Skip to main content
All-In with Chamath, Jason, Sacks & Friedberg

AI Sovereignty Wars, Palantir-Nvidia Deal, SCOTUS Birthright Ruling, Newsom's CA Budget Lie

102 min episode · 3 min read

Episode

102 min

Read time

3 min

Topics

Career Growth, Health & Wellness, Remote Work

AI-Generated Summary

Key Takeaways

  • Enterprise AI Data Risk: Frontier model providers including Anthropic have launched vertical products — Claude Code, Claude Design, Claude Legal, Claude Science — directly competing with customers who trained them on proprietary data. Anthropic's chief product officer sat on Figma's board before launching a competing design tool, causing Figma's stock to drop roughly 50%. Enterprises sharing proprietary workflows with these providers are effectively funding their own competition. The strategic response is to use open-source models on controlled infrastructure.
  • Open-Source Cost Arbitrage: Chamath's company 8090 ran a controlled benchmark comparing Claude Opus 4 alone versus Claude with a software harness versus an open-source frontier model with the same harness. The open-source plus harness combination delivered results 16.4x cheaper than standalone Claude, with a 3x slower runtime measured in hours. For non-time-critical enterprise workflows, this cost differential makes continued reliance on closed frontier models financially indefensible for most organizations.
  • AI Stack Power Dynamics: The model layer is consolidating into an Anthropic-OpenAI duopoly, with Anthropic simultaneously pursuing regulatory capture by framing open-source models as safety risks. Chip companies like NVIDIA and application companies like Palantir share a common incentive to keep the model layer competitive. NVIDIA's Nemotron open model, now available via Perplexity, is positioned to commoditize the model layer and create a long tail of enterprise buyers for GPU hardware.
  • Distributed Inference Architecture: Enterprises are moving from a large-hub, large-spoke cloud inference model toward a three-tier structure: major cloud providers handling roughly 70% of workloads, medium on-premise enterprise clusters handling 20%, and distributed local inference handling 10%. Friedberg projects that within months, most enterprises will run proprietary models on their own hardware for cost, security, and competitive differentiation reasons, with per-employee local compute becoming standard.
  • AI and Employment Data: A Ramp and Revelio Labs study of 21,000 U.S. firms found that high AI-spending companies grew headcount approximately 10% over two years following adoption, with entry-level hiring rising 12%. Low-intensity AI adopters showed flat headcount. The data contradicts the mass job-loss narrative, though the hosts acknowledge displacement in narrow categories — particularly level-zero customer support outsourced to the Philippines and autonomous vehicle markets where Waymo has reached critical mass.

What It Covers

Chamath, Sacks, and Friedberg analyze the Palantir-NVIDIA sovereign AI partnership and the enterprise data sovereignty crisis, where frontier labs like Anthropic systematically compete against their own customers. The episode also covers the Supreme Court's 6-3 birthright citizenship ruling, California's accounting-trick "balanced" budget masking $1.5–2 trillion in unfunded liabilities, and the AI job displacement debate.

Key Questions Answered

  • Enterprise AI Data Risk: Frontier model providers including Anthropic have launched vertical products — Claude Code, Claude Design, Claude Legal, Claude Science — directly competing with customers who trained them on proprietary data. Anthropic's chief product officer sat on Figma's board before launching a competing design tool, causing Figma's stock to drop roughly 50%. Enterprises sharing proprietary workflows with these providers are effectively funding their own competition. The strategic response is to use open-source models on controlled infrastructure.
  • Open-Source Cost Arbitrage: Chamath's company 8090 ran a controlled benchmark comparing Claude Opus 4 alone versus Claude with a software harness versus an open-source frontier model with the same harness. The open-source plus harness combination delivered results 16.4x cheaper than standalone Claude, with a 3x slower runtime measured in hours. For non-time-critical enterprise workflows, this cost differential makes continued reliance on closed frontier models financially indefensible for most organizations.
  • AI Stack Power Dynamics: The model layer is consolidating into an Anthropic-OpenAI duopoly, with Anthropic simultaneously pursuing regulatory capture by framing open-source models as safety risks. Chip companies like NVIDIA and application companies like Palantir share a common incentive to keep the model layer competitive. NVIDIA's Nemotron open model, now available via Perplexity, is positioned to commoditize the model layer and create a long tail of enterprise buyers for GPU hardware.
  • Distributed Inference Architecture: Enterprises are moving from a large-hub, large-spoke cloud inference model toward a three-tier structure: major cloud providers handling roughly 70% of workloads, medium on-premise enterprise clusters handling 20%, and distributed local inference handling 10%. Friedberg projects that within months, most enterprises will run proprietary models on their own hardware for cost, security, and competitive differentiation reasons, with per-employee local compute becoming standard.
  • AI and Employment Data: A Ramp and Revelio Labs study of 21,000 U.S. firms found that high AI-spending companies grew headcount approximately 10% over two years following adoption, with entry-level hiring rising 12%. Low-intensity AI adopters showed flat headcount. The data contradicts the mass job-loss narrative, though the hosts acknowledge displacement in narrow categories — particularly level-zero customer support outsourced to the Philippines and autonomous vehicle markets where Waymo has reached critical mass.
  • California Fiscal Collapse: California's state budget grew 65% from $215 billion in 2019 to $355 billion today. The top 1% of earners — 150,000 people — pay $70 billion, half of all personal income tax revenue. One to 1.5% of total personal income is leaving the state annually through corporate and individual exodus, with 2,100 mid-to-large companies relocated since 2019. Projected deficits reach $40 billion annually by 2028–2029, with $1.5–2 trillion in unfunded pension and retiree healthcare liabilities looming.
  • Birthright Citizenship Legal Framework: The Supreme Court ruled 6-3 in Trump v. Barbara that the 14th Amendment guarantees birthright citizenship, striking down Trump's executive order. The amendment was ratified in 1868 specifically to overturn the Dred Scott decision and grant citizenship to freed slaves, a context entirely distinct from modern immigration patterns. Friedberg's proposed framework — citizenship for children of legal permanent residents but not temporary visa holders or undocumented individuals — represents a middle-ground position Congress could legislate if the constitutional barrier were removed.

Notable Moment

During the Anthropic export control discussion, Sacks outlined three specific conditions that together triggered the government's letter to Anthropic: Dario publicly describing his own model as a cyber weapon, Amazon reporting that safety guardrails failed during testing, and Anthropic's apparent refusal to roll back the model proactively. Remove any single condition, Sacks argued, and the government action would never have occurred.

Know someone who'd find this useful?

Episode Transcript

Alright, everybody. Welcome back to the number one podcast in the world. It's your favorite podcast. It's your podcast. It's your favorite podcast. It's the all in podcast, episode two seven nine. With me, Friedberg, Sacks, Chamath. You know the the squad. You know the squad. We're here. It's the summer and we're ready to rock and roll. We got a power docket. We got a rocket docket Palantir and NVIDIA have announced. A sovereign AI partnership. Where have we heard that term before? Palantir is gonna use NVIDIA's Nemo Tron. Nemo Tron, like the Pixar film, open models to build a custom frontier quality model to serve the US government. Palantir is calling this new platform Sovereign AI Operating System. The US government agencies will own the hardware, the data, and the model weights. Palantir also shared a viral tweet manifesto laying out the concept, data retention is your treasure. Transfer it at your own peril. Transferring that data hands over access to your preexisting winning plays and yields the means of production for new ones. CEO Alex Karp went on CNBC to announce the partnership in a classic Karp Robin Williams style monologue. Here's a clip from his twenty minute interview where he where he basically went after the Frontier models like Anthropic. Play the clip. Our clients are just to say they're unhappy with the Frontier Labs is to say I'm welcome at the Berkeley faculty. It's like there's just a level of discomfort and loss of trust. Sam and and and Dario there's there's nothing more fun than debating Dario in private. So this is I'm not throwing shade at them. Okay. But something has gone completely wrong, and the basic view among enterprises in this country is I'm gonna chill wax, and waste my time with tokens. I'm gonna get no value, and they're gonna get my IP. When the d o Department of War goes to you and says, I need this application, do they get to control the weights to do it, or do you get to control the weights? Are we really gonna outsource the battlefield of this country to the consensus view in Silicon Valley? That is effing insane. Alright. And so some folks, refer to this as a televised nervous breakdown. We here at All In call that Alex Karp on a Tuesday. And we talked about this, Chamath, Sachs, Friedberg. We talked about this a whole bunch. Back in February, I coined the term intelligence sovereignty. Here, here's your victory fap. Do I wanna give all of the secrets in our organization, every piece of intellectual property to Sam Altman, who's gotta make a billion dollars a year to keep up with his spend? Right? He's going to build every application. I've been talking about AI sovereignty here for a bit just in terms of how much more cost effective it is and how you're not training other people's AIs with your knowledge and your insights. This is why …

Get the full transcript (19,483 words) + summary by email — free

One-time email with the complete transcript and AI summary of this episode. No account needed.

One email, no spam. We’ll also show you what SignalCast does.

Browse all All-In with Chamath, Jason, Sacks & Friedberg transcripts →

You just read a 3-minute summary of a 99-minute episode.

Get All-In with Chamath, Jason, Sacks & Friedberg summarized like this every Monday — plus up to 2 more podcasts, free.

Pick Your Podcasts — Free

Keep Reading

Books, tools, and gear mentioned in this episode

SignalCast may earn commission on purchases via these links.

Tools

  • by Anthropic

    Frontier model providers including Anthropic have launched vertical products — Claude Code, Claude Design, Claude Legal, Claude Science — directly competing with customers who trained them on proprietary data.
  • by Anthropic

    Chamath's company 8090 ran a controlled benchmark comparing Claude Opus 4 alone versus Claude with a software harness versus an open-source frontier model with the same harness.
  • by NVIDIA

    NVIDIA's Nemotron open model, now available via Perplexity, is positioned to commoditize the model layer and create a long tail of enterprise buyers for GPU hardware.
  • by Anthropic

    Frontier model providers including Anthropic have launched vertical products — Claude Code, Claude Design, Claude Legal, Claude Science — directly competing with customers who trained them on proprietary data.
  • by Anthropic

    Frontier model providers including Anthropic have launched vertical products — Claude Code, Claude Design, Claude Legal, Claude Science — directly competing with customers who trained them on proprietary data.
  • by Anthropic

    Frontier model providers including Anthropic have launched vertical products — Claude Code, Claude Design, Claude Legal, Claude Science — directly competing with customers who trained them on proprietary data.
  • NVIDIA's Nemotron open model, now available via Perplexity, is positioned to commoditize the model layer and create a long tail of enterprise buyers for GPU hardware.

More from All-In with Chamath, Jason, Sacks & Friedberg

We summarize every new episode. Want them in your inbox?

Similar Episodes

Related episodes from other podcasts

Explore Related Topics

This podcast is featured in Best Tech Podcasts (2026) — ranked and reviewed with AI summaries.

Read this week's Health & Longevity Podcast Insights — cross-podcast analysis updated weekly.

You're clearly into All-In with Chamath, Jason, Sacks & Friedberg.

Every Monday, we deliver AI summaries of the latest episodes from All-In with Chamath, Jason, Sacks & Friedberg and 192+ other podcasts. Free for one show.

Start My Monday Digest

No credit card · Unsubscribe anytime