Teahose.
SIGN IN
NEW HERE — WHAT TEAHOSE DOES
We read the entire AI & tech firehose — so you don't have to.
PODPodcastsAll-In, No Priors, Acquired…
NEWNewslettersStratechery, Newcomer…
PAPPapersPhysical AI research
PHProduct Huntdaily launches
VCInvestor ScoutSequoia, a16z, Benchmark…
CLAUDE DISTILLS →
7 reads, 30 sec each — free, 6 AM ET.
+ a live graph of the companies, people & themes underneath.
HOME/AXIOS PRO RATA/Axios Pro Rata: Bernie's AI plan
NEWS
// NEWSLETTER ISSUE
AXIOS PRO RATA

Axios Pro Rata: Bernie's AI plan

DATE June 18, 2026SOURCE AXIOS PRO RATAPARTICIPANTS DAN PRIMACK
// KEY TAKEAWAYS5 ITEMS
  1. 01Government Equity Seizure of AI Companies Is a Bipartisan Risk
  2. 02AI Talent Wars Are Intensifying
  3. 03OpenAI vs. Anthropic IPO Race Is the Defining Liquidity Event of the AI Era
  4. 04AI Infrastructure & Energy Is Drawing Serious Capital
  5. 05Defense & Dual-Use Tech Is a Durable Investment Theme
// SUMMARY

1. Key Themes

Government Equity Seizure of AI Companies Is a Bipartisan Risk

Sen. Bernie Sanders has proposed legislation requiring the U.S. government to take a 50% equity stake in any company with $200M+ in annual AI revenue, covering not just model companies but also "data centers, AI compute infrastructure, AI services, and advanced robotics." This would rope in Amazon, Nvidia, Microsoft, and SpaceX. The article warns investors not to dismiss this as fringe: "The White House is thinking along similar lines, just on a smaller scale. Strange bedfellows can make profound policy."


AI Talent Wars Are Intensifying — and Acqui-Hires Have a Structural Ceiling

Google paid $2.7 billion in 2024 to bring Noam Shazeer and part of his CharacterAI team in-house, only to see Shazeer leave for OpenAI two years later. The article frames this as a systemic problem: "This shows the limits of acqui-hires, in which the most valuable assets can walk out the door once retention periods end."


OpenAI vs. Anthropic IPO Race Is the Defining Liquidity Event of the AI Era

Goldman Sachs will lead OpenAI's IPO while Morgan Stanley leads Anthropic's, with both banks forming "bespoke banker teams to ensure sensitive info isn't shared." Meanwhile, Shazeer's defection to OpenAI is framed explicitly through the IPO lens: "Shazeer's hiring is a major win for OpenAI in the AI talent wars, as OpenAI competes to catch up with archrival Anthropic's most advanced models ahead of the two companies' hotly anticipated IPOs."


AI Infrastructure & Energy Is Drawing Serious Capital

Multiple deals this issue point to heavy VC conviction in the physical layer of AI: Verse (grid connections for data centers, $54M led by Bessemer with NVIDIA participating), Critical Energy (modular geothermal turbines, $19M seed), Form Energy (multi-day grid batteries, ~$2B raised, now pursuing IPO), and Deep Fission (underground nuclear reactors, Nasdaq IPO). The Sanders bill itself implicitly acknowledges this by defining AI companies to include "data centers, AI compute infrastructure."


Defense & Dual-Use Tech Is a Durable Investment Theme

Dream, an Israeli sovereign AI and cyber defense company, raised $260M at a $3B valuation. Earlybird and AVP launched a €500M growth equity fund "focused on defense and dual-use tech." EQT acquired Exolaunch, a launch mission management firm partnering with SpaceX. The deal flow across this issue signals sustained institutional capital formation in defense-adjacent sectors.


2. Contrarian Perspectives

The Sanders AI Bill Is More Dangerous Than It Looks — Because the Trump White House Agrees in Spirit

The consensus view would be to dismiss Sanders' 50% equity proposal as socialist grandstanding with zero chance of passage. But Primack explicitly pushes back: "Don't. Because the White House is thinking along similar lines, just on a smaller scale. Strange bedfellows can make profound policy." The tacit acknowledgment from both ends of the political spectrum that AI will cause "major labor disruptions" makes some form of government equity participation more plausible than markets may be pricing.


The Sanders SWF Would Hurt the Very People It's Designed to Help

The bill is framed as wealth redistribution, but the article surfaces a self-defeating structural flaw: if large tech companies are forced to dilute existing shareholders to give the government a 50% stake, that directly harms ordinary Americans who hold those stocks through retirement accounts. "Do these companies dilute existing investors, such as 401(k) managers and public pension funds? And, if so, isn't that taking from the same average Americans this bill is designed to help?"


The "AI Business" vs. "Non-AI Business" Separation Requirement Is Effectively Unenforceable

The bill requires companies to carve out their AI businesses for government equity purposes — but the article argues this is operationally incoherent for integrated tech giants: "How does a large, integrated tech company distinguish between its 'AI business' and 'non-AI business?' It reads a bit like online vs. offline in 1998, which soon became inextricably blurred." This suggests any enacted version of the bill would be mired in definitional litigation and regulatory arbitrage.


3. Companies Identified

OpenAI

  • AI model company pursuing IPO (Goldman Sachs leading)
  • Why mentioned: Hiring Noam Shazeer from Google; competing with Anthropic ahead of IPO; cited as a target of Sanders' equity bill
  • "Shazeer's hiring is a major win for OpenAI in the AI talent wars, as OpenAI competes to catch up with archrival Anthropic's most advanced models ahead of the two companies' hotly anticipated IPOs."

Anthropic

  • AI model company pursuing IPO (Morgan Stanley leading)
  • Why mentioned: Named as OpenAI's primary competitive rival; also a target of Sanders' bill
  • "OpenAI competes to catch up with archrival Anthropic's most advanced models ahead of the two companies' hotly anticipated IPOs."

CharacterAI / Google

  • Google paid $2.7B to acqui-hire Shazeer's team from CharacterAI in 2024
  • Why mentioned: Cautionary case study on the limits of acqui-hires
  • "This shows the limits of acqui-hires, in which the most valuable assets can walk out the door once retention periods end."

Dream

  • Israeli sovereign AI and cyber defense company
  • Why mentioned: $260M raise at $3B valuation; signals sovereign AI as an emerging investment category
  • Described as "an Israeli sovereign AI and cyber defense company for governments and critical infrastructure."

Verse

  • Grid connections company for data centers
  • Why mentioned: $54M Series B led by Bessemer, with NVIDIA as investor — highlights AI infrastructure investment theme
  • Listed under VC deals with NVIDIA participation noted

Form Energy

  • Multi-day grid battery developer
  • Why mentioned: ~$2B raised in VC; selected JPMorgan and Jefferies to lead IPO — clean energy storage IPO to watch
  • "It's raised nearly $2b in VC funding."

Rylo (fka Nagish)

  • Communication platform for the deaf and hard-of-hearing
  • Why mentioned: $85M raise at $500M valuation led by General Catalyst and Canaan — accessibility tech gaining institutional backing

XDOF

  • Robot training data company
  • Why mentioned: $70M raise from Thrive Capital, a16z, Lux Capital and others — signals strong VC conviction in physical AI/robotics data infrastructure

Dragos

  • Cybersecurity firm (OT/ICS focus)
  • Why mentioned: Accenture acquisition at $4.18B total deal value — major liquidity event for industrial cybersecurity
  • Had "raised over $450m from firms like Allegis Capital, Canaan Partners, Koch Disruptive Technologies, and Energy Impact Partners."

Thoma Bravo / Medallia

  • Customer feedback software; TB surrendered it to lenders
  • Why mentioned: Represents a "$5.1b loss for TB" — a notable PE write-down case study in overpriced SaaS buyouts

Capital Factory

  • Austin-based venture firm and startup accelerator
  • Why mentioned: Founder Joshua Baer died in a private plane crash; major figure in Austin's entrepreneurial ecosystem

SpaceX

  • Launch and space company
  • Why mentioned: Roelof Botha (Sequoia) joining the board; cited as now more valuable than OpenAI; also named as a potential target of Sanders' AI equity bill
  • Botha "told me during Axios BFD last fall that SpaceX could become more valuable than OpenAI, in which Sequoia also invested. And it's happened."

4. People Identified

Noam Shazeer

  • AI researcher, formerly co-lead of Google Gemini; departing Google for OpenAI
  • Why mentioned: High-profile talent defection illustrating the limits of acqui-hires and the stakes of AI talent competition
  • "Shazeer also is viewed as one of AI's top researchers, particularly in the area of pretraining, and was co-lead of Google's Gemini."

Sen. Bernie Sanders (I-Vt.)

  • U.S. Senator; introduced the American AI Wealth Fund bill
  • Why mentioned: Proposed legislation requiring government 50% equity in large AI companies — a policy risk for AI investors
  • "Sen. Bernie Sanders (I-Vt.) today unveiled his plan for the U.S. government to take 50% equity stakes in large AI companies."

Roelof Botha

  • Partner at Sequoia Capital; joining SpaceX board of directors
  • Why mentioned: His earlier public prediction that SpaceX would surpass OpenAI in value has proven correct; board appointment signals deepening Sequoia-SpaceX ties
  • "Botha told me during Axios BFD last fall that SpaceX could become more valuable than OpenAI, in which Sequoia also invested. And it's happened."

Joshua Baer

  • Founder of Capital Factory; died in a private plane crash on June 17, 2026
  • Why mentioned: Major figure in Austin's startup ecosystem; loss noted across the VC community
  • Bill Gurley: "No one tilted harder as an enthusiastic supporter of Austin entrepreneurism. Will be greatly missed."

Kevin Warsh

  • New Federal Reserve Chair
  • Why mentioned: Presided over his first Fed meeting, leaving rates unchanged; Fed projections suggest rate increases later in 2026 — a macro headwind for risk assets
  • "The Fed left interest rates unchanged at Kevin Warsh's first meeting as chair, and released projections showing that many Fed governors expect rates to rise later this year."

5. Operating Insights

Acqui-Hires Require Longer or Stronger Retention Structures — or They're Just Expensive Loans of Talent

Google's $2.7B acqui-hire of Shazeer lasted approximately two years before he walked to a competitor. For operators and acquirers, this underscores that retention periods must be calibrated to the realistic integration timeline of key talent — and that talent whose value is primarily individual (rather than embedded in systems or teams) represents a structurally higher flight risk. The article frames the lesson bluntly: "The most valuable assets can walk out the door once retention periods end."


AI Companies Should Begin Scenario Planning for Government Equity Demands Now

Even if the Sanders bill fails, the fact that both a progressive senator and the Trump White House are independently considering government equity in AI companies signals that this policy direction has momentum. The $200M annual AI revenue threshold is not a distant benchmark for many fast-growing AI businesses. Founders and CFOs should model the structural and governance implications — including forced business separation — before it becomes urgent. The article warns: "Strange bedfellows can make profound policy."


6. Overlooked Insights

The Sanders Bill's SWF Cannot Pay Its Promised Dividend — By Design

The article notes a quiet structural impossibility embedded in the bill: "The bill does not detail how the SWF would provide the annual 5% distribution, given that it's prohibited from selling any of its underlying equities." With no sell mechanism and equity stakes in companies that may not pay dividends, the promised 5% annual payout to Americans is mathematically unfunded. This detail is easy to miss but is central to evaluating the bill's credibility — and suggests any enacted version would require significant structural revision.


The Fed's Dot Plot Under New Chair Warsh Points to Rate Increases — Not Cuts

Most market participants have been focused on the rate-cut narrative. But the Fed's latest projections show "many Fed governors expect rates to rise later this year" — at Kevin Warsh's very first meeting as chair. This is a potentially significant macro shift that received minimal emphasis in the newsletter but has direct implications for VC deal valuations, IPO pricing windows, and leveraged buyout economics.