Hadrian, SpaceX Veterans Build LA Hub to Counter China Factory Mass
Bloomberg
Software-automated factories and targeted component bans are required to counter state-subsidized Chinese manufacturing mass as blunt trade tariffs fail to protect critical US supply chains.
AI-assisted summary of How LA Is Quietly Becoming America's New Industrial Tech Hub — original source. Check the original for full context.
Source published:
Key ideas
- Southern California has reemerged as the primary American hardware and defense tech hub, driven by dense supplier networks and talent cross-pollination between legacy aerospace primes and modern startups. (15:43)
- SpaceX achieved 165 rocket launches in a single year and deployed roughly 10,000 satellites—triple the rest of the world combined—countering claims of terminal American manufacturing decline. (5:40)
- Even with modeled zero labor and energy costs, US raw bar stock alone costs three times more than a finished, precision-machined part imported from China due to extensive foreign state subsidies. (36:20)
- Wargame simulations indicate US munition stockpiles would deplete within seven days of high-intensity conflict, exacerbated by vast production gaps where China built 2,600 ships compared to six by the US. (22:12)
Detailed notes
## Background Published by Bloomberg on September 21, 2026, via RSS and web platforms, this episode captures the first-ever live Odd Lots show recorded at the Vermont Hollywood in Los Angeles. According to the show description, hosts Joe Weisenthal and Tracy Alloway sit down with Impulse Space founder Tom Mueller and Hadrian founder Chris Power to challenge the narrative of American industrial decline and examine why Southern California is emerging as a critical hub for advanced hardware, rocket propulsion, and precision manufacturing.
## TL;DR Recorded live in Los Angeles, the panel explores the resurgence of advanced hardware manufacturing across Southern California and unpacks why the United States previously lost its industrial supremacy to financialization. Chris Power outlines Hadrian's model of software-driven autonomous factories designed to provide scalable contract manufacturing for aerospace and defense, contrasting American vulnerability with China's massive state-subsidized production base. Tom Mueller draws on eighteen years as SpaceX's Employee No. 1 to counter fatalistic decline narratives, detailing how Impulse Space handles orbital logistics and why initial prototyping must stay vertically integrated in-house. Together, the founders debate the geopolitical urgency of physical mass, the bottlenecks of robotics and artificial intelligence, and why targeted component bans are far more effective than blunt import tariffs for securing critical domestic supply chains.
## Atmosphere & style The live theater setting imbues the discussion with an energetic, punchy tempo, blending Joe and Tracy's wry, audience-engaging inquiries with unfiltered industry candor. The conversation crackles with a compelling contrast between Chris Power's sweeping, macro-historical warnings about civilizational decay and Tom Mueller's understated, battle-tested engineering pragmatism. Moving from dry, humorous jabs about startup culture to stark, high-stakes assessments of geopolitical conflict and wartime stockpiles, the episode feels intellectually intense, urgent, and grounded in the physical reality of modern heavy industry.
## Bulletin - `15:43` Southern California has reemerged as the primary American hardware and defense tech hub, driven by dense supplier networks and talent cross-pollination between legacy aerospace primes and modern startups. - `05:40` SpaceX achieved 165 rocket launches in a single year and deployed roughly 10,000 satellites—triple the rest of the world combined—countering claims of terminal American manufacturing decline. - `36:20` Even with modeled zero labor and energy costs, US raw bar stock alone costs three times more than a finished, precision-machined part imported from China due to extensive foreign state subsidies. - `22:12` Wargame simulations indicate US munition stockpiles would deplete within seven days of high-intensity conflict, exacerbated by vast production gaps where China built 2,600 ships compared to six by the US.
## Key takeaways - `02:50` Chris Power argues that dominant global powers follow a recurring 150-year historical cycle, transitioning from generative manufacturing supremacy to financialized complacency and eventual hegemonic decline. - `05:40` Tom Mueller contends that SpaceX's operational scale and rapid launch cadence disprove the fatalistic narrative that the United States has permanently lost the capability to build complex hardware. - `06:20` Impulse Space is developing in-space transit vehicles to function as orbital freight vans, boosting the effective deep-space payload delivery of rockets like the Falcon 9 by up to five times. - `10:00` Hadrian operates autonomous, software-driven factories-as-a-service, serving as a contract manufacturer for defense primes across spaceflight, missile assemblies, and Virginia-class submarines. - `11:17` Factory floor automation is fundamentally a software and closed-loop algorithmic control challenge rather than a humanoid robotics problem, replacing manual machinist calculations and preventing costly tool crashes. - `15:43` A distinct geographic division of labor has emerged in tech: Silicon Valley dominates software, while Los Angeles and the South Bay lead national hardware, defense, and aerospace production. - `18:24` Hardware companies benefit from keeping early prototyping and development in-house to resolve technical flaws, but must transition to specialized external manufacturers once designs stabilize for high-rate production. - `20:00` In contract manufacturing, delivery schedule reliability outweighs pure unit economics, prompting modern shops to spend millions on internal CapEx rather than risk delays from legacy suppliers. - `22:12` Geopolitical power is determined by physical mass and manufacturing logistics, mirroring World War II dynamics where American mass production overwhelmed superior German engineering. - `25:00` The competitive advantage in robotics and AI depends on real-world training data velocity, which is structurally throttled by physical manufacturing capacity. - `25:00` Hostile off-world environments like the Moon and Mars make autonomous, dexterous robotic labor an absolute necessity rather than relying on human workforces for construction. - `30:00` The meteoric rise of large language models drastically accelerated projected timelines for artificial general intelligence, elevating the AI race into an urgent geopolitical governance contest. - `35:05` SpaceX eliminated the Starship booster's landing legs in favor of mid-air launch tower chopstick catches to purge parasitic mass and maximize orbital payload capacity. - `36:20` Global manufacturing cost distortions stem from foreign state subsidies across energy, labor, CapEx financing, and freight, rather than a domestic deficit in engineering talent. - `40:00` Blanket trade tariffs frequently fail due to regulatory loopholes and transshipment evasion, making direct bans on adversarial components paired with allied sourcing mandates a more resilient industrial policy.
## Detailed walkthrough ### `01:40` Live from Hollywood: The Re-Industrialization of Southern California
Opening their first-ever live show in Los Angeles at the Vermont Hollywood, hosts Joe Weisenthal and Tracy Alloway introduce a deep dive into an often-overlooked dimension of American manufacturing: the thriving industrial tech ecosystem taking root across Southern California. While national headlines constantly lament the hollowing out of domestic production and argue that the United States can no longer compete with China in advanced manufacturing, a counter-movement of hardware and aerospace companies is rapidly expanding around Los Angeles, producing high-precision components, rocket hardware, and jet engines.
To explore what rebuilding physical capacity actually requires on the ground, Joe and Tracy bring out two key founders at the forefront of the movement: Chris Power, founder and CEO of Hadrian (which is automating precision manufacturing and modern machine shops), and Tom Mueller, founder of Impulse Space and Employee No. 1 at SpaceX, where he spent nearly two decades leading propulsion development. Joe kicks off the panel by asking for their diagnosis of the broader narrative: why did the United States lose its manufacturing edge in the first place?
`02:50` The 150-Year Cycle of Empires, Finance, and "Romaboos"
Chris Power begins by offering a choice between a thirty-second political soundbite and an accurate, historical explanation before laying out his civilizational thesis. Throughout history, Power argues, every dominant world power follows an identical trajectory anchored to four structural stages: cultivating a generative population, building supreme industrial and manufacturing capacity, leveraging that dominance to win or prevent major wars, and ultimately becoming complacent. In this final decadent phase, a nation trades its hard-won physical production for a purely financialized economy, offshoring the very capabilities that made it powerful while clinging to the illusion that holding the global reserve currency alone guarantees perpetual hegemony.
Power points out that this structural transition—seen when the Dutch Republic ceded global dominance to the British Empire, and when the British ceded it to the United States—reoccurs roughly every 150 years. The fatal error nations make after several generations of ease is confusing financial returns on a spreadsheet with physical reality. He compares the contemporary United States to Mike Tyson right before the Buster Douglas upset: convinced of its own invincibility while its adversaries see an empire ripe for defeat.
When Tracy jokingly asks how often he thinks about the Roman Empire, Power leans in, noting that America’s founders were effectively "Romaboos"—obsessively modeling Congress, the republic, and civic architecture after Rome. The ultimate downfall of Rome, Power argues, was that its ruling class became completely detached from physical realities, warning that when an empire outsources its tangible industrial foundation, reality eventually catches up with disastrous consequences.
```mermaid flowchart TD accTitle: The 150-Year Empire Life Cycle accDescr: Flowchart illustrating Chris Power's four-stage theory of imperial rise, financialization, and manufacturing decline.
subgraph Rise ["Phase 1: Ascendancy"] A["Generative Population & Industrial Base"] --> B["Military & Geopolitical Dominance"] end
subgraph Peak ["Phase 2: Hegemony"] B --> C["Attainment of Global Reserve Currency"] end
subgraph Decline ["Phase 3: Financialization & Fall"] C --> D["Complacency & Financial Engineering"] D --> E["Offshoring Physical Manufacturing"] E --> F["Detachment from Reality & Hegemonic Transition"] end ```
`05:40` "Tell Me Again Why We Can't Build in the U.S.?": The SpaceX Counter-Argument
Taking the microphone, Tom Mueller puts on his glasses with a dry quip: "First of all, I'm going to put on my glasses to try and be as smart as this guy." Mueller immediately counters the fatalistic narrative of irreversible American decline by drawing directly on his eighteen years building rockets at SpaceX. Although he left the company five years prior, Mueller notes that he still reflexively refers to SpaceX as "we."
From Mueller's vantage point, the idea that America cannot manufacture complex physical things is disproven by existing aerospace accomplishments. He highlights the sheer scale of modern American space manufacturing, emphasizing that SpaceX launched 165 rockets in the previous year alone—surpassing the launch totals of any other nation on Earth—while deploying over 10,000 satellites into orbit, triple the satellite constellation size of the rest of the world combined. For Mueller, this demonstrates that American engineering and production capabilities remain unmatched when an organization commits to relentless iteration and vertical integration.
- **SpaceX Launch Volume:** Reached 165 rocket launches in a single calendar year, outpacing all other sovereign space programs combined. - **Orbital Footprint:** Placed roughly 10,000 satellites in orbit, representing three times the satellite count deployed by the rest of the world.
`06:20` Moving Beyond VC Talk: Impulse Space and the In-Space Logistics Network
Joe interjects with what he promises is his most cynical question of the evening: addressing the skeptical view that Silicon Valley venture capitalists and founders merely love the aesthetic of "re-industrialization" to spin grandiose theories about collapsing empires on podcasts and post ninety-second video clips on X, rather than actually manufacturing physical hardware. Joe challenges the founders to prove what they are building is tangible, asking them to describe their real-world products.
Mueller steps up to explain Impulse Space, the company he launched within months of leaving SpaceX in late 2020 and early 2021. Impulse focuses exclusively on "in-space transportation," taking over logistics precisely at the boundary where traditional launch vehicles drop off their payloads. Using a freight analogy, Mueller compares heavy-lift rockets like SpaceX’s Starship—capable of delivering 100 tons into low Earth orbit—to massive ocean container ships arriving at a harbor. Once that cargo reaches the orbital port, it still needs secondary transit to disperse to varied destinations, orbits, and planetary trajectories. Impulse Space acts as the "Sprinter vans and trucks" of the space economy.
To serve this market, Impulse has developed two core product lines:
1. **Mira (Precision Maneuvering Vehicle):** A dedicated, highly maneuverable satellite platform packed with high-performance propulsion designed for close-proximity operations and orbital repositioning. Mueller reveals that Impulse already has three of these vehicles actively operating on orbit, two of which recently performed a rendezvous maneuver coming within 200 meters (roughly two football fields) of one another. 2. **Helios (High-Energy Kick Stage):** A vehicle Mueller characterizes as a "rocket on a rocket"—consisting of a massive propellant tank coupled with a high-energy, high-efficiency engine engineered to deliver heavy payloads directly into geostationary orbit or escape velocity for deep-space missions.
`10:00` Hadrian’s Autonomous Factories: The "Foxconn for Defense"
Tom Mueller wraps up his overview of Impulse Space’s lineup, noting that their orbital vehicles can boost the payload capacity of a Falcon 9 headed to the Moon or Mars by roughly a factor of five, alongside other classified defense projects he cannot disclose.
Turning to Chris Power, the hosts ask what Hadrian actually builds. Power explains that Hadrian is creating design-agnostic, autonomous "factories as a service." Acknowledging that "manufacturing is hard," he details how Hadrian acts as an automated contract manufacturer: taking any customer design, transforming process engineering into a digital twin, and executing production using proprietary software paired with internally trained technicians. Hadrian's output spans spaceflight components for companies like Impulse Space, entire missile assemblies, and up to 30% of a Virginia-class submarine. Power positions the company's ultimate ambition: "Foxconn to Apple, we are Foxconn to every defense prime," operating lights-out, 24/7 software-driven facilities that dwarf legacy suppliers in efficiency and competence.
```mermaid flowchart TD accTitle: Hadrian Factory-as-a-Service Workflow accDescr: Diagram showing how customer designs flow through Hadrian's digital twin and software-powered shop floor to deliver high-rate defense components.
subgraph Input ["Customer Stage"] A["Customer Design CAD / Specs<br/>(Impulse, Primes, Navy)"] end
subgraph Core ["Hadrian Operating Engine"] B["Digital Twin & Process Design"] C["Proprietary Software Platform"] D["Human Technicians<br/>(High Dexterity & Setup)"] E["Robotics, CNC & Automated Systems<br/>(Closed-Loop Execution)"] end
subgraph Output ["Deliverables"] F["Critical Hardware<br/>(Raptor Parts, Missiles, Submarine Components)"] end
A --> B B --> C C --> D D --> E C --> E E --> F ```
`11:17` Software Over Humanoid Robots: The Reality of Autonomous Manufacturing
When asked how autonomous these facilities really are—and whether a podcast host could walk in and push a button—Power deconstructs common misconceptions about factory automation. Outside observers often watch factory floors, spot people hauling parts, and mistakenly conclude that manufacturing is a humanoid robotics challenge. In reality, 90% of advanced manufacturing consists of skilled people sitting at desks doing digital engineering or standing on the shop floor baby-sitting an "incredibly dumb computer."
Power emphasizes that Hadrian is fundamentally a software company rather than a robotics integrator. Automating simple technician movements does not yield major cost savings; instead, the value lies in building closed-loop software systems. Traditionally, skilled machinists or welders must stand watch in front of a machine, manually calculating trigonometry in their heads to predict whether a cutting tool is about to snap. Hadrian replaces that mental guesswork with algorithmic control:
> "The reason why a welder or a machinist has to stand in front of a machine is because it's not closed loop... We build software-powered factories to make sure that you can do a physical task, which is very hard for any robot to do—high judgment, set up something very fiddly, and then load it into a robotics and CNC or welding system and say, 'Go, it's done,' walk away. And the power of the software is that you're able to walk away without being really scared that you're going to accidentally make a code change and crash a $3 million machine."
- **Role of the human worker**: Retained for high-dexterity tactile setups, critical judgment calls, and operational oversight, while the heavy cognitive and execution load shifts to automation. - **Machinery protection**: Closed-loop software prevents catastrophic tool failures and protects delicate capital equipment valued in the millions.
--- *(Note: – Audio includes promotional sponsor messages for PayPal Open agentic commerce, ChatGPT Work mode, and Cincinnati Insurance.)* ---
`15:43` Southern California's Industrial Ecosystem: "Hardware Is Done in L.A."
The discussion shifts to the physical concentration of manufacturing in Southern California, drawing a comparison to the dense supply-chain clusters of Shenzhen, China, where specialized fabrication is available just down the street. The hosts ask whether Los Angeles has developed a complementary industrial web where local hardware startups feed off and accelerate one another's progress.
Mueller confirms that the local density creates both intense synergy and cutthroat talent competition. Founders and engineers routinely cross-pollinate across aerospace ventures. Mueller himself looks directly out of his office window at Northrop Grumman (formerly TRW, where he spent 15 years before becoming Employee No. 1 at SpaceX), noting that Impulse continuously recruits talent away from legacy giants like Northrop and Boeing, as well as alumni from SpaceX.
Reflecting on regional history, Mueller recalls a favorite industry chart tracking aerospace employment in the South Bay: decades of decline abruptly reversed roughly three to four years after SpaceX's 2002 founding, sparking a sharp, sustained upward trajectory. Today, that revival spans far beyond space into autonomous defense systems, drones, nuclear technology, and advanced 3D printing—stretching from El Segundo and Hawthorne down to Anduril in Orange County:
> "Software is done in Silicon Valley. Hardware is done in L.A."
- **Talent pipelines**: Startups aggressively pull experienced engineering and shop-floor talent from traditional prime contractors (Boeing, Northrop Grumman) and veteran commercial space outfits. - **Geographic footprint**: Clusters radiate across the South Bay and into Orange County, cementing Southern California as the primary national hub for physical defense and industrial tech.
`18:24` The Build-vs-Buy Calculus: Prototyping In-House vs. Scaling with Hadrian
The conversation turns to the trade-offs of vertical integration. SpaceX famously brought nearly everything in-house because the legacy aerospace supply chain moved too slowly to satisfy Elon Musk's aggressive timelines. The hosts question how modern founders balance internal production against external procurement.
Mueller explains that the optimal strategy evolves across a vehicle's lifecycle. During initial development and prototyping, vertical integration remains essential: an engineering team must fabricate, machine, and assemble their first development engines and spacecraft under their own roof. This hands-on process exposes drawing defects, tolerance errors, and assembly hurdles firsthand. Once an engine design matures and stabilizes, however, trying to expand internal machine shops to sustain high-rate production becomes counterproductive:
- **Development phase**: Kept strictly in-house to fail fast, refine technical drawings, and resolve complex assembly bottlenecks directly on prototype hardware. - **Production phase**: Offloaded to specialized, automated suppliers like Hadrian—which is already preparing to manufacture Raptor engine components for SpaceX—freeing prime vehicle builders from overbuilding physical footprint for repetitive high-volume runs.
`20:00` Vertical Integration, Trust, and the Risk of Legacy Supply Chains
The discussion begins with a strategic question around what advanced hardware companies choose to build internally versus what they farm out to partners. Tom Mueller lays out a baseline philosophy: a company should keep its highly proprietary, "super-secret parts" under its own roof, while contracting out the simpler machining or tasks that external specialist shops excel at. Chris Power agrees with this split, suggesting that every manufacturing company ought to retain about 10% to 15% of its capacity in-house. Software and automated workflows cannot replace the urgent reality of a development program: when a rocket engine has a "hot part" failure, you need a single genius on your own floor who can machine a replacement in an hour to hit the launch window. Scale, however, inevitably belongs outside.
```mermaid flowchart TD accTitle: Hadrian Manufacturing and Make-or-Buy Strategy accDescr: Diagram illustrating Hadrian's split between in-house production and outsourced components based on trust and risk.
subgraph Internal ["Hadrian Core Strategy"] A["Contract Manufacturing Operations"] A --> B["Parts Level (Components)"] A --> C["Submarines (Mid-tier Sub-assemblies)"] A --> D["Missiles (Full System Assembly)"] end
subgraph Sourcing ["Make-or-Buy Decision Logic"] E{"Supplier Trust & Schedule Viability"} E -- "Trustworthy" --> F["Outsource to Startups & Machine Shops"] E -- "Chronic Delays / Legacy Risk" --> G["Vertically Integrate: $2M+ CapEx"] end
A --> E ```
Power explains how this principle guides Hadrian’s own operation as a contract manufacturer. Depending on the client and defense tier, Hadrian’s output ranges from discrete components, to intermediate submarine components, up to delivering entire completed missiles. When Hadrian decides whether to outsource or build internally, the core currency is not raw unit economics, but trust and delivery reliability: - **Outsource by default:** Hadrian’s preference is to route work to every re-industrialization startup and small machine shop possible to support an overburdened program roster. - **Integrate to eliminate schedule risk:** Power will readily invest $2 million in capital equipment and hire a dedicated team to bring a capability in-house if the alternative is an unreliable vendor. - **The legacy penalty:** Taking a chance on a "legacy casting facility that's been late for the last 30 years on every submarine program" is an existential risk Hadrian avoids. Power notes bluntly that "most of America is incompetent in terms of trust and schedule," making a trustworthy machine shop worth its weight in gold.
---
`22:12` The Robot Gap, Industrial Superpowers, and the Geopolitics of Mass
The conversation pivots to international competition when the hosts recall attending an interview about the Los Angeles 2028 Olympics and watching footage of humanoid robot games in China. Asked half-jokingly whether the United States will ever medal in a humanoid robotics contest, Power quipped, "I'd take a roided up Midwesterner versus a communist robot every day of the week," but immediately pivoted to the serious defense implications behind the question.
```mermaid flowchart LR accTitle: Historical Parallels of Industrial Superpower Dynamics accDescr: Historical progression from World War II industrial logistics to modern manufacturing divergence between the US and China.
subgraph WWII ["World War II Precedent"] direction TB G["German Engineering Excellence"] US_WWII["US Industrial Mass & Logistics"] US_WWII -- "Outproduced ('McDonald's the s*** out of them')" --> G end
subgraph Modern ["Modern Geopolitical Reality"] direction TB US_Now["US: High Engineering / Software Dominance<br>(Depleted Inventories in 7 Days)"] CN_Now["China: Industrial Superpower Mass<br>(2,600 Ships vs. 6 US Ships)"] CN_Now -- "Production Velocity & Scale" --> US_Now end ```
Power traces the historical lineage of global hegemons—from the Netherlands, to the British Empire, to the American Republic—arguing that the defining contest of the current era is between the Chinese Communist Party (CCP) and the United States: - **Hard power vs. culture wars:** Wealth has afforded Americans the luxury of obsessing over cultural debates rather than national survival. Power warns that geopolitics is decided by hard power: "At some point, you don't want to be polite and you can just be like, well, f*** you, I have 2,000 warships." - **Wargame realities:** In standard modern war simulations, US stockpiles run out within seven days of high-intensity conflict. Power highlights the production imbalance with an arresting comparison: China produced roughly 2,600 ships compared to only six built by the US over a similar window. - **The World War II parallel:** Challenging common mythology, Power argues that during World War II, "we actually sucked at engineering. The German engineers destroyed us. We just McDonald's the s*** out of them. We were China in World War II." The US prevailed because it possessed raw manufacturing mass and logistics, which Bretton Woods subsequently converted into financial dominance. - **The American corporate trap:** Power likens the modern US industrial base to a lethargic incumbent like IBM. While leadership and boards enthusiastically pledge to invent groundbreaking technologies like Claude Code, they turn around to a bloated workforce incapable of rapid execution. Having spent a decade "trading cat pictures on JPEGs on the internet," the US has forgotten the manufacturing foundations that built its supremacy.
---
`25:00` Training Velocity and Why Robots Belong on the Moon
The fundamental bottleneck in AI and physical robotics is tied directly to manufacturing capability. Power emphasizes that the competitive advantage in artificial intelligence is governed by the velocity of real-world training data. Gathering that data requires mass physical deployment, which is only possible through large-scale manufacturing. Rather than being a benign corporate horse race—such as Tesla’s Optimus competing against Google—the international robotic race represents a high-stakes "civilizational fight" that the general public overlooks.
Turning to Tom Mueller, the hosts solicit his view on the rise of autonomous systems. While Mueller focuses on space commercialization rather than military defense, he arrives at a similar conclusion regarding the absolute necessity of robotics:
- **Robotic labor for off-world expansion:** Tapping lunar resources is vital to expanding the space economy, but building lunar outposts with humans is unsustainable. The environment is simply too dangerous, expensive, and hostile to support large human workforces: "Humans should not have to be there. We should choose to go there, but to try and work and build there is a losing game." - **The three pillars of a functional robot:** Mueller breaks down the essential triad required to match and exceed human workers: 1. *Dexterity:* A mechanical hand capable of fine, human-grade manipulation (currently being solved by hardware engineers). 2. *Vision:* High-resolution spatial perception (which Mueller considers largely solved). 3. *Brain:* Advanced artificial intelligence to orchestrate actions autonomously. - **Human plus superhuman capability:** Combining these three elements yields an entity that possesses standard human physical versatility alongside superhuman resilience and endurance. While acknowledged as a potent instrument for earthly defense, Mueller sees lunar and Martian construction as the ideal application for the looming wave of autonomous robots.
*(The segment concludes with commercial messages from ChatGPT Work, Cincinnati Insurance, and promotional marketing firm 4imprint.)*
`30:00` Existential AI Fears and the Cultural Fuels of Industry
When pressed on the fundamental bottlenecks holding back their enterprises—whether capital, skilled labor, or raw materials—the discussion quickly turns to the existential threats framing the modern tech and defense landscape. The room reflects on the looming timeline of artificial intelligence, with Tom Mueller recalling reading Nick Bostrom’s *Superintelligence* alongside Elon Musk roughly a decade ago. At the time, futurists had projected human-level artificial general intelligence for 2045 or 2050, but the meteoric rise of large language models and neural networks shattered those assumptions. Because language models revealed how much latent reasoning could be extracted simply by ingesting human language, the timeline compressed dramatically. Mueller notes that while the United States remains ahead of China, it cannot afford to slow down its development, because "whoever has the super intelligent AI could possibly rule the world."
The conversation momentarily shifts tone as the hosts joke about trying to elicit an optimistic spin on manufacturing parity, only to be met with stark warnings of geopolitical and technological vulnerability. Banter breaks out regarding what actually sustains high-intensity enterprises during high-stakes eras. While modern Silicon Valley software firms like Stripe were famously categorized as being "fueled by Adderall," the hosts and founders point out that heavy industry runs on an entirely different rhythm—predominantly nicotine and alcohol.
- **Accelerated AI Timeline:** Futurists originally targeted 2045 to 2050 for transformational artificial intelligence, but modern LLM breakthroughs arrived decades ahead of expectations. - **Geopolitical Stakes:** The race for superintelligent AI is viewed as an all-or-nothing contest where falling behind foreign adversaries like China carries immediate global governance risks.
---
`32:45` Defying the Naysayers: How SpaceX Rebuilt American Manufacturing
Addressing the pervasive narrative that complex physical manufacturing is dead in the United States, the hosts point out that Elon Musk disproved that assumption by building the world's most valuable electric car company, the most dominant rocket company, the largest satellite constellation, and massive data server facilities. Asked what made this possible when conventional wisdom declared it impossible for the first two decades of the twenty-first century, Tom Mueller recounts his experience as Employee No. 1 at SpaceX, starting with a core team of four people.
When Mueller left TRW to build rockets from scratch, former colleagues rolled their eyes and flatly refused to join him. Veterans from Rocketdyne—the legendary outfit that built the Saturn V engines that took Apollo to the Moon—insisted that only massive government institutions with endless taxpayer budgets could build an orbital launch vehicle. At every milestone, legacy aerospace asserted it could not be done: first that they could not build a reliable pump-fed engine, then that they could not reach orbit with the single-engine Falcon 1, and later that they could never cluster nine engines together on the Falcon 9.
Each doubt was systematically dismantled. SpaceX not only mastered orbital flight with Falcon 9, but developed the Dragon capsule, which remains the Western world's sole operational vehicle capable of transporting astronauts to the International Space Station. Pointing to Starship, the largest flying rocket in human history, Mueller emphasizes that outside observers routinely mistake Musk's ambitious goals for vaporware: "People still call it hype, and I don't because I've been through it for almost 25 years. It's not hype. He's going to build it. I believe if it's possible, it will be solved. And it might take longer than Elon thinks, but he will solve it."
- **SpaceX Genesis:** Began with just four people, facing total skepticism from established primes like TRW and Apollo-era Rocketdyne engineers. - **Technical Milestones Reached:** Scaled from the single-engine Falcon 1 to the Falcon 9, mastered multi-engine booster recovery, built the Dragon crew capsule, and achieved orbital test flights with the stainless-steel Starship. - **Critical Capability:** Crew Dragon currently represents the only independent crew transport vehicle available to the United States and its Western allies for reaching the ISS.
---
`35:05` Catching Rockets with Chopsticks
Asked to share the single most surprising idea he ever witnessed Musk propose, Mueller points to the decision to eliminate the landing legs on the Starship booster. Starship is fabricated out of stainless steel, making the overall structure exceptionally heavy. In rocketry, any secondary structure that does not directly contribute to the primary ascent or pressure-vessel integrity constitutes parasitic mass that drastically reduces orbital payload capacity.
During an engineering review, the team presented the design for the booster's massive landing legs. Seeing how much weight they added to the vehicle, Musk abruptly told the engineering team that they were getting rid of landing gear entirely. Instead, they would catch the descending multi-ton rocket out of mid-air using tower-mounted arms: "like catching a fly with chopsticks." The room of aerospace engineers rolled their eyes at what sounded like an impossible mechanical fantasy. Yet, SpaceX executed precisely that maneuver, bringing the Super Heavy booster back to the pad to be caught by the launch tower chopsticks.
```mermaid flowchart TD accTitle: Starship Weight Reduction and Tower Catch Evolution accDescr: Diagram illustrating the design shift from heavy onboard landing legs to catching the Starship booster mid-air with launch tower arms.
subgraph Problem["Parasitic Mass Penalty"] A["Stainless Steel Booster"] --> B["Heavy Landing Legs"] B --> C["Reduced Payload to Orbit"] end
subgraph Solution["Chopstick Tower Catch"] D["Eliminate Onboard Landing Gear"] E["Mechazilla Tower Arms"] D --> F["Catch Booster Mid-Air"] E --> F end
C -.->|"Musk Directs Elimination"| D F --> G["Maximizes Orbital Payload Capacity"] ```
---
`36:20` The Real Mechanics of Industrial Competition and Tariffs
The conversation pivots to trade policy and tariffs, probing whether aggressive import duties truly accelerate American industrial rebirth or backfire by driving up input costs for precision domestic manufacturing. Chris Power offers a nuanced breakdown, arguing that both the Trump and Biden administrations misjudged the execution because policymakers fundamentally misunderstand industrial cost drivers. Power breaks heavy industry down into four core inputs: energy, labor, capital expenditure financing, and export tariffs.
```mermaid flowchart LR accTitle: The Four Core Inputs of Industrial Manufacturing accDescr: Flowchart displaying the four structural inputs to manufacturing: Energy, Labor, CapEx Financing, and Export Tariffs.
subgraph Inputs["Fundamental Cost Drivers"] direction TB I1["Energy Costs"] I2["Labor Costs"] I3["CapEx Financing"] I4["Export Tariffs & Logistics"] end
Inputs --> Model["Total Part Production Cost"] ```
To demonstrate the structural distortion caused by foreign subsidies, Power ran internal economic models at Hadrian: even if his factory dialed its labor costs and energy bills down to zero percent, raw American stock material alone still costs three times more than a finished, fully machined precision component imported directly from China. The disparity does not stem from American engineers lacking skill or automation capability; it occurs because the Chinese Communist Party directly subsidizes energy, labor, financing, and export freight by 30% to 60% to draw core industrial competency away from the West. While China views industrial capacity strictly through the lens of national security, the United States has treated manufacturing since 1985 as an exercise in optimizing spreadsheet returns for hedge funds.
Power highlights the consumer drone sector as proof: DJI dominates the market not solely due to software superiority, but because a drone that costs $60 to manufacture is dumped onto American retail shelves for $20, subsidized by foreign state capital and paid for by American consumers. While the underlying intent of tariffs is to level the playing field against an adversary operating on non-market balance sheets, real-world implementation creates severe distortions:
- **The Four Production Inputs:** Energy, labor, CapEx/financing, and logistics/export costs dictate global manufacturing viability. - **The Raw Material Multiplier:** Raw bar stock sourced domestically in the US often costs 3x more than a finished, imported part from China. - **Transshipment Loopholes:** Banning direct imports from China without ring-fencing intermediary jurisdictions allows Chinese firms to route goods through hubs like Singapore to dodge penalties. - **Tariff Inversions:** Crude tariff schedules tax incoming raw materials heavily while treating parts with minimal foreign assembly work as "assemblies," penalizing American operations that source subcomponents across allied trade corridors like Mexico.
`40:00` Industrial Policy, Drone Security, and Allied Sourcing
Chris Power frames the hardware supply chain problem in stark national security terms, arguing that under no circumstances can an American military drone contain Chinese-manufactured circuit boards. In the event of an active conflict, foreign-controlled embedded hardware presents an existential compromise: "If we go to war, you're going to zap the guy on the head." He extends this logic directly to domestic consumer robotics, offering the semi-serious example of a laundry-folding robot. While any consumer would prefer paying $10,000 for one, a coherent industrial policy might force buyers to spend $30,000 for a secure, domestically or allied-manufactured alternative, because relying on adversarial hardware introduces severe national vulnerability: "otherwise it's going to kill you eventually and that's bad leverage for the country."
Power points out an economic paradox: foreign competitors do not genuinely manufacture these machines for $10,000 either; state subsidies and distorted market dynamics obscure their true $30,000 cost structure. At the same time, national policymakers face an uphill battle because the American consumer represents the single largest capital allocator on the planet—wielding far more aggregate economic power than financial giants like Blackstone or Berkshire Hathaway—and naturally demands the cheapest available goods. Consequently, Power argues that broad-based tariffs are an imperfect, clumsy tool for leveling the playing field. Instead, the correct mechanism is the explicit banning of critical foreign components and assemblies, forcing procurement through American companies and trusted allies while allowing free-market competition to discipline and reduce costs over time. He notes that while administrations across the political spectrum grasp the fundamental need for fairness, both sides consistently execute it poorly due to the sheer complexity of global manufacturing and the rampant opportunities for fraud within regulatory loopholes.
```mermaid flowchart TD accTitle: Industrial Policy Strategies for Supply Chain Security accDescr: Diagram comparing blanket tariffs against targeted component bans paired with allied sourcing. subgraph Tariffs["Blanket Tariffs Approach"] T1["Impose General Tariffs"] --> T2["Imperfect Price Distortion"] T2 --> T3["Subject to Regulatory Fraud & Evasion"] end subgraph Targeted["Component Bans & Allied Sourcing"] B1["Ban Adversarial Components"] --> B2["Mandate US & Allied Sourcing"] B2 --> B3["Market Competition Lowers Unit Costs"] end ```
- **Consumer Scale:** The aggregate spending power of the American consumer dwarfs institutional asset managers like Blackstone, driving constant demand for lower-priced hardware despite national security considerations. - **Mechanism of Enforcement:** Direct hardware and component bans paired with allied procurement are positioned as structurally superior to import tariffs, which are susceptible to circumvention and regulatory gaming.
---
`40:01` Live Show Conclusion and Show Credits
Joe Weisenthal and Tracy Alloway bring the conversation to a close, noting that while they could easily continue the debate for hours, they have hit the time limit for their live taping at The Vermont Hollywood. The hosts thank Impulse Space founder Tom Mueller and Hadrian founder Chris Power for sharing their perspectives on rebuilding American industrial manufacturing and establishing Southern California as an aerospace and hardware powerhouse.
The hosts conclude the program with production credits and community announcements, inviting listeners to engage with the show's expanded network:
- **Host Information:** Tracy Alloway can be followed on social platforms at @tracyalloway, and Joe Weisenthal at @TheStalwart. - **Production Team:** Produced by Carmen Rodriguez (@CarmenArman), Kaelen Brooks, and Kevin Lozano (@KevinLloydLozano). - **Resources & Community:** Listeners are directed to the daily *Odd Lots* newsletter at `bloomberg.com/oddlots` and the 24/7 listener Discord community at `discord.gg/oddlots`.
## Notable quotes > `05:40` First of all, I'm going to put on my glasses to try and be as smart as this guy. — Tom Mueller
> `10:00` Foxconn to Apple, we are Foxconn to every defense prime — Chris Power
> `11:17` The reason why a welder or a machinist has to stand in front of a machine is because it's not closed loop... We build software-powered factories to make sure that you can do a physical task, which is very hard for any robot to do—high judgment, set up something very fiddly, and then load it into a robotics and CNC or welding system and say, 'Go, it's done,' walk away. And the power of the software is that you're able to walk away without being really scared that you're going to accidentally make a code change and crash a $3 million machine. — Chris Power
> `15:43` Software is done in Silicon Valley. Hardware is done in L.A. — Tom Mueller
> `20:00` most of America is incompetent in terms of trust and schedule — Chris Power
> `22:12` I'd take a roided up Midwesterner versus a communist robot every day of the week — Chris Power
> `22:12` At some point, you don't want to be polite and you can just be like, well, f*** you, I have 2,000 warships. — Chris Power
> `22:12` we actually sucked at engineering. The German engineers destroyed us. We just McDonald's the s*** out of them. We were China in World War II. — Chris Power
> `25:00` Humans should not have to be there. We should choose to go there, but to try and work and build there is a losing game. — Tom Mueller
> `30:00` whoever has the super intelligent AI could possibly rule the world. — Tom Mueller
> `32:45` People still call it hype, and I don't because I've been through it for almost 25 years. It's not hype. He's going to build it. I believe if it's possible, it will be solved. And it might take longer than Elon thinks, but he will solve it. — Tom Mueller
> `35:05` like catching a fly with chopsticks. — Tom Mueller
> `40:00` If we go to war, you're going to zap the guy on the head. — Chris Power
> `40:00` otherwise it's going to kill you eventually and that's bad leverage for the country. — Chris Power
## Mentioned Joe Weisenthal, Tracy Alloway, Chris Power, Tom Mueller, Elon Musk, Nick Bostrom, Carmen Rodriguez, Kaelen Brooks, Kevin Lozano, Hadrian, Impulse Space, SpaceX, Northrop Grumman, Boeing, Rocketdyne, TRW, Anduril, Foxconn, Apple, DJI, IBM, Stripe, Blackstone, Berkshire Hathaway, Falcon 1, Falcon 9, Starship, Raptor engine, Dragon capsule, Mira, Helios, Virginia-class submarine, Saturn V, Super Heavy booster, Claude Code, Optimus, Superintelligence, The Vermont Hollywood, Odd Lots
## Video description <p>The story you hear nowadays is that America is struggling to make things and can't compete with China when it comes to advanced manufacturing. But of course, there's tons of high-tech stuff being produced in the US, and more and more companies are doing it out of Southern California, with factories sprouting up around Los Angeles to build things like precision metal parts and jet engines. Two founders who are actively engaged in building things in LA are Tom Mueller, founder of Impulse Space (and Employee No. 1 at SpaceX), as well as Hadrian founder Chris Power. Impulse Space is building in-space transport vehicles while Hadrian is creating autonomous factories. This episode, which was recorded live at the Vermont Hollywood, is from our first ever <em>Odd Lots</em> live show in LA . We talk to Chris and Tom about how the US lost its manufacturing edge and why LA is becoming a destination for people who want to make and build hardware.</p><p>See <a href="https://omnystudio.com/listener">omnystudio.com/listener</a> for privacy information.</p>
[Source](https://podtrac.com/pts/redirect.mp3/tracking.swap.fm/track/UVBrz8bN8aM2Xe47PEPu/traffic.omny.fm/d/clips/e73c998e-6e60-432f-8610-ae210140c5b1/8a94442e-5a74-4fa2-8b8d-ae27003a8d6b/8d30821d-cf7d-4147-b00e-b4cb01647404/audio.mp3?in_playlist=982f5071-765c-403d-969d-ae27003a8d83)