‘for 2 Years, We’ve Held Our Tongues’: Democrats Clap Back At 1 Of Silicon Valley’s Biggest Donors

A dominant origin story explaining Silicon Valley’s pro-Trump lurch in the 2024 election is getting a belated rebuttal from former President Joe Biden’s ex-aides — a dispute arising as Democrats struggle to define their stance on artificial intelligence.

The crux of the fight is backward-looking: Who said what during a May 2024 White House meeting on AI policy? But it has big implications for this year’s midterms and the 2028 presidential race, including whether the tech billionaires dumping oceans of campaign cash have more to fear from pro-regulation Democrats or from President Donald Trump’s chaotic approach to AI.

In an opinion piece published Thursday by POLITICO Magazine, ex-Biden staffers Bruce Reed and Ben Buchanan contradict venture capitalist Marc Andreessen’s account of that 2024 meeting — a sit-down where Andreessen contends it became clear that the Biden administration intended to impose suffocating regulations and eliminate AI start-ups.

Andreessen, who says AI will “save the world” and champions a laissez-faire approach to regulating it, later called the conversation so “horrifying” that it caused him and his business partner Ben Horowitz to back Trump’s return to the White House. He framed it as a pivotal moment in the unraveling of the tech industry’s ties with the Biden administration and Democrats more broadly.

“And that’s the day we walked out and stood in the parking lot of the West Wing and took one look at each other, and we’re like, ‘Yep, we’re for Trump,’” Andreessen told Ross Douthat on his New York Times’ podcast, in one of multiple interviews the veteran tech investor recorded soon after Trump’s election.

But Reed, Biden’s ex-deputy chief of staff for policy, and Buchanan, a former White House special adviser on AI, write in POLITICO Magazine that Andreessen’s characterization of the meeting is inaccurate.

They say they never told Andreessen and Horowitz that Biden intended to ensure that “there will be no start-ups” in AI, and that only two or three large companies controlled by Washington would be allowed to dominate the industry. They also deny saying the administration would ban so-called open-source or open-weight AI, a lower-cost variant of the technology that allows users greater control and flexibility while raising security risks.

“The experience he describes as essential to his political conversion simply never happened,” Reed and Buchanan write of Andreessen. As the “AI policy debate once again centers on questions of corporate power and government oversight,” they say, “it’s time to set the record straight.”

Andreessen’s venture capitalist firm, Andreessen Horowitz, said in a statement that it “stands by its account of the meeting and rejects the authors’ claims.” An Andreessen Horowitz spokesperson said concerns about the Biden administration’s potentially onerous approach to AI rules were widespread across Silicon Valley at the time, pointing to a November 2023 letter signed by top industry representatives that warned the White House against overregulating open-weight AI models.

2 years too late?

The opinion piece represents the first time that Biden’s former aides have pushed back in detail on Andreessen’s description of the White House meeting.

“For two years, we’ve held our tongues,” write Reed, now the head of AI at the influential kids’ safety group Common Sense Media, and Buchanan, a professor at the Johns Hopkins School of Advanced International Studies. Buchanan is also an adviser to AI companies, including Anthropic.

“We don’t know whether Andreessen has simply misremembered our discussion or purposely misrepresented it,” they add.

Two years later, Democrats appear poised to take back at least one chamber of Congress in the November elections. But they are still trying to chart a coherent path on AI, even as polls show voters souring on the technology.

While moderate Democrats weigh taking a less intrusive approach to AI governance — in part to preserve their party’s remaining financial ties with Silicon Valley — progressives are urging aggressive measures, such as a new proposal by Sen. Bernie Sanders (I-Vt.) and Rep. Greg Casar (D-Texas) to permanently ban “superintelligent” AI models “that humanity cannot fully control.” Lawmakers have yet to get any guidance from the House Democratic Commission on AI and the Innovation Economy, a panel launched late last year by Minority Leader Hakeem Jeffries in hopes of crafting a consensus plan.

Republicans, meanwhile, are less than unified on whether AI is a boon or a threat to humanity. And Trump’s own AI policies have been more heavy-handed than his most ardent tech backers had hoped, with his administration imposing abrupt restrictions earlier this year on access to advanced models from Anthropic and OpenAI on national security grounds.

Trump also signed an executive order in June to set up voluntary federal vetting of advanced AI models, which Reed and Buchanan call similar to a Biden-era policy that Trump rescinded on Inauguration Day. White House officials have debated creating a regulatory agency to handle the vetting — and briefly floated giving it the same kind of strict authority that the Food and Drug Administration exerts over prescription drugs.

Coalitions of tech companies even expressed alarm this summer that the Trump administration may be preparing to curtail the spread of open-weight AI models — the same models that had inspired Andreessen’s worries about Biden’s policies more than two years ago. The White House did not directly say at the time whether those fears were accurate, but said the administration “is doubling down on innovation” to ensure America maintains its AI lead.

“Ironically,” Reed and Buchanan write, the Trump administration’s inconsistent approach to regulations could create an AI ecosystem dominated by a handful of top companies — a market that would look “a lot like the one Andreessen claims to fear.”

Democrats’ scrambled Silicon Valley strategy

As that debate plays out, tech investors including former Trump AI czar David Sacks continue to cite the May 2024 meeting as a key reason for the tech sector’s tilt toward the GOP. “I do believe that Marc was correct in his interpretation of what he heard, because I heard similar things,” Sacks said in a podcast posted last month.

The shift toward Trump has had a sizable impact on the political leanings of the tech world’s deepest pockets.

After years of serving as a reliable source of donations for Democrats, several Silicon Valley titans defected to the GOP in 2024 — most notably Elon Musk, who spent hundreds of millions of dollars in support of Trump and other Republican candidates and is on track to spend millions more in the 2026 elections.

Andreessen also spent heavily to elect Trump and has since poured money into the midterms, including through a super PAC network set up to elect pro-AI industry candidates. The Andreessen Horowitz firm, also known as a16z, emerged earlier this year as the biggest single donor in this year’s elections, The New York Times reported in May.

Still, other tech billionaires have remained loyal to Democrats. Last year, LinkedIn co-founder Reid Hoffman admitted the Democratic Party “really did alienate” parts of the industry — but that hasn’t stopped him from spending heavily for the party in this year’s elections.

A burgeoning civil war in Silicon Valley over AI policy is also shifting how Democrats approach the sector for campaign cash.

Just a few months ago, many AI safety advocates worried Andreessen’s super PAC network would cause legislators to hesitate before regulating the fast-moving technology. But now other tech donors are funneling gobs of money to candidates who back stricter rules on AI.

Public First, a competing super PAC network backed by top Anthropic executives and other deep-pocketed tech interests, says it has raised at least $80 million since its launch late last year. While the group is ostensibly bipartisan, much of that money has gone to Democratic congressional candidates who pledge to rein in AI.