May 5, 2026

Welcome to The Hidden Layer. I’m Ian Krietzberg, fresh off watching Sam Altman’s P.R. pivot in real time on Twitter over the weekend. Just a few months ago, Altman was claiming that “whole classes of jobs” would be “going away.” Then, OpenAI bought TBPN and adopted a more positive messaging strategy; now Altman wants to “build tools to augment and elevate people, not entities to replace them.” (If you missed it, I interviewed Chris Lehane about Sam’s not-a-pivot pivot last week…)

Anyway, today we’re digging into the Elon Musk–shaped legal conflict that has enveloped OpenAI, and what’s really at stake. Could the company survive if a judge nixes an I.P.O.? Plus, news and notes on the Pentagon’s A.I. buddies and the Oscars’ new rules of the road. And, perhaps more importantly: Go Knicks.

Also mentioned in this issue: Greg Brockman, Alnoor Ebrahim, Katie Stanton, Trump, Jack Altman, Gary Marcus, Ellen Aprill, Ashwini Jayaratnam, and more…

Let’s get into it…

Three Things You Should Know…

A.I.’s crash test dummies: Earlier today, Common Sense Media launched its Youth A.I. Safety Institute, which is focused on ramping up independent testing of A.I. products for children’s use to safeguard the “development of a generation growing up with A.I.” For the past three years, the organization has been studying the harms associated with the steady proliferation of artificial intelligence; the Youth A.I. Safety Institute, Common Sense founder and C.E.O. James Steyer told me, has been in the works for the past six months. “The companies are moving so quickly, and yet the federal government is doing nothing,” he said. “We felt there was a desperate need. Someone’s got to do it. Someone’s got to do a good job protecting kids and families in school.”

The institute, which is modeled after the vehicle crash-test programs that enabled car safety innovations in the 20th century, will begin by red-teaming and stress-testing models “under real-world, adversarial, multi-turn, and youth-specific scenarios,” according to a statement. The institute’s backers include the OpenAI Foundation and Anthropic, in addition to a long list of philanthropic funders. Asked about the project’s scale, Steyer said that “the initial incremental budget increase will be $20 million in year one.” He expects it to “grow from there.”

A MESSAGE FROM OUR SPONSOR

Retailers can now process millions of transactions in minutes—not hours. Toshiba Tec partnered with McKinsey using Nvidia accelerated computing to enable real-time recommendations, faster promotion testing, and measurable lifts in sales, profit and long-term customer value.

Read the case study.

More Pentagon A.I. deals: On Friday, the Department of Defense entered into a series of agreements with frontier A.I. companies SpaceX, OpenAI, Google, Nvidia, Reflection, Microsoft, Amazon Web Services, and Oracle to deploy their technology on the Pentagon’s classified network. This follows Anthropic’s refusal to allow the government to use its technology for “all lawful purposes,” and comes despite an enormous amount of backlash, including from within several of the companies themselves. While D.O.D. said in a statement that the technology will help soldiers make decisions “in complex operational environments,” among other benefits, safety engineers have been warning for years that there’s not enough evidence the tech is fit for military use, highlighting the risks to civilians and soldiers alike.

To wit: In the past five months alone, per D.O.D., the department’s official A.I. platform has deployed “hundreds of thousands of agents,” cutting the time to complete unspecified “tasks” from “months to days.” The Pentagon, it seems, has adopted Silicon Valley’s famous “move fast” mentality. Let’s just hope they’re not breaking things, too.
The Oscars crack down on A.I.: If you missed it, the Academy just issued a few substantive rule changes for the 99th Oscars, some of which deal implicitly, or explicitly, with A.I. tools. Acting awards can now go only to “roles credited in the film’s legal billing and demonstrably performed by humans with their consent,” and screenplays “must be human-authored to be eligible” for writing awards. And just to make sure no one sneaks a script in under that wonderful French mononym Claude, the Academy “reserves the right to request more information about the nature of the use and human authorship.”

The Recording Academy did something similar with the Grammy Awards last year, adding rules that only human creators are eligible for an award. The Grammys did allow for generative A.I. to be used so long as the human contributed more than the A.I.—surely a tough issue to adjudicate these days.

Hallucination of the Week: RoboPods

Half of all new podcasts are being made without human hosts, according to The Podcast Index. Bloomberg reported that over a recent nine-day period, roughly 11,000 new podcast feeds were created… and nearly 40 percent of them were A.I.-generated. Is anyone actually listening to this stuff?

Capital Intelligence

Aidoc, a medical A.I. platform, raised $150 million last week, bringing the company’s total funding to more than $500 million. General Catalyst, SoftBank, and Nvidia (through its V.C. arm, NVentures) led the Series E round.
Financial A.I. platform Rogo raised $160 million in Series D funding last week, bringing its total funding to over $300 million. The round included participation from a few of the major Valley V.C.s—Sequoia, Thrive Capital, and Khosla Ventures—in addition to Jack Altman, Sam Altman’s younger brother.
Scout, an A.I. company aiming to enable “unmanned warfare,” closed a $100 million Series A round last week. Scout described the financing as “the largest defense-tech Series A in U.S. history.” The 18-month-old company has already booked $11 million in contracts with the Department of Defense.

And now for the main event…

What If Elon Wins the OpenAI Trial?

As the messy Musk v. Altman trial enters week two, the A.I. industry has yet to grapple with the improbable but earth-shattering consequences if the jury actually sides with Elon. With an I.P.O. off the table, would OpenAI’s financial house of cards fall apart?

Ian Krietzberg

While media reporters and industry insiders are gabbing about the irresistible drama of Musk v. Altman—including OpenAI president Greg Brockman’s testimony yesterday that he’s worth nearly $30 billion, and Musk referring to himself as a “fool” for funding the startup—few have really grappled with the possibility, however remote, that Musk could actually win. Yes, most of his claims have been tossed, narrowed, or dismissed (some at Musk’s own insistence). But three of his core demands will be considered: $134 billion in damages, the removal of Altman from the board of directors, and a requirement that OpenAI return to its original nonprofit structure.

That last demand, obviously, would have the most seismic impact. OpenAI has raised more than $100 billion since ditching its initial nonprofit model, making it the most heavily capitalized private company in history. And since the business is likely years away from profitability given the preposterous costs of A.I. compute, a public offering remains essential. After all, much of the money that OpenAI has raised over the past two years was predicated on its conversion into a for-profit entity; not only would Elon’s preferred remedies complicate the company’s ability to raise funds—nonprofits can’t I.P.O.—but they might also jeopardize earlier deals.

A MESSAGE FROM OUR SPONSOR

Read the case study.

Notwithstanding Musk’s moral posturing—in testimony last week, he accused Altman and Brockman of conspiring to “steal a charity”—a victory would enable “xAI or some of the others to increase their market share,” said Dr. Alnoor Ebrahim, a Tufts University professor and expert on N.G.O.s. But if Musk loses, Ebrahim added, “that’s also a bad outcome—because it will send a strong signal that nonprofit charities can be converted to for-profit businesses at potentially a substantial loss to the public.”

“The Most Hated Men in America”

It now seems quaint, but there was a palpable idealism when Musk, Brockman, Altman, and others joined hands to co-found OpenAI in 2015. “Our goal is to advance digital intelligence in the way that is most likely to benefit humanity as a whole, unconstrained by a need to generate financial return,” the company wrote in its first-ever blog post. “Since our research is free from financial obligations, we can better focus on a positive human impact.” Musk was a key reason for this financial freedom: He was the nonprofit’s largest donor, contributing roughly $40 million of his $1 billion pledge.

But after a few years, everyone involved came to the conclusion that they’d need a lot more money to keep going. In 2018, Musk suggested a merger with Tesla, then left the board after the others demurred; a year later, OpenAI created a capped for-profit subsidiary to help fill the funding void created by Musk’s departure, and Microsoft soon invested its first $1 billion tranche. (Microsoft is named as a defendant in the lawsuit.) From there, OpenAI released ChatGPT and went on a fundraising spree. Last year, as Elon was busy scaling xAI and merging it with X into SpaceX, Altman completed the conversion of that aforementioned subsidiary into a for-profit public benefit corporation, something the company said was essential for its endurance. (The nonprofit arm owns a 26 percent stake in the for-profit entity.) Now OpenAI is approaching a trillion-dollar valuation.

Musk, who launched xAI in 2023, wasn’t thrilled about how things played out. In 2024, he filed suit, claiming he was “deceived” and “assiduously manipulated” by Altman and Brockman into funding a charity the two men later “stole.” OpenAI, for its part, has consistently dismissed Musk’s case as a “baseless and jealous bid to derail a competitor.” Indeed, two days before the trial began, Musk texted Brockman to see whether he might be interested in pursuing a settlement. “When Mr. Brockman responded with a suggestion that both sides drop their respective claims,” according to a Monday filing, “Mr. Musk shot back: ‘By the end of this week, you and Sam will be the most hated men in America. If you insist, so it will be.’”

Of course, it’s easy to view the trial as “a battle of the egos,” the A.I. scientist Gary Marcus wrote. But, he added, the trial is also “about whether OpenAI should be held to its promises to be a nonprofit working for the benefit of humanity, which it clearly no longer is.” Ebrahim, the N.G.O. expert, agreed that a Musk victory wouldn’t be the worst outcome. “I do think there is a reasonable case to be made that the public interest has not been adequately protected in OpenAI’s conversion,” he said.

The Classic Lose-Lose

Yet a win for Musk is still unlikely, according to the scholars and litigators I spoke to. For one thing, the attorneys general of the two states with jurisdiction—Delaware, where OpenAI is incorporated, and California, where it’s headquartered—have signed off on its conversion into a for-profit company. “My own view is that Musk is not the proper person to assert breach of charitable trust,” said UCLA legal scholar Ellen Aprill, who specializes in nonprofits. “The involvement of two A.G.s cuts against his claims. A.G.s have primary responsibility for use of charitable funds.” But as the litigator Ashwini Jayaratnam of DarrowEverett pointed out, “Juries are tricky. You don’t know what they’re going to do.” (She thinks Musk has about a 30 percent chance of pulling out a victory, though even then, the fallout hinges entirely on what kind of remedies the judge deems fair.)

Either way, the implications will be vast—and for Ebrahim, society loses regardless of who wins. “We can’t trust Musk,” he said. “We can’t trust Altman. We can’t trust any individual regardless of their personality, and so we’re left to rely on government.” That’s why the state A.G.s are so crucial to the case, and why several scholars I spoke with are worried about setting a precedent that will undercut their role. “My eyes are not so much on Musk or Altman,” Ebrahim said, “but on the California A.G. and legislature, in terms of what they decide to do.”

That’s all for today. I’ll see you on Thursday.

Ian

Join Emmy Award-winning journalist Peter Hamby, along with the team of expert journalists at Puck, as they let you in on the conversations insiders are having across the four corners of power in America: Wall Street, Washington, Silicon Valley, and Hollywood. Presented in partnership with Audacy, new episodes publish daily, Monday through Friday.

An essential, insider-friendly Hollywood tip sheet from Matthew Belloni, who spent 14 years in the trenches at The Hollywood Reporter and five before that practicing entertainment law. What I’m Hearing also features veteran Hollywood journalist Kim Masters, as well as a special companion email from Eriq Gardner, focused on entertainment law, and weekly box office analysis from Scott Mendelson.

Need help? Review our FAQ page or contact us for assistance. For brand partnerships, email ads@puck.news.

You received this email because you signed up to receive emails from Puck, or as part of your Puck account associated with {{customer.email}}. To stop receiving this newsletter and/or manage all your email preferences, click here.

Puck is published by Heat Media LLC. 107 Greenwich St., New York, NY 10006