All-In Podcast

Why the datacenter backlash could cap America's AI lead

Ratepayer groups, doomer activists, and local politicians are blocking permits; Allbirds' AI-pivot stock surge shows how scarce compute has become.

JCJason CalacanishostCPChamath PalihapitiyahostTKTravis KalanickguestDSDavid Sackshost

Apr 16, 2026·1h 30m·Watch on YouTube ↗

🙋 FREQUENTLY ASKED QUESTIONS

Direct answers grounded in the episode transcript. Tap any timestamp to verify against the source.

1. 1

   Why does David Sacks say Mamdani's pied-a-terre tax targets elastic buyers?

   David Sacks says the tax targets buyers who can most easily take their money elsewhere. In his explanation, a pied-a-terre owner is not locked into New York the way a primary resident is. They may want a second or third home, but they could buy it in another city. That makes them the most price-sensitive segment, so putting the heaviest tax on them would, in his view, hit demand for second homes in New York. He later adds that buyers considering a new pied-a-terre would now worry the rate could keep rising. His broader concern is development: high-end buyers like Ken Griffin can make a project pencil by paying a high price per foot, and removing that demand could dry up new construction.

   ▸ 1:16 in transcript
2. 2

   What was in Denise Dresser's leaked OpenAI memo about Anthropic?

   Jason Calacanis says the leaked memo framed Anthropic as both a revenue and strategy threat. He describes a four-page memo from OpenAI Chief Revenue Officer Denise Dresser that called out Anthropic, saying its $30 billion run rate was inflated by $8 billion because of revenue share and accounting with AI model providers. He also says the memo accused Anthropic of building a story around fear, restriction, and elite control of AI. The memo then laid out OpenAI's own push toward business customers and the agent platform layer. Jason connects that push to OpenAI hiring Peter Steinberger, the architect of the open source project OpenClaw, and to criticism from anonymous investors who thought OpenAI was losing focus while ChatGPT still had a huge consumer business.

   ▸ 11:23 in transcript
3. 3

   Why does enterprise coding matter in the OpenAI vs Anthropic race?

   David Sacks argues enterprise coding monetizes in a way consumer chat subscriptions do not. His point is that Anthropic focused on enterprise, especially coding, where businesses pay for code tokens on a metered basis, almost like electricity. As usage rises, revenue can rise with it. He contrasts that with consumer subscriptions, which OpenAI prioritized: consumers have lower willingness to pay, only a small share may convert to premium, and they tend to want a $20 per month all-you-can-eat plan. That makes consumer revenue scale differently from enterprise revenue. Sacks says OpenAI still should pursue enterprise, but with more focus on coding and business customers. If it does not catch up soon, he thinks Anthropic could build a lead over the next one or two years that becomes hard to overcome.

   ▸ 23:44 in transcript
4. 4

   Why did Allbirds' AI pivot make Chamath think compute is scarce?

   Chamath Palihapitiya treats the Allbirds stock spike as a canary for compute scarcity. After the hosts joke about Allbirds becoming New Bird AI, Chamath says the capital markets are getting something right: AI companies are massively compute constrained. He breaks the bottleneck into power and the physical land-and-shell problem for data centers. He points to Bloom Energy as an example of on-site natural gas power that can move faster than waiting years for grid access, and he says approvals for land and shells are becoming difficult as sentiment toward AI shifts. His warning is that Anthropic and OpenAI need land, power, and shells urgently. Otherwise, their revenue could slow or hit a wall because of infrastructure limits, not because customers stopped adopting the products.

   ▸ 39:50 in transcript
5. 5

   Why does David Sacks think data centers create blue-collar jobs?

   David Sacks says data center construction can support a long wave of blue-collar work. Responding to the idea that data centers do not create jobs, he argues that the build-out is already a boon for construction and could continue for a decade or two because AI capital expenditure is still increasing. He points to Jensen's projections and says the projects are creating tens of thousands of new construction jobs. He also says wages are rising 25 to 30% for electricians, carpenters, drywall crews, concrete workers, road builders, and equipment installers. Jason Calacanis pushes back that those jobs are not the same as permanent fab jobs, but Sacks' core claim is that the CapEx cycle itself is not a one-year event, so the construction employment effect can persist.

   ▸ 49:58 in transcript

Answers are AI-generated from the transcript and may contain errors. Tap a question to verify against the source.