Prof G Markets – Trump’s $2,000 Tariff Dividend Doesn’t Add Up — Here’s Why

Date: November 12, 2025 | Host: Ed Elson | Guests: Justin Wolfers, Alex Heath

Episode Overview

This episode dives deep into two major stories shaking the markets:

1. Trump's $2,000 Tariff Dividend Proposal: President Trump’s weekend announcement of a $2,000 tariff ‘dividend’ for all but the top earners. Is it economically viable, and what are its implications?
2. OpenAI vs Anthropic – Business Model Showdown: New Wall Street Journal reporting reveals divergent business strategies and spending philosophies between AI giants OpenAI and Anthropic.

Through expert commentary from economist Justin Wolfers and tech journalist Alex Heath, host Ed Elson breaks down the economic realities and market implications—delivered with the pod’s signature irreverence.

Segment 1: Trump’s $2,000 Tariff Dividend—Economic Nonsense or Strategic Masterstroke?

Background and Proposal

• Trump proposes a $2,000 annual “tariff dividend” for most Americans, allegedly funded by tariff revenues.
• No official government plan, numbers, or details have been released (02:08).

Expert Take: Prof. Justin Wolfers Analysis (from 04:05)

Key Criticisms

• Fundamental Flaw: The proposal simply takes money from Americans through tariffs, runs it through the IRS, and sends it back as checks.

  “In the best possible scenario, we come out even. The best thing we can do is not harm people.” (05:00)

• Tariffs as Consumption Taxes:

  “A tariff is a form of a consumption tax... exactly the same logic Republicans use to say we shouldn’t tax income also applies to taxing consumption, also applies to tariffs.” (05:35)

• No Economic Justification:

  “If you want to reduce how much tariffs are hurting Americans, I have this whole much simpler plan: don’t levy the tariffs.” (06:26)

Common Misconceptions Dispelled

• Foreigners Don’t Pay the Tariff: Addressing the argument that ‘money comes from China and flows to Americans’:

  “Tariffs are a tax on Americans, a tax on importers as well.” (06:47)

• Math Doesn’t Add Up: US collects about $250–300B in tariffs—it would need double that to pay out the proposed ‘dividend’:

  “He’s planning on sending out $600 billion in checks... The dividend he’s declared is more than all [the revenue]... There’s no rest. It’s negative. The national debt’s going up.” (12:03)

Populist Theater, Not Policy

• No Work Shown, No Seriousness:

  “My son is 13… He has to show his work… The stakes of whether he gets question five on his middle school math test are far lower than the macroeconomics of a $2,000 rebate per person.” (09:40)

• Process Breakdown: Unlike legitimate policy, no analysis or background released; “populist nonsense.” (08:25)

Irony: Proposal Helps the Poor

• Regressive Tax, Progressive Payout:

  “Pull in this money in a regressive way, but… giving it out as a flat $2,000 per, which is incredibly positive in terms of redistributing towards working or middle class Americans. This is extremely progressive. So even though it’s a regressive tax, if he sends it out as a flat amount, working class Americans would come out on top. I love that for him. And guess what? That’s how I know it’s not going to happen.” (12:53)

Memorable Soundbites

• “Dumbest idea I’ve heard in my head.” (04:40, Justin Wolfers)
• “If you want to reduce how much tariffs are hurting Americans… don’t levy the tariffs.” (06:26, Wolfers)
• “My son is showing more of his work than the President.” (09:40, Wolfers)

Meta-Commentary: How Seriously Should We Take Nonsense? (14:20)

• Justin and Ed address media responsibility:

  “What truths can we tell? How seriously should we take nonsense? I mean, it’s the President, but it’s nonsense. But sometimes he does nonsense.” (14:20, Wolfers)

• Conclusion: Ridicule is sometimes the only appropriate response to unserious proposals.

Segment 2: OpenAI vs Anthropic—AI Titans, Divergent Strategies

(Starts 19:11, Guest: Alex Heath)

New Insights from Wall Street Journal Leaks

• Business Models:
  • Anthropic: Focused, primarily B2B, aiming for break-even by 2028; spends far less, sells mainly API access.
  • OpenAI: Consumer-first, massive losses projected ($74B in losses for 2028), “spends like crazy,” burns 14x as much cash as Anthropic.
• Strategic Focus:
  • OpenAI aims to be a “multi-trillion dollar multiplatform tech conglomerate,” fostering a broad, splashy products approach.
  • Anthropic hones in on coding, enterprise software, avoids spending big on hardware, consumer products, or multi-modal AI.

The Sam Altman Meltdown (22:01)

• Altman’s “frantic” viral response when pressed on how to pay for $1.5T in spending was likely staged for effect, but “not a good look”:

  “Saying, basically, trust me bro, does not endear confidence.” (22:26, Heath)

• Sarah Friar floats idea of federal ‘backstop’ for AI funding—a “trial balloon”—which the market rejected.

  “If you go back and listen... it’s not a flub… There was [a] moment... she was like, yeah, yeah, like a backstop.” (23:54, Heath)

SoftBank Sells Nvidia to Fund AI Investments (24:40)

• SoftBank cashes out nearly $6B in Nvidia shares; likely intends to invest proceeds in OpenAI.
• Market Impact: Nvidia stock falls, causing broader tech declines.
• Heath’s take:

  “Very cynically, the take could be that SoftBank is selling its Nvidia stake to fund companies whose main capex is buying Nvidia chips... There is a very circular nature to all of this.” (25:07)

Market Sentiment: AI Bubble Anxiety

• Recent stock volatility, rising bond spreads, and search trends all signal investor nerves:
  • Nvidia dropped as much as 9% this week; CoreWeave fell >20%.
  • Google searches for “AI bubble” surged 10x in 6 months.
  • Big players (Bezos, Gates, Solomon) hint openly at a potential bubble.

Key Quote (Ed Elson, 28:00)

“Now, up until this point, OpenAI has been the winner... the undisputed heavyweight champion of AI. But... the narrative on AI is beginning to waver... and that may well mean pain for OpenAI... more importantly, it might mean new opportunities... companies that are reining in their spending, companies that are getting their costs under control... that winner could well be Anthropic.”

Notable Quotes

• “If you want to reduce how much tariffs are hurting Americans… don’t levy the tariffs.” (06:26, Justin Wolfers)
• “My son is showing more of his work than the President.” (09:40, Justin Wolfers)
• “Saying, basically, ‘trust me bro’ does not endear confidence.” (22:26, Alex Heath)
• “Very cynically… SoftBank is selling its Nvidia stake to fund companies whose main capex is buying Nvidia chips... there is a very circular nature to all of this.” (25:07, Alex Heath)
• “The narrative on AI is beginning to waver… that may well mean pain for OpenAI… that winner could well be Anthropic.” (28:00, Ed Elson)

Timestamps & Key Segments

• 01:50 – Quick stats, Market wrap, and Trump tariff proposal intro
• 04:05 – Justin Wolfers joins, eviscerates tariff dividend logic
• 09:40 – “Show your work”: Justin’s son vs. the President
• 12:00 – The numbers behind tariff revenue vs. payout
• 13:46 – Tariff payout is accidentally progressive (but won’t happen)
• 15:11 – Meta-discussion: Should we even seriously debate unserious proposals?
• 19:11 – OpenAI vs. Anthropic rivalry and strategy breakdown with Alex Heath
• 22:01 – Sam Altman's on-stage drama and OpenAI's federal ‘backstop’ balloon
• 24:40 – Softbank’s Nvidia sale, market effects, and AI capital recycling
• 28:00 – Ed’s wrap: AI bubble, investor nerves, why focus may trump scale in coming years

Summary Takeaways

• Trump’s $2,000 Tariff Dividend is economically incoherent:

  • Tariffs are consumer taxes; you can’t “give back” what you just took unless you want a costly, pointless exercise.
  • The math is off by hundreds of billions, and no one in government has shown foundational work behind the idea—because none exists.
  • The proposal is more campaign theater than policy, serving as “populist nonsense.”

• OpenAI faces new headwinds as AI narrative shifts:

  • Investors and insiders are anxious about profligate AI spending and the lack of clarity on profitability.
  • Anthropic, focused and frugal, may emerge as a more sustainable AI winner if the bubble pops.

For listeners who missed the episode:
You’ll find sharp economic analysis, market-moving news, and blunt, often hilarious commentary—especially on the emptiness of some recent political proposals and the tectonic clashes shaping the AI industry.