TBPN
TBPN

Oct 27, 2026 · 2h 11m

AI reshapes factories, finance, and the infrastructure trade

Anduril Laptop, DoorDash + Instacart, Gavin Baker Joins | Miles Arnone, Rahul Sengottuvelu

The episode connects reshoring, enterprise software, compute markets, and AI-enabled science to a broader question about who captures the technology’s economic gains.

3 key takeaways
  1. 1American industrial revival depends on rebuilding specialized supply chains, from drone motors and magnets to robotics.
  2. 2Enterprise AI is moving from experiments toward shared coding tools embedded in finance and operational workflows.
  3. 3Tight compute markets and expanding AI infrastructure are redirecting capital toward capacity, biology, and industrial systems.

Don't miss

Gavin Baker explains how tightening GPU markets, spot pricing, and forward contracts are changing the economics of AI infrastructure.

The brief

Anduril’s American-made laptop prompts a broader reckoning with the history of U.S. computer manufacturing, offshoring, and the difficulty of rebuilding production capacity.

OpenAI’s mathematical discoveries lead to a sharper question: which breakthroughs get shared publicly, and which advances in biology, materials, and cybersecurity remain commercially guarded?

Anthropic’s governance structure, China’s AI market, autonomous vehicles, and Harley-Davidson’s electric motorcycle struggle show how technology adoption depends on institutions, culture, and capital—not capability alone.

Miles Arnone describes Rebuild’s strategy for restoring industrial capacity by acquiring and incubating manufacturers tied to drone motors, rare-earth magnets, robotics, and resilient supply chains.

Rahul Sengottuvelu and Gavin Baker shift the focus to enterprise adoption and compute economics, from collaborative coding agents and CFO software to GPU pricing and infrastructure investment.

The episode closes on responsible optimism: AI may reshape ordinary work and ownership, but its benefits will depend on how broadly the resulting wealth and capacity are distributed.

What was said on this episode

80 statements · 52 positive · 17 negative · 3 mixed · 8 neutral

  1. John Cooganon American-made hardwarePositive3:14

    Jordi Hays is bullish on new, modifiable hardware and American manufacturing.

    “I'm very bullish on new hardware, new mutable hardware. And I love the fact that this one is going to be made in America.”

    Listen at 3:14

  2. John Cooganon American laptop manufacturingNeutral4:10

    The last American-made laptop may have entered production around 1993.

    “it is like way over 20 years. I think the last one that American made laptop went into production like 93.”

    Listen at 4:10

  3. Gavin Bakeron Commercial AI discoveriesNegative27:16

    Commercially valuable AI discoveries are likely being withheld for monetization.

    “You have to assume there's many other highly commercial discoveries happening in the background that people, both the labs. And people individually are not sharing broadly because they made a discovery and there's an opportunity to monetize it.”

    Listen at 27:16

  4. Gavin Bakeron Commercial AI breakthroughsNegative27:33

    John Coogan believes commercially valuable AI breakthroughs are often withheld.

    “anytime there's like a big breakthrough, my head is just going to everything that's not being shared.”

    Listen at 27:33

  5. John Cooganon AI-run businessesPositive28:39

    AI labs could eventually autonomously create, manufacture, distribute, and sell products.

    “you could have a big lab, say, like, go and build a business. And it could just do that. And it could do the market analysis. And then it could, like, create a new product and model it in CAD and send it to a factory and send it to a distribution center and pitch it to Walmart and get it on the shelves and, like, make the money and run that flywheel.”

    Listen at 28:39

  6. John Cooganon AI companiesPositive29:01

    Fully AI-run businesses could become economically significant.

    “you might have something economically interesting.”

    Listen at 29:01

  7. Gavin Bakeron AI-human collaborationNegative30:49

    Greater AI capability could make adding humans slow or derail projects.

    “there's a possibility that you get such a increase in capabilities that adding humans actually does just slow it down and derail the potential of a certain. project.”

    Listen at 30:49

  8. Gavin Bakeron NotionPositive34:02

    John Coogan thinks Notion may position itself as an AI-centered collaborative workspace.

    “my theory was Notion because they want to be a collaborative workspace where you have a bunch of agents coming in. They put themselves in the center. They're positioning themselves as an AI company.”

    Listen at 34:02

  9. Gavin Bakeron NotionPositive34:09

    Gavin Baker believes Notion positions itself as an AI company.

    “They put themselves in the center. They're positioning themselves as an AI company.”

    Listen at 34:09

  10. Gavin Bakeron Anthropic governancePositive37:53

    Anthropic has designed governance with an unusually long-term, centuries-oriented perspective.

    “this feels quite different. And it's cool to see a company thinking in centuries.”

    Listen at 37:53

  11. Gavin Bakeron Anthropic governancePositive37:53

    Anthropic has designed governance with a centuries-long perspective.

    “this feels quite different. And it's cool to see a company thinking in centuries.”

    Listen at 37:53

  12. John Cooganon Monkeys for preclinical testingPositive40:04

    Monkey demand and prices will rise because they are needed for preclinical testing.

    “monkeys are irreplaceable. So a bull market in monkeys is coming.”

    Listen at 40:04

  13. John Cooganon Research monkeysPositive40:06

    Demand and prices for research monkeys will rise.

    “a bull market in monkeys is coming.”

    Listen at 40:06

  14. Chinese AI companies have fallen about 40% since June.

    “since June, like AI companies broadly are down like 40%.”

    Listen at 40:34

  15. Jordi Hayson UnitreeNegative40:42

    Unitree is down about 50% since its IPO.

    “Unitree, since the IPO'd a few months ago, is down like 50%.”

    Listen at 40:42

  16. John Cooganon Tech-insider mediaPositive45:59

    Tech-insider media may occupy an underserved content niche.

    “there are a lot of news shows. There are a lot of interview shows. Where is the actual white space?”

    Listen at 45:59

  17. John Cooganon InstacartNeutral50:43

    Instacart reached approximately $13 billion market value after its 2023 IPO.

    “Instacart went public in 2023, had a market value around $13 billion.”

    Listen at 50:43

  18. John Cooganon DoorDashPositive50:52

    DoorDash is valued at approximately $73 billion.

    “DoorDash has a value of around $73 billion.”

    Listen at 50:52

  19. John Cooganon DoorDash acquisition of InstacartPositive50:59

    Acquiring Instacart would strengthen DoorDash's grocery-delivery offerings.

    “Buying Instacart would add a major grocery delivery partner to its growing roster of services.”

    Listen at 50:59

  20. Gavin Bakeron DoorDashPositive52:08

    DoorDash is an exceptionally strong company with continuing momentum.

    “I just think that DoorDash is generally sort of an unstoppable force.”

    Listen at 52:08

  21. John Cooganon Autonomous-vehicle companiesPositive54:03

    Autonomous-vehicle companies should organize racing events.

    “I think all the AV companies should be doing, like, race events.”

    Listen at 54:03

  22. John Cooganon LivewireNegative58:43

    LiveWire shares rose 2% that day but fell 62% year-to-date.

    “Livewire shares closed up 2%, though they are down 62% since the start of the year.”

    Listen at 58:43

  23. John Cooganon LiveWire sharesNegative58:43

    LiveWire shares rose 2% that day but were down 62% year-to-date.

    “Livewire shares closed up 2%, though they are down 62% since the start of the year.”

    Listen at 58:43

  24. Miles Arnoneon RebuildPositive1:01:06

    Rebuild employs about 2,000 people across 15 manufacturing facilities.

    “We're about 2000 people now across 15 facilities that can engineer and build complex products”

    Listen at 1:01:06

  25. Miles Arnoneon Production offshoringNegative1:01:57

    Offshoring production eventually erodes a company's product-design capability.

    “when you offshore production is you make some extra money for a while, but then you very quickly lose the ability to design the product effectively.”

    Listen at 1:01:57

  26. Miles Arnoneon U.S. drone BLDC motor manufacturingPositive1:04:05

    U.S. drone BLDC motor manufacturing is being restored relatively quickly.

    “on the motors, which are called BLDC motors, which are used on these drones, that problem is largely being solved, actually, in a pretty short period of time.”

    Listen at 1:04:05

  27. Miles Arnoneon Rebuild drone corePositive1:05:48

    Rebuild expects to wholesale its drone core in the U.S. for about $1,600.

    “we think we can sell in the U.S. wholesale for about $1,600.”

    Listen at 1:05:48

  28. Miles Arnoneon Industrial robotsNegative1:07:27

    Industrial robots may reduce total production costs only modestly while requiring high capital expenditure.

    “Bringing robots in might take a few percentage points off your total cost at a very high capital expense.”

    Listen at 1:07:27

  29. Miles Arnoneon Toyota Production SystemPositive1:07:33

    Lean production is more effective than robotics for eliminating manufacturing waste.

    “You're much better off trying to implement something like Toyota's lean production system to eliminate all the steps along the way that are wasteful”

    Listen at 1:07:33

  30. Miles Arnoneon Industrial robotsNegative1:08:13

    Robots alone will not solve U.S. manufacturing problems.

    “That won't happen.”

    Listen at 1:08:13

  31. Miles Arnoneon Slate truckPositive1:08:50

    Slate is developing an electric pickup truck priced below $25,000.

    “they're developing a sub-$25,000 electric pickup truck.”

    Listen at 1:08:50

  32. Miles Arnoneon RebuildPositive1:11:36

    Rebuild aims to become a multibillion-dollar manufacturing company.

    “build this business to be, you know, the multibillion dollar manufacturing company”

    Listen at 1:11:36

  33. Rahul Sengottuveluon Artificial IntelligencePositive1:13:06

    Gavin Baker believes AI is a real and valuable technology.

    “I think it's totally real.”

    Listen at 1:13:06

  34. Rahul Sengottuveluon Artificial IntelligencePositive1:13:06

    AI is a real and valuable technology for businesses.

    “I think it's totally real.”

    Listen at 1:13:06

  35. Rahul Sengottuveluon AI adoption by CFOsNeutral1:15:13

    AI will increasingly become a concern for chief financial officers.

    “AI is a CFO problem. It will become more of a CFO problem.”

    Listen at 1:15:13

  36. Rahul Sengottuveluon AI adoption in finance departmentsNeutral1:15:13

    AI will increasingly become a chief financial officer concern.

    “AI is a CFO problem. It will become more of a CFO problem.”

    Listen at 1:15:13

  37. Rahul Sengottuveluon Ramp InspectPositive1:16:20

    Ramp Inspect is designed for company-wide collaborative use through Slack.

    “we made it extremely collaborative multiplayer first. Works in Slack. Every session is public by default”

    Listen at 1:16:20

  38. Rahul Sengottuveluon Ramp InspectPositive1:16:20

    Ramp's Inspect is designed for highly collaborative, multiplayer coding workflows.

    “we made it extremely collaborative multiplayer first.”

    Listen at 1:16:20

  39. Rahul Sengottuveluon Ramp InspectPositive1:18:57

    Ramp Inspect supports multiple model providers and its own inference.

    “we have every provider running inference here. And so we also run our own inference for InSpec”

    Listen at 1:18:57

  40. Rahul Sengottuveluon Ramp InspectPositive1:18:57

    Inspect supports inference from multiple providers and Ramp's own models.

    “we have every provider running inference here. And so we also run our own inference for InSpec”

    Listen at 1:18:57

  41. Rahul Sengottuveluon Ramp InspectPositive1:20:28

    Ramp Inspect supports cybersecurity automation.

    “Cyber security is a big one for Inspect.”

    Listen at 1:20:28

  42. Rahul Sengottuveluon Ramp InspectPositive1:20:28

    Inspect supports cybersecurity automation within Ramp.

    “Cyber security is a big one for Inspect.”

    Listen at 1:20:28

  43. Rahul Sengottuveluon AnthropicPositive1:21:58

    Anthropic’s reported burn, profitability, cash generation, and unit economics were encouraging.

    “The fact that they've only burned $10 billion, the profitability, the cash generation, the unit economics, they're all very encouraging.”

    Listen at 1:21:58

  44. Rahul Sengottuveluon AnthropicPositive1:22:16

    About half of Fortune 500 companies adopted Anthropic at 0.1% of expenses.

    “half of the Fortune 500 has adopted Anthropic. And it's 0.1% of their expenses.”

    Listen at 1:22:16

  45. Rahul Sengottuveluon AnthropicPositive1:23:23

    Anthropic's annualized revenue may be approximately $67 billion.

    “Whatever the number is, $67 billion annualized”

    Listen at 1:23:23

  46. Rahul Sengottuveluon AnthropicPositive1:23:23

    Anthropic’s annualized revenue may be about $67 billion.

    “Whatever the number is, $67 billion annualized”

    Listen at 1:23:23

  47. Rahul Sengottuveluon AnthropicPositive1:25:58

    Anthropic's revenue comes from a highly diversified customer-spending base.

    “it shows a very diversified spending base for Anthropic.”

    Listen at 1:25:58

  48. Rahul Sengottuveluon AnthropicPositive1:25:58

    Anthropic has a diversified customer spending base.

    “it shows a very diversified spending base for Anthropic.”

    Listen at 1:25:58

  49. Rahul Sengottuveluon Open-weight AI infrastructure companiesPositive1:26:40

    Several open-weight AI infrastructure companies are growing faster than OpenAI or Anthropic.

    “there are so many companies, you know, whether it's Fireworks, whether it's Base 10, whether it's Together, whether it's Modal, whether it's Cognition, that are at scale, really open weight first. And they're all growing faster than either OpenAI or Anthropic.”

    Listen at 1:26:40

  50. Rahul Sengottuveluon Open-weight AI companiesPositive1:26:51

    Fireworks, Base10, Together, Modal, and Cognition are growing faster than OpenAI or Anthropic.

    “they're all growing faster than either OpenAI or Anthropic.”

    Listen at 1:26:51

  51. Rahul Sengottuveluon GrokBotPositive1:29:57

    Gavin Baker considers GrokBot his firm’s most impactful AI product so far.

    “Rockbot is the most impactful AI product to me and to the firm that we've had yet.”

    Listen at 1:29:57

  52. Rahul Sengottuveluon GrokBotPositive1:32:01

    Gavin Baker says GrokBot effectively routes information to the right people.

    “GrokBot is so... good at that.”

    Listen at 1:32:01

  53. Rahul Sengottuveluon GPU compute marketNegative1:34:03

    GPU rental prices have risen daily as spot compute supply tightens.

    “the price to rent, essentially any kind of a GPU, has gone up every day. So the spot market for compute is getting tighter.”

    Listen at 1:34:03

  54. Rahul Sengottuveluon GPU rental marketNegative1:34:03

    GPU rental prices have risen daily during October.

    “the price to rent, essentially any kind of a GPU, has gone up every day.”

    Listen at 1:34:03

  55. Rahul Sengottuveluon Compute marketNeutral1:34:52

    The compute market is currently in contango, with future capacity priced higher.

    “the compute market is now in contango.”

    Listen at 1:34:52

  56. Rahul Sengottuveluon CoreWeaveMixed1:36:31

    CoreWeave operates GPUs well but locked in long-term contracts below current market prices.

    “CoreWeave runs GPUs well. They bring them out. They are very, very good. They just sadly locked in a vast amount of compute and long-term contracts. that are now way, way under market.”

    Listen at 1:36:31

  57. Rahul Sengottuveluon GPU spot-market capacityPositive1:37:28

    Selling GPU capacity at spot prices can produce sub-one-year payback periods.

    “prices are going up. If we are willing to sell into the spot market, we can have paybacks well under a year”

    Listen at 1:37:28

  58. Rahul Sengottuveluon GPU spot-market capacityPositive1:37:28

    Selling GPU capacity in spot markets can produce sub-one-year paybacks and large prepayments.

    “prices are going up. If we are willing to sell into the spot market, we can have paybacks well under a year, get big prepayments.”

    Listen at 1:37:28

  59. Rahul Sengottuveluon DSX PodPositive1:39:33

    Nvidia's DSX pod will simplify GPU deployment through a preassembled seven-rack system.

    “DSX pod, which is going to make it a lot easier. You know, it's seven racks. It comes pre-assembled.”

    Listen at 1:39:33

  60. Rahul Sengottuveluon GPU cloud operatorsPositive1:40:44

    Several major cloud operators have proven GPU deployment capabilities.

    “these are companies with a proven track record of execution”

    Listen at 1:40:44

  61. Rahul Sengottuveluon 2027 compute clustersNegative1:42:04

    Securing a compute cluster in mid-to-late 2027 will remain very expensive.

    “you have to pay a truly vast amount to bring on a cluster in mid to late 27.”

    Listen at 1:42:04

  62. Rahul Sengottuveluon AI-enabled medicinePositive1:43:57

    AI has already helped save many lives through drug discovery or repurposing.

    “AI has already saved a lot of lives.”

    Listen at 1:43:57

  63. Rahul Sengottuveluon AI in medicinePositive1:43:57

    AI has already saved many lives through medical applications.

    “AI has already saved a lot of lives.”

    Listen at 1:43:57

  64. Rahul Sengottuveluon tomorrowxsummit.comPositive1:45:12

    Tomorrow X Summit aims to broaden access to investment information.

    “the whole goal is to kind of democratize investing.”

    Listen at 1:45:12

  65. Rahul Sengottuveluon Artificial IntelligencePositive1:51:16

    AI is currently benefiting ordinary and blue-collar Americans and small towns.

    “a lot of awesome things are happening today because of AI for ordinary Americans, for blue-collar Americans, for these small towns.”

    Listen at 1:51:16

  66. Rahul Sengottuveluon AI-driven U.S. reindustrializationPositive1:51:25

    AI-related activity is driving U.S. reindustrialization and major power-capacity expansion.

    “We're re-industrializing America. you know, massively increasing our electrical generation capacity”

    Listen at 1:51:25

  67. Rahul Sengottuveluon AnthropicMixed1:51:46

    Anthropic lost market share in summer but is reaccelerating and gaining share in October.

    “Anthropic, I called it like I saw it in the summer. They were losing market share. I'm calling it like I see it in October. They are re-accelerating and gaining share.”

    Listen at 1:51:46

  68. Rahul Sengottuveluon AnthropicMixed1:51:46

    Anthropic lost market share in summer but is reaccelerating and gaining share in October.

    “Anthropic, I called it like I saw it in the summer. They were losing market share. I'm calling it like I see it in October. They are re-accelerating and gaining share.”

    Listen at 1:51:46

  69. Rahul Sengottuveluon Rare-earth refiningNegative1:54:58

    Rare-earth supply is constrained mainly by difficult, dirty refining rather than scarcity.

    “the problem with rare earths, it's not that they're rare. It's that refining them is a very dirty process.”

    Listen at 1:54:58

  70. Rahul Sengottuveluon Rare-earth elementsNegative1:54:58

    Rare-earth supply is constrained more by dirty refining than geological scarcity.

    “the problem with rare earths, it's not that they're rare. It's that refining them is a very dirty process.”

    Listen at 1:54:58

  71. Rahul Sengottuveluon SpaceX and Anthropic IPOsPositive1:56:49

    SpaceX and Anthropic IPOs will unlock substantial mutual-fund buying power for private markets.

    “SpaceX and Anthropic going public is actually going to free up a huge amount of buying power in private markets”

    Listen at 1:56:49

  72. Rahul Sengottuveluon SpaceX and Anthropic IPOsPositive1:56:49

    SpaceX and Anthropic IPOs would unlock substantial mutual-fund buying power.

    “SpaceX and Anthropic going public is actually going to free up a huge amount of buying power in private markets”

    Listen at 1:56:49

  73. Rahul Sengottuveluon AI cloud providersPositive1:58:37

    Cloud providers are expected to show accelerating growth in the next reported quarter.

    “you're going to see like there's a pretty big acceleration in the clouds last quarter”

    Listen at 1:58:37

  74. Rahul Sengottuveluon AI infrastructure external financingNegative1:59:24

    External financing for AI infrastructure may be nearing its peak.

    “I think we're getting close.”

    Listen at 1:59:24

  75. Rahul Sengottuveluon AI infrastructure financingNeutral1:59:24

    The peak of external financing for AI infrastructure may be approaching.

    “I think we're getting close.”

    Listen at 1:59:24

  76. Rahul Sengottuveluon Gavin BakerNeutral1:59:45

    Gavin Baker says he will revise his AI views when evidence changes.

    “I will change my mind when the facts change”

    Listen at 1:59:45

  77. Rahul Sengottuveluon Broad-based company ownershipPositive2:01:29

    Gavin Baker recommends companies and wealthy Americans distribute ownership stakes broadly.

    “I would encourage more company and more you know, more wealthy people in America to do what Elon and Michael Dell and, and of course, you know, Michael and Brad Kirshner have done.”

    Listen at 2:01:29

  78. John Cooganon NanotyrannusNeutral2:04:56

    Sotheby's sold the Jody Nano Tyrannus skeleton for $47 million.

    “Sotheby's just sold Jody, a nano Tyrannus, a smallish friend of the T-Rex, for $47 million.”

    Listen at 2:04:56

  79. John Cooganon NanotyrannusPositive2:05:03

    The skeleton sold for five times Sotheby's high estimate of $9 million.

    “The high estimate was $9 million. It went for five times what they were predicting.”

    Listen at 2:05:03

  80. Gavin Bakeron Artificial intelligence terminologyNegative2:09:56

    Renaming artificial intelligence will not make people like it more.

    “I don't think people are going to start loving it because it has a new name.”

    Listen at 2:09:56

Statements are attributed to the speaker as said on the episode and reflect their view at the time, not PodLume's. They are not advice.

Weekly email

State of the Pod, every Monday

The week's most-discussed names and stories from podcasts, with the moments behind them. Free, five minutes, unsubscribe in one click.

Listen to the full episode and explore every guest, topic, and moment on PodLume.