The David Sacks AI Scandal Hiding in Plain Sight
The AI Czar and his firm still have hundreds of investments in AI-related software and hardware companies, according to an analysis of his public disclosures.
Up until a few weeks ago, venture capitalist David Sacks was having a rollicking good year. He’s become legitimately famous, which is a remarkable achievement for a mean-spirited man with a barren personality. The “All-In” podcast—where Sacks speaks with a marble-mouthed disdain for his co-hosts and everyone else—is reportedly garnering 6 million downloads per month. On X, Sacks has amassed more than 1.4 million followers. Best of all, Sacks has been able to keep his job atop the firm Craft Ventures while also serving as President Trump’s special advisor on AI and crypto.
Sacks has long claimed that his position with the White House is above-board because he and Craft sold most of their crypto investments and divested from AI-oriented companies. The sourcing for this claim is David Sacks on his self-reported financial disclosures, followed by an assessment from White House counsel. Sacks’s “trust me, bro” line of defense has never sufficed, but it’s hard for any one Trump-related enrichment scandal to break through the noise, since there are so many happening simultaneously.
Sacks’s hubris might be catching up to him, though.
Last week, a bunch of outlets (including Hard Reset and our newest contributor Jacob Ward!) reported on a leaked draft of an executive order intended to stave off states’ regulation of AI. Preempting the states on AI laws is a deeply unpopular concept across the political aisle, which is presumably why the draft leaked and has since been smothered, rather than enacted. Another reason the draft leaked is that it would’ve given even more power to Sacks, who seems to have pissed off his right-wing allies. That’s the educated guess of Tina Nguyen, senior reporter at The Verge, and I agree with her. From Nguyen’s analysis:
“[On] that rare occasion that an actual document leaks out of the [Trump] administration, it’s a sign that someone has carefully weighed the risks of undermining their enemy, considered the cost of doing so, and decided they hated them enough to do it anyway.”
Sacks drew more unwanted attention when he made the peculiar decision to weigh in on the debate over whether there is an “AI bubble.” His contribution came in the form of a November 24 post on X, where he wrote, “According to today’s WSJ, AI-related investment accounts for half of GDP growth. A reversal would risk recession. We can’t afford to go backwards.” This raised a number of follow-up questions: What else does America’s AI czar know about the strength of the economy? Why is Sacks framing AI investment like we’re all part of a prisoner’s dilemma? And who is Sacks referring to when he says we can’t afford to go backwards?
Sacks later tacked on an unhelpful addendum: “Puzzled that anyone could interpret this post as supporting a bailout,” he wrote with his usual, smarmy flair. “I’ve already opposed that. Nor do I believe one is needed.”
Some free PR advice for prominent tech investors and politicians: If you trapeze around with a “kick me” sticky note on your forehead, an editor at one of the remaining media institutions will eventually notice and say, hey, what’s up with that guy? That’s how we got Sunday’s five-bylined New York Times piece about Sacks, which according to the feature subject himself has been in the works for a while.
I believe it! Despite Sacks’s protestations (more on that in a second), there are lots of juicy, newsworthy tidbits in the Times report. Entitled “Silicon Valley’s Man in the White House Is Benefiting Himself and His Friends,” the story details Sacks’s increasingly close relationship with Nvidia CEO Jensen Huang, who is probably the most important figure in the entire AI industry. It also explores how Sacks’s “government work has boosted the profile of the [‘All-In’] podcast,” and alleges that Sacks lobbied the Trump administration to eliminate Biden-era restrictions on the export of Nvidia’s AI chips to other countries. On the latter point, the Times repeatedly raises “national security” concerns—the idea/objection being that the Nvidia “components could bolster China’s economy and military.”
I find those concerns to be overwrought saber-rattling, and they unfortunately distract from important research that the Times unearthed. Though the White House says Sacks self-reported the sale of “over 99 percent” of “holdings in companies that could potentially raise a conflict of interest concern,” Sacks and Craft still hold “20 crypto and 449 AI-related investments,” according to a Times examination of Sacks’s public ethics filings.
More, from their report: “Of the AI-related investments, 11 were designated in one waiver as ‘AI Interests.’ The other 438 were classified as software or hardware makers, even though they promote AI offerings or services on their websites.”
The Times cites a number of hilarious examples. For instance: Sacks’s ethics filings list Palantir as a software service, as opposed to an AI company he might need to divest from, despite Palantir marketing its “AI-Powered Automation for Every Decision.” Of the 438 AI-related companies deemed as providing software or hardware services, 41 literally have “AI” in their names, the Times found.
In other words, while Sacks is playing an influential role shaping AI policy, he and his firm are invested in dozens and dozens and dozens of “software” and “hardware” companies that, at a minimum, are likely to integrate AI into their services (if they haven’t already).
David Warrington, counsel to the president, attempted to head off this obvious tension in the waiver he granted Sacks, which allows Sacks and Craft to retain hundreds of investments. I’m going to quote his explanation in full, because it’s unbelievably shallow:
“You have also disclosed that both you and the Craft Funds have interests in various Software-as-a-Service companies (the ‘SaaS Interests’). These SaaS companies currently use artificial-intelligence-related applications acquired from foundation model companies, or integrate with foundation models through an API. Many of the SaaS Interests do not currently use Al-related applications in their core business in any material way, but by their nature many of them are likely to at some point in the future. Your direct SaaS Interests are listed in the attached Schedule C; Craft’s SaaS Interests are listed in the attached Schedule D. Combined, the SaaS Interests amount to just over 13% of your total assets.”
Warrington’s assessment is the real scandal hiding in plain sight. It’s a classic example of how financial disclosure forms—even ones that have been diluted or obscured—are often more valuable than direct quotes or expert testimony. I’m not clear on who originally designated the 438 AI-related investments as software and hardware companies—was it Sacks? His own counsel? White House counsel? A combination? Regardless, the takeaway is the same. Sacks can tout his divestments all day every day, but hundreds of AI-related gaps remain unaddressed.
You can go through the list of Sacks’s/his firm’s investments for yourself, and make your own tally of which software and hardware companies are related to AI. Your number might differ from the analysis done by the Times, but rest assured it will be much higher than zero.
Sacks responded to the Times report on Sunday night via an X post and an attached letter from a law firm he’s apparently hired to signal he might sue. I have given up on predicting when right-wing politicians and their allies are actually going to file lawsuits, and when they might extract concessions from the parent companies of media outlets; Trump and feckless executives at CBS and ABC ruined all of that.
I will say that the threatening letter Sacks posted on X reads like a series of grievances about the normal fact-checking process. The law firm representing Sacks alleges that the Times sent a number of questions and fact-checking inquiries, and then pivoted away from some of those inquiries after receiving replies from Sacks’s camp. That is how fact-checking works. Sometimes you are presented with information, so you take it to the source for comment, and they have the opportunity to counter or even disprove the information. My editors and I have killed stories after hearing back from reporting subjects. I have drastically altered the substance of stories after hearing back from reporting subjects. It’s a normal part of the process, not an admission of wrongdoing or malice.
What was ultimately published by the Times—mainly the section about the hundreds of AI-related software and hardware companies that Sacks and Craft are still invested in—is obviously newsworthy. And wouldn’t you know it, that section isn’t mentioned in the whiny letter posted by Sacks and his law firm.
Here’s what else we’re reading this week:
Slate’s Mark Joseph Stern discovered that Wisconsin Supreme Court Justice Annette Kingsland Ziegler cited a fake quote in her dissent about whether the state should potentially redraw its right-wing gerrymandered maps. Ziegler is functionally a conservative politician (she’s an elected official serving her third 10-year term), and she’s become increasingly antagonistic towards her more liberal colleagues as the Wisconsin Supreme Court has shifted leftward to a 4-3 liberal majority. The made-up quote Ziegler cited was supposed to be from a recent U.S. Supreme Court decision; she claimed the quote supported her dissenting opinion. How that quote ended up in her dissent is an open question—was AI hallucination afoot? I do not know!
OpenAI has operated with impunity around copyright issues, but lawsuits against the company are really starting to pile up. The Hollywood Reporter has an interesting update on one such case, where a district judge ruled that “OpenAI must hand over documents revealing the company’s motivations for deleting” two datasets of pirated books. The decision means OpenAI’s own in-house legal team can be deposed, according to The Hollywood Reporter. I am not a lawyer, but let me tell you: when your legal team gets deposed, it’s not good, folks.
More than 1,000 Amazon workers anonymously signed a letter warning that the company’s growing investment in data centers and AI systems could soon prove to be catastrophic for the world. “The current generation of AI has become almost like a drug that companies like Amazon obsess over, use as a cover to lay people off, and use the savings to pay for data centers for AI products no one is paying for,” one Amazon employee told Wired, which has more on the letter.
Two reports from the Wall Street Journal and Reuters about Anduril—the weapons firm run by Palmer Luckey, the guy who dresses like an extra on Tim Robinson’s I Think You Should Leave—allege serious safety issues and setbacks with the company’s autonomous offerings.