AI Doom vs Boom, EA Cult Returns, BBB Upside, US Steel and Golden Votes
Speakers
David Friedberg
— CEO, The Production Board
Chamath Palihapitiya
— CEO, Social Capital
David Sacks
— General Partner, Craft Ventures
Jason Calacanis
— Angel Investor / Founder, LAUNCH
Summary
The original All-In quartet records from Washington DC and spends the first half on AI doomerism, with David Sacks arguing it is an organized, effective-altruism-funded lobbying ecosystem tied to Anthropic and former Biden AI staff whose real goal is global compute governance. David Friedberg counters the job-destruction thesis by arguing AI raises the return on invested capital, so more capital gets deployed and more jobs, growth and deflationary abundance follow. The second half revisits the Big Beautiful Bill, CBO scoring, the long end of the bond market and Chamath's charts showing the United States is already short power, with renewables and storage the only fast source of new supply. They close on the cleared Nippon Steel and US Steel deal, golden votes and national champions, and a disagreement over government stakes in markets and Social Security reform.
- Sacks maps the AI-doomer network to Open Philanthropy, effective altruism, Anthropic personnel and former Biden AI officials, and says the agenda is global compute governance.
- Friedberg argues AI raises the return on invested capital, so companies deploy more capital, create more jobs, and drive a deflationary drop in the cost of goods.
- Chamath says new grads were the old autocomplete, so entry-level hiring shrinks while revenue per employee rises and operating expense as a share of revenue falls.
- Sacks defends the Big Beautiful Bill: Byrd rules keep discretionary DOGE cuts out of reconciliation, and CBO scoring misstates the 2017 tax-cut baseline.
- Chamath calls the long end of the bond market the second biggest actor after Trump because it sets the US cost of capital, and notes the curve has already steepened.
- Chamath's power charts show the US consuming all the electricity it produces with demand growing 3 percent a year, SMRs a decade out and new gas plants four years out, leaving renewables and storage as the fastest supply.
- Chamath backs the golden-vote structure in the Nippon Steel and US Steel deal and wants US national champions in steel, pharma precursors, AI, chip equipment, batteries and rare earths; Sacks would reshore steel, aluminum and rare earths with subsidies.
- Friedberg opposes government equity stakes and prefers trade incentives, while the group agrees Social Security's 4.5 trillion dollars should be invested rather than lent to the Treasury before the 2032 funding cliff.
Ideas
AI boosts capital returns, not unemployment
Friedberg rejects the premise that AI destroys jobs. His argument is that AI raises the return on invested capital: one software engineer using AI generation tools can output 20 to 50 times more than before, so the same salary now buys far more value. When the return on deployed capital moves from 2x to 20x, companies and investors deploy more capital rather than less, which means more startups, more venture funding flowing in, more work getting done and more jobs created. The second leg is deflation and abundance: automation in software, food preparation and services drops the cost of goods, so people can work 30 hours instead of 60 and keep the same or a better standard of living. He also frames AI as an infinite game with no finish line, where the productivity gains accrue to every economy rather than to one winning nation.
Watch the long end of Treasuries
Chamath frames the fiscal debate around who actually prices the Big Beautiful Bill. After President Trump, he says the second biggest actor is the long end of the bond market: central bankers, long bond holders and macro hedge funds, because they ultimately determine the cost of capital for the United States irrespective of what the headline deficit number turns out to be. He points out that they have already steepened the curve and made it more expensive for the country to borrow, and that the bill only works if GDP growth comes in well above the low CBO assumptions. Long-end yields are therefore the variable he says everyone should be watching as the bill progresses.
US power is short and demand rising
Chamath argues the binding constraint on the GDP growth the whole fiscal plan depends on is electricity. The United States consumes essentially every unit of power it produces, there is no slack in the system, and demand is growing roughly 3 percent a year, so on the margin the country is already short power, with brownouts appearing because capacity was not added. At the same time hundreds of billions of dollars of data-center capital are being committed, including the 1 gigawatt data center in Arizona he announced that week. If energy supply contracts, the GDP number the bill relies on does not materialize, so keeping the power market robust and heavily invested is, in his view, the single most critical policy item and the one provision of the bill that has to be fixed.
Only renewables and storage can scale now
Chamath walks through the deployment timeline of every new source of power as of May 2025. Small modular reactors are a 2035-plus story even if permitting goes well and the technology is derisked, so they are a decade away. A newly planned natural gas plant is at least four years out, and 24 gigawatts of gas is already sitting in a queue that cannot be turned on. Restarting a mothballed nuclear reactor, of which only three exist, lands somewhere between 2027 and 2030. His conclusion is that renewables plus storage are the only ready supply that can be switched on fast enough to close the gap between demand and utilization, so that is where the incremental build has to come from regardless of partisanship.
Only renewables and storage can scale now
Chamath walks through the deployment timeline of every new source of power as of May 2025. Small modular reactors are a 2035-plus story even if permitting goes well and the technology is derisked, so they are a decade away. A newly planned natural gas plant is at least four years out, and 24 gigawatts of gas is already sitting in a queue that cannot be turned on. Restarting a mothballed nuclear reactor, of which only three exist, lands somewhere between 2027 and 2030. His conclusion is that renewables plus storage are the only ready supply that can be switched on fast enough to close the gap between demand and utilization, so that is where the incremental build has to come from regardless of partisanship.
US Steel gains national-champion state backing
On the Nippon Steel acquisition of US Steel that Trump just cleared, Chamath argues the United States has spent 20 years on the wrong side of these deals, showing up only when an asset is stranded or run into the ground, as with the auto bailouts and TARP, and getting almost nothing in return. This deal inverts that: the government keeps control through a golden vote and partial ownership while a strong foreign partner funds and recapitalizes the asset, the same structure Brazil, the UK and China use with their national champions. He expects the template to be repeated and suggests US Steel could even receive preferred capital from the government in the future, which turns the company into a state-backed national champion rather than an ordinary cyclical.
America needs national champions in strategic industries
Chamath wants the United States to copy the plan Hu Jintao laid out in 2003, which designated ten Chinese national champions in the industries that would matter for the next fifty years and then let them thrive and crush it. He names the categories he considers strategic for America: steel, the precursors for pharmaceuticals, AI, upstream lithography and deposition and chipmaking capability, batteries, and rare earths together with the specialty chemical supply chain. Controlling those five or six areas is what allows a country to keep its citizens healthy and manufacture everything else, so he expects and favors category-by-category government support, golden votes and preferred capital in exactly those industries rather than across the whole economy.
Steel, aluminum, rare earths must be reshored
Sacks argues that the offshoring of US manufacturing to China over the past 25 years was not a free-market outcome, because WTO developing-country status let China subsidize its national champions, dump product globally and drive competitors out of business as the low-cost producer. He then draws a line between industries that do not need to be reshored, such as clothes and toys, and those that do. Steel production, aluminum and rare earth capacity are strategic because the country cannot be completely dependent on China for that supply chain, and he is willing to use subsidies and protection for those specific sectors on national security grounds while keeping government out of the industries where private markets work fine.
Social Security money belongs in equities
Chamath says the 4.5 trillion dollars that Social Security beneficiaries have had deducted from their paychecks should be flipped into an investment account where retirees own equities, make investments in the markets and participate in the upside of American industry and the GDP growth he expects. Instead that money is lent to the federal government and earns roughly 3.8 to 4.5 percent in Treasuries that now carry the lowest credit rating the United States has ever had, while the deficit spending it funds inflates the value of the dollar away. The comparison is explicit: equity ownership of American industry beats holding downgraded government paper at four percent, and the program is functionally bankrupt by 2032 if nothing changes.
Up Next
This All-In Podcast video, published May 31, 2025, features David Friedberg, Chamath Palihapitiya, David Sacks discussing Artificial Intelligence, TLT, Power infrastructure, SOLAR, Small modular reactors, X, SLX, REMX, BATTERIES, SMH, Aluminum, VTI. 9 trade ideas extracted by AI with direction and confidence scoring.
Speakers: David Friedberg, Chamath Palihapitiya, David Sacks · Tickers: Artificial Intelligence, TLT, Power infrastructure, SOLAR, Small modular reactors, X, SLX, REMX, BATTERIES, SMH, Aluminum, VTI