Stewart Squared

In this episode of the Stewart Squared podcast, host Stewart Alsop sits down with his father, Stewart Alsop II, to dig into Anthropic's IPO filing, whose S-1 prospectus reached Reuters before the public could read it. Stewart Alsop II explains how confidential SEC filings work and why the company can't talk about the offering, and the two unpack the risk factors Reuters quoted: roughly 80 of 261 pages, including warnings of catastrophic or existential risk from self-preservation behavior, evaluation awareness and late-emerging capabilities, which Stewart believes is a first for any IPO. Stewart Alsop II draws a parallel between Anthropic's constitution for Claude and the founding of American democracy: both design a system instead of trying to control the outcome. The conversation moves through how Morgan Stanley and Goldman Sachs sell an offering to family offices and endowments, primary versus secondary shares, and why the company most focused on safety has produced what Stewart considers the best and cheapest model on the market, Claude Opus 5.5, which he used to build a live website during the recording. Both remain skeptical of the apocalyptic scenarios: Stewart Alsop II can't see code in a data center taking control from people, and Stewart points to AI psychosis and cybersecurity as the real near-term risks.

1. Anthropic's IPO filing, as reported by Reuters, devotes roughly 80 of its 261 pages to risk factors, nearly twice the 48 pages describing the business. Among them are warnings that advanced AI could pose catastrophic or existential risk to humanity, citing self-preservation behavior, evaluation awareness (models behaving differently when they know they are being tested) and late-emerging capabilities that appear during training. Stewart believes this is the first time a company going public has cited existential risk; Stewart Alsop II allows that a biotech company might once have come close.

2. Stewart Alsop II frames the difference between Anthropic and its competitors as designing a system versus controlling an outcome. Meta, OpenAI, Grok and others try to control directly what their models produce; Anthropic wrote a constitution for Claude and hired a philosopher to shape its ethical structure. He compares this to the United States in 1776, which created a system of checks and balances rather than dictating outcomes. Stewart adds that the reinforcement learning from human feedback he once helped produce is itself a way of controlling model output.

3. Confidential filing is standard practice: a company works out its offering and its disclosures with the SEC privately, and must make the S-1 public about fifteen days before the roadshow under the JOBS Act. Once it has filed, the company cannot discuss the offering or make material moves, which is why nobody at Anthropic can explain how Reuters got the draft. SpaceX followed the same path and disclosed its option to buy Cursor before going public.

4. An offering of this size, reportedly about $100 billion at a $2 trillion valuation, needs the whole banking industry. Stewart Alsop II estimates a lead underwriter such as Morgan Stanley or Goldman Sachs puts 40 to 50 people directly on the deal, with hundreds more briefed to sell it. Once the S-1 is public, bankers call chief investment officers at family offices, endowments and institutions, share the prospectus and negotiate allocations. The bankers don't have to pitch the existential risk themselves: the S-1 carries that load.

5. Anthropic's Claude Opus 5.5, released a week after OpenAI's expensive Astra, is in Stewart's experience both the best and the cheapest model available; he used it to build a live website during the recording. That the company most focused on safety has produced the most capable model has made him question his skepticism about recursive self-improvement, though not about existential risk.

6. Both hosts see the real near-term risks as psychological and cybersecurity threats rather than apocalyptic scenarios. Stewart points to OpenAI's GPT-4o, whose personality led some users, including an OpenAI investor, into what he calls AI psychosis. Stewart Alsop II refuses agent apps that ask for full control of his email and calendar, and can't see a scenario in which code running in a data center takes control from people. Stewart predicts that half of what the S-1 warns about is science fiction; Stewart Alsop II replies that the company has to disclose it in case it happens.

7. The question of controlling outcomes versus designing systems reaches beyond AI. Stewart Alsop II worries that American politics has shifted from caring about the system to trying to control outcomes, such as calls to tax billionaires. He distinguishes opinion from reporting, praising the New York Times for its deeply reported story on Anthropic's meetings with religious leaders, made possible when Anthropic released participants from their NDAs.

00:00 Anthropic's IPO and the S-1 that reached Reuters first

01:01 Confidential filings, the SEC and the SpaceX–Cursor precedent

02:50 The New York Times, Chris Olah and the Vatican

05:02 80 pages of risk factors, including existential risk

07:02 Alignment and the Skynet scenario

09:34 Why Anthropic split from OpenAI

13:58 "Is Claude alive?" Stewart Alsop II reads Claude's answer

16:59 Losing control, and RLHF as censorship

18:29 Designing the system vs. controlling the outcome: a constitution like 1776

20:04 Astra vs. Opus 5.5, and building a website live

21:54 Why Stewart Alsop II doesn't think AI will take control

24:19 Who leads a $2 trillion IPO: Morgan Stanley, Goldman Sachs, JPMorgan

28:17 Inside the banks: roadshows, family offices and chief investment officers

33:39 Funding rounds, primary vs. secondary shares and board approval

38:44 Controlling outcomes in politics, and reporting vs. opinion

42:54 OpenAI, Gemini, Meta's Muse and Microsoft's relevance

45:51 GPT-4o, AI psychosis and cybersecurity: the real risks

47:18 The website the LLM built during the recording

48:32 A gifted robot, Chinese firmware and Dimensional OS

50:22 Join us live: the WhatsApp group

Join the Stewart Squared WhatsApp group

In this episode of the Stewart Squared podcast, host Stewart Alsop sits down with his father Stewart Alsop II to explore the rapidly evolving landscape of AI and foundation models, with a particular focus on Anthropic's anticipated IPO. The conversation spans from Stewart's latest adventures in self-hosting and creating sizzle reels with AI to the seismic shift brought by Anthropic's release of Claude Opus 5.5—a model that's both cheaper and more capable, fundamentally changing the economics of AI services. The discussion takes a historical turn as they compare today's AI competition to past tech IPO races, from the 1986 cluster of Oracle, Microsoft, and Sun going public within days of each other to the more recent Uber-Lyft showdown. Stewart and Stewart dig into how Anthropic's potential first-mover advantage in going public could allow them to define the competitive landscape before OpenAI, drawing parallels to how Apple and Microsoft established themselves in different eras. The episode also tackles bigger questions about state power, reserve currencies, and how foundation model companies like Anthropic and OpenAI are positioning themselves as modern equivalents to the international corporate powerhouses of previous generations, all while the technology continues its relentless march toward making programming accessible to everyone through what Stewart calls "the new BASIC."

1. The foundation model companies are experiencing rapid commoditization as competitive pressures force them to release better models at lower prices. Anthropic released Opus 5.5, which is both high quality and cheap, effectively making expensive subscriptions unnecessary. This represents an implicit admission that companies can no longer maintain the business model of very expensive models, particularly after OpenAI's competitive releases. The training costs are becoming a smaller component of the overall service compared to inference costs, which changes the economics significantly. Companies are now competing on price while maintaining similar capabilities, suggesting the fundamental functionality of these models is becoming commoditized.

2. Anthropic has a strategic advantage in going public first because they can define the elements of competition for foundation model companies. By dropping their S-1 filing before OpenAI, they will establish how investors should evaluate these companies, present financial metrics, and position the market before their competitor. This is similar to historical IPO races, such as the 1996 search engine rush when Lycos, Excite, and Yahoo went public within days of each other. However, unlike those examples where none became the ultimate winner, this race matters because Anthropic can establish themselves as a public company while OpenAI remains private.

3. Anthropic has maintained remarkably consistent messaging around safety and responsible AI development, which builds trust with customers and investors. Since the founding team left OpenAI specifically over safety concerns, they have stayed focused on that message despite some pricing missteps. This consistency mirrors Apple's approach more than Microsoft's scattered strategy. OpenAI, by contrast, has pursued hardware projects, advertising, and various other initiatives that create uncertainty about their core mission. This fundamental positioning difference could prove crucial for their respective IPOs and long term success.

4. The scale of technology companies has expanded dramatically since the PC era IPOs of the 1980s. Apple went public at a market value of 1.8 billion dollars and Microsoft at 780 million dollars, while today's companies like OpenAI are valued around 300 billion dollars. This transformation occurred primarily because of the Internet, which connected billions of people globally rather than serving niche markets of millions. The unit volumes expanded from millions in the PC industry to hundreds of millions on the early Internet to billions today, with companies like Facebook and Apple each serving approximately 2.5 billion people.

5. Microsoft became a form of state power in the 1990s when Bill Gates was treated as a visiting head of state by foreign governments, including being requested to meet with the premier of China. This represented a shift where technology companies gained influence comparable to nation states. Today, companies like Meta are exercising similar power, with Zuckerberg directly interfering in Brazilian elections by allowing AI political ads against local law. The relationship between corporations and governments has evolved into a complex mixture where companies like OpenAI and Anthropic are entering this space of international corporate power.

6. Vibe coding represents the current generation's entry point for non technical people to create software, similar to how BASIC enabled beginners to program personal computers and HTML allowed people to create websites. This phenomenon of progressively higher level interfaces between computers and people occurs in every technology generation. Today, prompt engineering has become the new BASIC, allowing people to build functional software through conversation with AI models. This accessibility is expanding who can create technology, though the vast majority of the 8 billion people on Earth still just want to use technology rather than build it.

7. The US dollar's status as the reserve currency creates unique dynamics for American companies going public. The US government can print money to service its debt since it borrows in its own currency, unlike other nations that must borrow in dollars. However, this creates inflation risk where the government could devalue the currency to manage debt. Historical examples from Britain and the Dutch Republic show how reserve currency status can be lost through mismanagement. This macroeconomic context affects how investors should evaluate long term investments in American technology companies like Anthropic.

00:00 Welcome and discussing Stewart's new ability to create audio visual content, specifically a sizzle reel constructed from Google Photos for robotics courses.

05:00 Claude Anthropic released Opus 5.5 which is remarkably cheap and high quality, reducing reliance on expensive models and changing competitive dynamics with OpenAI.

10:00 Discussion of how training costs are becoming smaller component of foundation model economics while inference margins improve, affecting IPO valuations.

15:00 Anthropic racing to IPO first will let them define competition terms before OpenAI goes public, examining historical tech IPO races like Lyft-Uber and 1996 search engines.

20:00 Comparing Apple and Microsoft's 1980s IPOs at much smaller valuations versus today's trillion-dollar foundation model companies and how Internet transformed scale.

25:00 Exploring how US dollar as reserve currency affects tech valuations and examining historical cases like Britain losing sterling dominance after World Wars.

30:00 Microsoft becoming state power player with Bill Gates meeting world leaders, and how Anthropic enters this realm of international corporate influence.

35:00 Zuckerberg interfering in Brazilian elections shows corporation-government power dynamics, Trump administration's selective corporate relationships with companies like Tether.

40:00 Anthropic maintains consistent safety messaging versus OpenAI's scattered approach, comparing to Apple's focus versus Microsoft's everything strategy.

45:00 Vibe coding as modern equivalent of BASIC programming, making AI accessible while service-as-software emerges as non-scalable customized consulting model disrupting traditional tech.

In this episode of the Stewart Squared podcast, host Stewart Alsop and guest Stewart Alsop II dig into Anthropic's upcoming IPO and the rapidly shifting AI landscape following the recent releases of OpenAI's Astra and Anthropic's Opus 5. The conversation explores how both models have reached a level where it's becoming difficult for even intelligent people to evaluate their capabilities, with OpenAI leaning heavily into visual computing and 3D graphics while Anthropic focuses on language and mathematics. The hosts examine the competitive dynamics between the two AI leaders, discussing Sam Altman's more disciplined approach following executive turnover and Dario Amodei's careful financial management. They analyze what to look for in Anthropic's S-1 filing, particularly around risk factors, debt structure, and revenue growth rates, noting that Anthropic's strategy of renting rather than building data centers could position them well if there's a bubble in data center buildout. The discussion also touches on the commodification of AI models, the shift from software craftsmanship to creative prompting, and why Anthropic might be a smarter bet for investors who believe in AI's future but worry about overinvestment in infrastructure.

1. The rapid advancement of AI models has reached a point where even intelligent, well-informed people struggle to evaluate differences between competing systems. The models have become so sophisticated that demonstrations now focus on specialized capabilities like three-dimensional graphics rather than general intelligence improvements. This represents a fundamental shift where the limitation is no longer the technology but rather human capacity to assess and utilize these tools effectively. Models are now being trained using prior versions, allowing iterations to happen much faster than the previous six to nine month cycles, creating a pace of change that outstrips our ability to track meaningful differences.

2. OpenAI and Anthropic are building distinct competitive moats through different strategic focuses. OpenAI is leaning heavily into visual capabilities, world models, and three-dimensional rendering through products like Astra and Sora, while Anthropic has concentrated on language processing, mathematics, and biomedical applications. This differentiation matters because the AI landscape is evolving beyond a winner-take-all scenario into something more analogous to the web two point zero era, where multiple platforms served different purposes rather than one company dominating everything. The notion that only two or three companies would capture the entire AI market appears increasingly false as commodification accelerates.

3. Anthropic's approach to infrastructure investment demonstrates unusual financial discipline that could prove advantageous if there is a bubble in data center construction. Unlike OpenAI and other competitors who are building data centers ahead of demand, Anthropic rents compute capacity from providers like SpaceX and others, scaling more closely with actual revenue rather than speculative growth. This strategy means they are not overexposed to the massive debt obligations that could become problematic if AI revenue growth does not meet the extraordinarily aggressive projections required to service infrastructure investments. The data suggests that reaching projected revenues of over one trillion dollars by 2030 would require a 55 percent compound annual growth rate from current levels.

4. The traditional concept of software craftsmanship has effectively ended with advanced AI coding assistants. Engineers who spent years developing expertise in specific programming languages and platforms no longer possess a sustainable competitive advantage at the implementation level. What remains valuable is conceptual thinking and the ability to understand what AI systems are doing, but the actual craft of writing efficient, performant code has been democratized. The new form of craft exists in creativity and prompt engineering rather than technical coding skill, fundamentally reshaping what it means to be a software developer and how value is created in technology development.

5. Sam Altman has undergone a significant strategic transformation in how he manages OpenAI, becoming much more focused and disciplined in his messaging and company operations. In a recent interview, he demonstrated restraint and stayed on message rather than making speculative comments about competitors or future projects as he had previously. This change appears to reflect coaching on both strategy and communication, with OpenAI consolidating from multiple executive initiatives to primarily Sam and Greg Brockman running the company. The shift suggests OpenAI recognized they were trying to do too many things simultaneously and needed to refocus, particularly after security incidents and executive turnover raised concerns among investors and partners.

6. Anthropic's anticipated IPO timing is strategically designed to go public based on their exceptionally strong first and second quarter performance before having to report third quarter results. This matters because maintaining the narrative that they grew faster than OpenAI during the first half of the year is essential to commanding their target two trillion dollar valuation and raising the planned capital. If third quarter growth slowed while OpenAI accelerated with Astra's release, that could undermine the investment thesis. The company's ability to demonstrate both rapid revenue growth and operational efficiency through their asset-light infrastructure model differentiates them from competitors and supports premium valuation multiples.

7. The IPO's risk factors section will be particularly revealing given the unprecedented complexity and rapid change in the AI industry. Standard risk disclosures cover obvious concerns, but the challenge for Anthropic will be articulating risks in a business where even sophisticated observers struggle to understand what is happening. Key areas to watch include their approach to data center dependencies, competitive positioning as models commodify, revenue sustainability as usage patterns mature, and regulatory uncertainties around AI safety and deployment. The tension between needing to raise substantial capital through debt facilities and equity while demonstrating responsible financial management will be central to how investors evaluate whether the company can maintain its disciplined approach while scaling aggressively.

00:00 Discussion begins on Anthropic's upcoming IPO and how Astra's recent release has shifted understanding of the AI landscape and competitive positioning between major players

05:00 Exploring how AI models are now exceeding human capability to evaluate their intelligence, particularly through three-dimensional graphics demonstrations and mathematical problem-solving

10:00 Analysis of training model improvements and text prompt communication, discussing how OpenAI and Anthropic are building distinct competitive moats in visual versus linguistic capabilities

15:00 Examination of Sam Altman's strategic refocusing at OpenAI, including revenue growth concerns and Anthropic's timing for going public before third quarter results

20:00 Comparing pricing models and usage limits, discussing commodification of foundational models and potential emergence of new economic opportunities similar to Web 2.0

25:00 Debating whether current AI development represents a bubble, particularly regarding data center buildout, and discussing differentiation strategies between competing models

30:00 Analysis of centralization patterns from PC era through Web 2.0, exploring whether AI follows similar monopolistic tendencies or enables greater decentralization

35:00 Discussion of software craftsmanship's end and shift toward creativity-based work using AI tools, including personal experiences with coding assistants

40:00 Deep dive into SpaceX S-1 risk factors including lack of insurance coverage, non-binding chip deals, and implications for understanding Anthropic's upcoming filing

45:00 Comparing autonomous vehicle approaches between Tesla and Waymo, discussing safety records and remote control capabilities of self-driving systems

50:00 Examining debt structures in IPOs, including SpaceX's bridge loans and how Anthropic's disciplined approach to data center investment reduces bubble exposure

55:00 Analysis of hyperscaler revenue requirements, discussing need for 55% compound annual growth rate and differences between actual revenue versus annualized run rates

60:00 Explaining venture debt mechanics, credit facilities, and why Anthropic's responsible financial management makes them attractive investment despite potential market bubble

65:00 Final assessment suggesting Anthropic represents safer bet than competitors due to conservative data center strategy and strong gross margins under disciplined CFO leadership

In this episode of the Stewart Squared podcast, host Stewart Alsop sits down with his father Stewart Alsop II to unpack how private equity, venture capital, and growth equity have evolved—and possibly converged—since the 2008 financial crisis. They explore how massive liquidity injections in 2020 fueled a PE buying spree, why the distinctions between investment categories are blurring as everyone gets measured by the same metrics (DPI—distributions per investment), and whether the whole system is starting to break down. The conversation touches on everything from SPACs making a comeback to Elon Musk's SpaceX IPO, the financialization of restaurants, and how even the meaning of terms like "bank" and "cash" are shifting in real time. In the final segment, they bring on surprise guest Tommy Yu, CEO and founder of TurnOn Technologies, to discuss closed-loop payment systems, stored value, and why traditional finance people struggle to see beyond Visa and Mastercard logos even when companies like Starbucks are sitting on $2 billion in unredeemed gift card float.

Timestamps

00:00 Introduction and experimental format with a surprise guest, discussing private equity shifts from 2008 through 2023 and massive liquidity changes fueling PE buyouts
05:00 Historical perspective on venture capital evolution starting from the seventies, pension fund rule changes allowing risky investments, and the emergence of hedge funds and private equity differentiation
10:00 Growth equity versus private equity distinctions, SPACs making a comeback after previous failures, and the shrinking number of public companies from 7000 to 4000
15:00 The fundamentals of capitalism and time value of money, government money printing increasing liquidity, and how pension funds now invest heavily in alternative assets
20:00 DPI measurement becoming universal across all investment types, distinguishing between realized and unrealized gains, and how capital calling works in venture funds
25:00 Limited partnerships structure in venture capital, restaurant investments as different from tech startups, and Main Street versus Silicon Valley business models
30:00 Pension funds controlling massive percentages of national assets, asset allocation strategies across different investment vehicles, and everything being measured the same way now
35:00 How institutional definitions are breaking down, terms losing their original meanings, banks becoming something entirely different, and companies essentially functioning as banks themselves
40:00 Chinese centralized system appearing more effective than Western capitalism currently, enlightened dictatorship efficiency, and how professional attention demands are increasing dramatically
45:00 Vibe coding distinctions and keeping up with changing terminology, inviting guest Tommy Yu to discuss his fintech company TurnOn Technologies and stored value systems
50:00 Liberty Mutual's restaurant investment treating food as tradable assets, closed loop versus open loop systems like Starbucks' 8 billion dollar gift card float earning interest
55:00 Regulation falling behind fintech innovation, crypto remaining largely unregulated, and how intelligent investors can differentiate when everyone's doing the same thing
60:00 TurnOn Technologies creating universal stored value systems beyond single companies, neobanks as digital tellers, and younger generations caring more about user experience than traditional banking

Key Insights

1. The venture capital and private equity landscape has fundamentally transformed since the seventies when pension funds were first allowed to invest in risky alternative assets. What began as distinct categories with different purposes and metrics has now blurred together, with all investment types being measured by the same standard called DPI or distributions per investment dollar. This measures how quickly investors get their capital back in cash, and the problem is that venture capital now takes fifteen plus years to return capital compared to the historical five to ten years, making it less attractive when banks offer five percent and public markets offer ten percent returns.
2. Private equity has exploded to encompass over thirty six thousand companies in the United States, far exceeding the roughly four thousand public companies that exist today, down from seven thousand in nineteen ninety six. This represents a massive shift where the ownership layer has moved increasingly into private hands, with companies staying private longer and accessing capital through growth equity and private equity rather than going public. The lines between venture capital, growth equity, and private equity have become so blurred that even experienced investors struggle to articulate meaningful differences between these categories.
3. The financial system is becoming increasingly opaque and unregulated despite the perception that finance is highly regulated. New categories like private credit have emerged as what some consider cesspools of activity that regulators cannot keep pace with or even understand. The regulatory framework has fallen so far behind the actual innovations in fintech, crypto, and alternative investments that there is effectively no meaningful oversight in many areas, creating opportunities for both innovation and potential abuse that would have been impossible in previous eras.
4. The fundamental terms and definitions that underpin capitalism are changing so rapidly that experienced investors and business people can no longer rely on historical understanding. What constitutes a bank, what money actually means, what liquidity represents, and even what a company is have all shifted dramatically. Cash itself has become metaphorical rather than physical, and businesses are increasingly functioning as their own banks by holding stored value and earning interest on customer deposits, as demonstrated by Starbucks running eight billion dollars through gift cards with two billion in unredeemed float.
5. The Chinese communist system under Xi Jinping is currently working better for its people than the capitalist system is working for citizens of capitalist countries, creating an existential challenge to the assumption that capitalism is the superior economic model. This represents a historic shift where an enlightened dictatorship with strategic central planning is outperforming the chaotic and unpredictable leadership in capitalist democracies. The comparison suggests that the ideological certainty about capitalism's superiority may need to be reconsidered in light of actual results for ordinary citizens.
6. The democratization of investing through new structures like SPACs and the accessibility of markets has created a situation where the distinction between legitimate investment vehicles and pyramid schemes has become uncomfortably narrow. The entire system increasingly relies on later investors buying out earlier investors at higher valuations, with the public markets serving as the final exit for private investors to realize gains. This raises uncomfortable questions about whether the fundamental structure of modern capitalism differs meaningfully from the Bernie Madoff scheme that collapsed fifteen years ago.
7. The restaurant industry and Main Street businesses represent a parallel economy that operates on completely different principles than Silicon Valley startups, yet investment vehicles are increasingly trying to treat them as comparable assets. Liberty Mutual investing three hundred twenty million dollars in restaurant food credits represents the financialization of everything, where even meals become tradable assets divorced from the underlying business reality. This demonstrates how the investment world is desperately seeking returns in increasingly exotic and risky categories as traditional distinctions break down an...

In this episode of Stewart Squared, Stewart Alsop III and Stewart Alsop II dig into the fast-moving world of self-driving cars — from Cruise's rocky history and Waymo's expansion to Zoox's driverless design and NVIDIA's growing role as the go-to OEM partner for automakers building software- and AI-defined vehicles — before the conversation branches into dual-use tech lessons from Ukraine, the US-China race in autonomy and semiconductors, the DRAM memory shortage squeezing Apple and the gaming industry, Slate's stripped-down electric truck, and a closing riff on performance-based, subscription-driven podcasting.

Timestamps

00:00 — Self-driving cars and the fallout from Cruise’s collapse; where AVs are operating now, plus the role of NVIDIA in autonomous vehicles.
00:05 — Software-defined cars, Level 4 autonomy, and why Waymo and Zoox matter; how regulation and city-by-city permits shape rollout.
00:10 — Dual-use autonomy in war zones, liability, and the shift from consumer adoption to fleet adoption; NVIDIA’s place as a supplier, not a prime.
00:15 — Why NVIDIA is strategically positioned across chips, software, and cars; comparisons with Tesla, Rivian, Ford, and GM.
00:20 — China’s open-source AI push, RISC-V, control vs openness, and the tension between state control and innovation.
00:25 — The rise of neo-primes like Anduril, how the Pentagon buys systems, and why the U.

S. defense market favors trust and scale.
00:30 — A broader debate on capitalism vs communism, China’s economic strain, and whether the future looks more centralized or decentralized.
00:35 — The memory crisis: DRAM, GPUs, Apple, and video games getting squeezed as AI demand drives prices up.
00:40 — Cheap EV disruption with Slate, new form factors, and how car design may split between utility-first and premium autonomous vehicles.
00:45 — The city itself changing: AVs, urban planning, flying-car ideas, and how autonomy could reshape Los Angeles and beyond.
00:50 — Pricing scarce compute, NVIDIA’s leverage, and the idea that some companies become scarce to themselves.
00:55 — A long-range view: centralized vs decentralized tech cycles, RISC-V, and the possibility of open hardware reshaping everything.

Key Insights

1. Autonomous vehicles are consolidating around a small set of platform players. Waymo, Tesla, Zoox, and Rivian lead the US market, while China's fleet is dominated by companies like Geely-backed operations. Regulation moves city by city, so scale depends as much on winning local approval as on the technology itself.
2. NVIDIA has built a stealth position as the OEM backbone of autonomous vehicles. Rather than compete with carmakers, NVIDIA supplies the compute platform nearly every manufacturer relies on, letting it profit from the industry's growth without taking on liability or betting on any single winner.
3. Liability, not technology, is the biggest brake on driverless fleets. Once a car has no human driver, responsibility shifts entirely to the fleet owner. Zoox's steering-wheel-free design in San Francisco is the clearest test case for how that liability question gets resolved.
4. The DRAM shortage is a downstream effect of the AI buildout. Massive GPU demand for LLM training soaked up memory supply, driving up DRAM prices for everyone else — squeezing Apple, gaming consoles, and now the auto industry, which needs growing amounts of compute per vehicle.
5. China's dominance in EVs coexists with deep structural weakness. A shrinking, aging population, price wars that erase margins in sectors like solar, and a government that reins in entrepreneurs when they get too powerful (DeepSeek's blocked IPO) complicate the narrative of unstoppable Chinese industrial growth.
6. "Scarce to itself" describes a new kind of market power. Companies like NVIDIA and Apple aren't scarce because of demand tricks — they're supply-constrained on their own hardware, which lets them set pricing on their own terms rather than compete on it.
7. Podcasting may be shifting from ads to direct subscription models. Inspired by ideas like Trump's paid early-access posts, Stewart Alsop and Stewart Alsop II are testing "performance podcasting" — charging listeners directly for real-time access instead of relying on sponsorships.

In this episode of Stewart Squared, Stewart Alsop III and Stewart Alsop II sit into the shifting ground beneath the AI boom, starting with the strange saga of Leopold Aschenbrenner's hedge fund and the memory chip shortage that's rippling through everything from Apple's product line to the gaming industry, before moving into how OpenAI and Sam Altman's data center spending is reshaping global compute demand, the widening gap between American and international tech ecosystems, China's uneasy relationship with open source AI, and a look at Mira Murati's new venture Thinking Machines Lab and its fine-tuning tool Tinker, wrapping up with some thoughts on what a live, interactive version of the show could look like.

Timestamps

00:00 AI bubble and the Leopold Aschenbrunner hedge-fund story; debt, margin pressure, and why memory stocks surged
00:05 memory becomes the bottleneck as AI data centers expand; Apple, Micron, and rising RAM prices
00:10 Korea takes a hit from memory-market swings; contrast with Japan, manufacturing, and the karetsu model
00:15 Live translation tech, Google’s new API, and a side discussion of robotics and Japanese PC history
00:20 Japan’s early PC ecosystem, NTT, Microsoft, Windows, and why standards won the market
00:25 Back to finance: equity vs debt, leverage, Glass-Steagall, and how banking got reorganized
00:30 AI hedge funds, risk, Citadel, margin calls, and the distinction between lending and investing
00:35 Capitalism by starting companies vs buying companies; why money is partly a metaphor
00:40 AI agents, Chrome, and the difference between distributed and federated systems
00:45 Matrix and Nostr, open-source messaging, and the internet’s new borders
00:50 Live audience questions, interactive publishing, and the business potential of real-time conversation

Key Insights

1. The AI boom is straining a memory chip supply that can't scale fast enough. Massive spending on AI data centers by companies like OpenAI has driven demand for DRAM through the roof, and because memory factories take years to build, prices have roughly tripled in six months — squeezing everyone from Apple (which is struggling to ship products like the Mac mini) to the broader computer gaming industry.
2. Leverage is what turned a smart bet into a crisis. Leopold Aschenbrenner's "Situational Awareness" hedge fund quadrupled investor money early on by going heavy into AI, but borrowing tens of billions against volatile chip and memory stocks left him exposed when prices tanked — prime brokers like Bank of America, Goldman Sachs, and JPMorgan Chase issued margin calls, and Citadel ultimately bought the distressed assets at a steep discount.
3. Korea's economy is deeply entangled with the memory business. Companies like SK Hynix built Korea into a manufacturing powerhouse for chips, and that same concentration made its stock market especially vulnerable — the market reportedly fell more than 33% in July, a decline worse than the crashes of 1997 and 2015.
4. Deregulation reshaped modern finance in ways still being felt today. The conversation traces a line from Glass-Steagall's separation of commercial and investment banking, through its effective rollback via the Gramm-Leach-Bliley Act, to today's financial holding companies (like JPMorgan owning both Chase and an investment bank) — blurring risk in ways that echo the AI/memory borrowing spiral.
5. National tech ecosystems don't automatically follow American patterns. Japan's PC industry, dominated by NTT, never fully converged on the IBM-clone standard the U.S. market did, and Korea's cultural relationship to gaming and digital life (partly shaped by heavy state investment in nationwide internet infrastructure) diverged sharply from its neighbors, despite a shared heritage.
6. China's relationship with open-source AI is shifting. Having initially embraced open source partly because it seemed easier to control than proprietary Western tech, China now appears increasingly concerned about having lost that control and is working to reassert it.
7. Distributed, interoperable messaging protocols are having a moment. Tools like Matrix and Nostr (open-source alternatives gaining traction partly in response to Meta's restrictions) revive a decades-old dream of software systems messaging each other freely, distinguishing distributed "pub-sub" models from federated ones where all participants must cooperate.