President Trump has clearly and repeatedly articulated his goal: ensure continued American dominance in AI."

What to watch:

• Early August: The EO's deadline for NSA's classified cyber-capability benchmark and Commerce's voluntary framework — or another month of guest-list governance.
• Lekas's coordinated push: Whether labs can agree on a standardized safety approach before Congress writes one for them.
• IPO disclosures: Whether OpenAI and Anthropic can price public listings against a regime where the legal basis for model shutdowns still isn't published.

Convina's view: Deregulation was the product. Licensing by letter is the invoice. Silicon Valley spent two years and tens of millions telling Washington that written AI rules would kill American innovation — then discovered that no rules means any cabinet secretary with an export pen can turn off your flagship model on a Friday night. The industry isn't asking for safety regulation anymore. It's asking for predictability — which is what grown-up markets require before they underwrite the next trillion-dollar float.