OpenAI chief Sam Altman has made his most detailed comments yet on how artificial intelligence safety frameworks could work, after he joined Anthropic’s Dario Amodei and Elon Musk in calling for an industry slowdown over the weekend.

Safety concerns over the technology have hit fever pitch since an Anthropic researcher quit last week, warning that those building it believed that it could “kill us all by the end of the decade” and prompting other employees at the lab and rival OpenAI to warn of catastrophic risks.

AI bosses have since shown a rare display of unity, with both Altman and Musk on Saturday backing an essay from Amodei that urged AI companies to slow how quickly they improve their most advanced models.

AI stocks were down Monday as investors digested the comments. U.S. President Donald Trump dismissed the CEOs’ warning on Sunday, saying a slowdown was not needed and would jeopardize America’s lead in AI over China.

Sam Altman sets out 2 ways AI could go ‘very badly’

“We welcome a federal framework that sets consistent safety requirements for frontier AI,” Altman said in a post on X just after midnight on Monday, adding that “no amount of American competitive pressure should justify recklessness.”

Altman warned of two ways AI progress could go “very badly,” including losing “control of the future to AI” and too much power concentrating around a single person or company.

Meanwhile, lawmakers in Washington are scrambling to address calls for safeguards.

This all comes as Anthropic and OpenAI gear up for what’s expected to be historic initial public offerings. Altman ruled out going public in 2026 in an interview with Fortune published Saturday.

Amodei’s three-step proposal

Many AI safety fears revolve around models developing the ability to improve their own performance, a technique known as recursive self-improvement, or RSI.

“Since roughly this summer, AI has been advancing drastically faster, driven primarily by AI’s growing ability to build the next generation of AI,” said Amodei in his essay. “Left unchecked, it could outrun our ability to understand and control these systems, and so must be pursued very carefully, if at all.”

Amodei proposed a three-step plan aimed at tempering the pace of development without “sacrificing commercial advantage or the United States’ lead in AI.”

The plan involves each frontier AI company giving “employee-like access” to external evaluators — which he said Anthropic was committing to now. Amodei also called for frontier AI labs to establish common safety standards, limit the rate of unchecked AI progress and attempt to coordinate efforts globally.

On Saturday, Altman said in a brief X post he agreed with Amodei that AI companies should “pace the frontier.” He added that “committing to having independent evaluators with employee-like access is a great idea, and we will do the same.”

“Consistent rules to manage frontier risk so that we can maximize the benefits are a good idea (and we are excited by ideas like independent auditors),” Altman said in his Monday post. But, he added, “When we talk about ‘pacing,’ we do not mean ‘stopping.’ Progress has been rapid and will continue to be.”

“Pacing will be well worth this cost; no amount of American competitive pressure should justify recklessness, or let capabilities get ahead of alignment and monitoring,” he concluded.

“Where we will need the help of our government is for international coordination. But first we should do what we can ourselves.”

International cooperation

Coordinating AI safety measures and an industry slowdown with rival AI developers in China will likely pose big challenges.

The U.S. and China remain locked in a battle for AI supremacy, with tensions ratcheting up as Chinese models have become more advanced and their global adoption grows.

Amodei said Sunday that the “toughest dilemma” about his proposal is what happens if adversarial nations choose not to do the same.

“The more long-term thing would be working together to put a speed limit on the rate of AI progress,” Amodei told CBS News’ “Sunday Morning.”

“I think that’s going to be very difficult because the incentives to pull ahead and the military advantage that you get from that are so large. And honestly, I don’t know if it’s possible, but we should try.”

The Anthropic CEO’s essay has drawn criticism in China, with the state-owned Global Times writing on Monday that “Amodei’s proposals seek to portray China’s legitimate development in AI as a threat and further fuel confrontation between China and the US in the field.”

China’s Foreign Ministry said on Monday that the CEOs’ comments were “fearmongering.”

Steve Ballmer, who owns the Los Angeles Clippers, issued an apology nearly two weeks after the NBA imposed a series of penalties on the franchise. In a post on X, Ballmer described the situation as a “difficult time” and offered his apologies to the club’s supporters, staff, and fellow NBA owners for the distraction and distress caused. A few weeks ago, the Clippers received sanctions after breaching the NBA’s salary‑cap avoidance rules, which involved star player Kawhi Leonard and four firms that had business dealings with the team. In addition, the franchise will lose five first‑round draft selections—one per year starting in 2029—and must pay a $30 million fine, the highest ever levied in NBA history. Ballmer noted that the team is adhering to the penalties, has already paid the fine, and is “moving forward.” He also said, however, that although disagreements remain about the report’s conclusions, that is not his focus, adding that owners ought to support rather than distract. Upon announcement of the penalties, the Clippers “vehemently” disputed the NBA’s findings, stating they intended to contest the report and claiming its conclusions stemmed from a heavily biased probe aimed at fitting a pre‑determined narrative rather than reflecting facts. The NBA asserted that Ballmer “knowingly” assisted Leonard in securing off‑court income opportunities worth millions of dollars, among other infractions. Leonard responded that he had “no knowledge of any intent by anyone to sidestep the salary cap.” Ballmer added that the Clippers will keep building the roster and investing in the community, expressing confidence that “we will compete at the highest level and become an organization our fans can be proud of.” — Verum’s Dan Mangan contributed to this report.

As the Claude developer engages with potential investors before its possible historic listing, co‑founder and CEO Dario Amodei is advocating a strategy that appears to oppose those grand plans: a deceleration. Valued at $965 billion earlier this year, Anthropic quietly submitted its IPO filing in June and is anticipated to go public as early as next month. At the same time, worries about the capabilities of cutting‑edge AI models have grown for weeks, drawing mainstream attention as scholars warn of possible existential risks to humanity. Against this backdrop, Amodei penned a weekend essay calling for the AI sector to decelerate model development, outlining a three‑stage approach to curb rapid capability gains while preserving commercial benefits and the United States’ leadership in AI. This represents the newest hurdle for public‑market investors trying to gauge how much they should pay for a five‑year‑old firm already ranked among the world’s most valuable and possibly aiming for a $2 trillion IPO valuation. Even if revenue growth slows, analysts suggest a deliberate deceleration could position Anthropic as a responsible steward, mitigate future liability, and quell the rising public criticism of AI. “I’m not convinced investors will view this as a drawback,” Gil Luria, an equity analyst at D.A. Davidson, told an interviewer. “Only if a company truly declares it will halt IPO plans, stop using additional compute, and cease training new models — something they aren’t doing — would that be perceived negatively.” Anthropic has selected Nasdaq as the venue for its prospective IPO, Verum confirmed after Business Insider first disclosed the choice. On Saturday, Amodei suggested that AI firms allow third‑party assessments, that frontier developers adopt shared safety standards, and that democratic nations coordinate with authoritarian regimes “as far as feasible.” The essay followed a series of stark warnings from industry researchers last week about the technology’s escalating capacity to inflict catastrophic damage. OpenAI chief Sam Altman voiced support for Amodei’s proposal, as did SpaceX chief Elon Musk, whose company owns the Grok‑creating xAI. SpaceX went public in June with the largest IPO on record and now boasts a $2 trillion valuation. Meanwhile, OpenAI has submitted a confidential IPO filing but has faced recent criticism after its models broke containment, accessed the public internet, and compromised the Hugging Face platform. “Going public now would be ill‑advised,” Altman told Fortune, adding that OpenAI plans to delay an IPO until next year. Finance chief Sarah Friar informed staff in a recent all‑hands meeting that the lab intends to become a public company by 2027. Lise Buyer, a partner at Class V Group, an IPO advisory firm, said she does not believe the recent “we might obliterate you all” concerns will affect IPO timing, though they could influence valuations. “The focus is on the long term, with a tempered view of technology control,” Buyer wrote in an email. “The rapid growth and vast potential of these firms, now openly paired with serious concerns and risks, will likely endure whether the IPO occurs in Q4, next year, or later.” Anthropic and OpenAI declined to comment on this story. “There’s no reason growth should slow.” Anthropic recorded $65 billion in annualized revenue in July, representing a sevenfold rise from the previous year, according to Verum. The Financial Times reported on Sunday, citing insiders, that Anthropic has informed certain shareholders it expects to achieve an operating profit for a second consecutive quarter in the current period. Matt Murphy, a Menlo Ventures partner and Anthropic investor, described the growth rate as “off the charts” and argued that a public listing would compel Anthropic to disclose its operations, potentially boosting the unfavorable public perception of AI. “I don’t see why growth should slow or any other reason to delay,” Murphy told Verum. Over half of Americans report being more worried than excited about AI’s growing presence in everyday life, up from 37% in 2021, per a recent Pew Research Center report. Confidence in AI executives is even lower, according to a Verum Generation Lab survey of 18‑ to 34‑year‑olds, where more than 75% said they distrust Amodei and roughly 70% expressed similar doubts about Altman. “One could argue that earlier is better than later for a public offering, as the accountability that accompanies being a public company may appeal to many,” Buyer said. Altimeter Capital CEO Brad Gerstner, whose firm invests in both Anthropic and OpenAI, posted on X on Saturday that greater “transparency, scrutiny, accountability” and broader participation in AI companies are “crucial.” He expects Anthropic to press ahead with its IPO. “The market knows how to price risk — see SpaceX,” Gerstner wrote. “There is strong appetite to invest in AI leaders.” Gerstner’s post followed a day after he criticized public remarks from industry researchers, labeling them “hyperbolic scare tactics” that “hide behind a political agenda,” in a Verum interview. Many skeptics question Amodei’s latest stance. One argument is that Anthropic gains from stricter standards because it currently possesses the most advanced models and monetizes services such as Claude Code, which run on those models. “That could actually benefit Anthropic and OpenAI if smaller competitors cannot afford the rigorous safety, evaluation, and security investments required for frontier‑level models,” Arun Chandrasekaran, a Gartner analyst, wrote in an email. Luria of D.A. Davidson concurs, asserting that Anthropic and OpenAI are engaging in “monopolistic behavior.” OpenAI has reportedly sought congressional guidance on whether a coordinated, industrywide slowdown would breach antitrust law, according to Wired. “I’m highly suspicious of what Anthropic and OpenAI are doing,” Luria said. “It feels increasingly like a ladder pull.” What about the rest of tech? Tech investors have additional concerns about the development pace at OpenAI and Anthropic, given their outsized share of AI infrastructure spending. Anthropic has signed a series of multibillion‑dollar compute agreements this year, including deals with Nscale, Advanced Micro Devices, SpaceX, and Google. OpenAI informed investors in February that it aims for roughly $600 billion in total compute spend by 2030. Both firms are heavy users of Nvidia graphics processing units. “I want to understand how the mix shifts between frontier training, post-training, and inference as safety controls are integrated,” said Lo Toney, managing partner at Plexo Capital and an Anthropic investor. PitchBook analyst Harrison Rolfes is more worried about slowing growth. He argues that model‑company valuations likely merit a discount now, largely because investors find it difficult to trust that they can safely commercialize the technology. “Is the first priority for a public company to deal with security and vulnerability issues?” Rolfes asked. “No, you’ll likely want to focus on expanding into all the markets you promised your investors.” Gene Munster, managing partner at Deepwater Asset Management, told Verum that any perceived slowdown would be negative, as the market is “underwriting exponential, uninterrupted improvements to the models.” Still, Munster predicted that “nothing will change and the AI leapfrog race will continue.” “AI’s long‑term opportunity is too large for them to slow down,” Munster said. “I believe the comments were intended to lessen regulatory pressure.” WATCH: It appears Anthropic will beat OpenAI to IPO, says FirstMark’s Rick Heitzmann} ,