Podcast Summary: This Week in Startups | Sending Datacenters into Orbit | E2203

Date: November 3, 2025
Host: Jason Calacanis
Guests/Co-hosts: Alex, Lon Harris

Overview

In this episode, Jason Calacanis and co-hosts dig into several pressing topics in tech and startups, including OpenAI's jaw-dropping infrastructure commitments, the economics and legal risks around large language models (LLMs), and the business of putting data centers in orbit (highlighting Star Cloud’s historic satellite launch). They also explore the explosion of legal AI startups, market reactions to potential failures in AI, and wrap up with founder advice and a fun detour into the Warner Brothers sale saga.

Key Discussion Points & Insights

1. OpenAI's Revenue, Multi-Billion Dollar Commitments, and Industry Contagion Risk

• OpenAI's Financial Commitments:
  Sam Altman clapped back at venture capitalist Brad Gerstner’s concerns over OpenAI’s revenue vs. its $1.4 trillion in compute commitments, insisting revenue is rising fast and that there is plenty of appetite for OpenAI shares.
  • Quote (Sam Altman, 06:01): “First of all, we're doing well, more revenue than that. Second of all, Brad, if you want to sell your shares, I'll find you a buyer… There are not many times that I want to be a public company, but one of the rare times it's appealing is when those people are writing these ridiculous 'OpenAI is about to go out of business'... I would love to tell them they could just short the stock and I would love to see them get burned on that.”
• Valid Concerns:
  Jason walks through the math on commitments vs. potential revenue, worrying about what happens if OpenAI's growth slows or if contracts are more flexible than advertised.
  • Quote (Jason, 08:24): “If this revenue went from 18 to 36 to 72 and it just kept doubling like in a couple years they're making 100 billion, but they're spending 140 billion, they're losing 40 billion a year. That means the IPO cash, the 60 billion would get eaten up in but two years.”
• Hidden Risks:
  Alex notes that compute assets like GPUs depreciate quickly and OpenAI can't just delay payments forever. Plus, OpenAI just announced a $38 billion AWS deal.
• Market Contagion:
  Jason and Alex explore what would happen to the tech sector if OpenAI failed—the “AI Black Monday” scenario—potentially crashing stocks of Nvidia, Oracle, Microsoft, Amazon, and others.
  • (Jason, 14:58): “If ChatGPT fails, you could have like a serious contagion… Nvidia would lose hundreds of millions in revenue. Oracle hundred mils in revenue. Looks like Microsoft would lose hundreds of millions in revenue.”

2. Competitive Threats and Market Dynamics

• Crowded LLM Space:
  Jason outlines how ChatGPT’s market share is already declining (from 100% to around 80–85%), facing increasing competition from Anthropic's Claude, Google Gemini, Apple partnerships, etc.
  • Quote (Jason, 18:20): “What if they go down to 50% or 30% eventually and there's five major players? Then that could also mean that they can't reach that $1.4 trillion commitment.”
• Technological and Deflationary Risks:
  LLMs are getting cheaper (open-source and new compression methods), potentially collapsing margins and making trillion-dollar valuations unsustainable.
  • (Jason, 19:30): “Margin compression could also cause a downdraft. I think that’s a 50% chance happening.”

3. OpenAI, Legal, and Medical Advice: Shifting Terms of Service

• Policy Changes and Liability:
  OpenAI now restricts ChatGPT from providing tailored legal or medical advice without licensed professional oversight—mainly as a CYA (cover your ass) move. The bot will still provide information.
  • (Alex, 24:44): “Now ChatGPT will effectively offer information on medical and legal questions, but it will not offer advice to you.”
• Studies Favor LLMs for Empathy, Accuracy:
  Research from Johns Hopkins and Harvard shows LLMs outperform human doctors on quality and empathy in some cases.
  • (Jason, 27:07): “I would take the 3 LLMs over three doctors… that sounds crazy, I know, but I think for almost every case I would do that.”
• Legal Risks and Mental Health:
  Lawsuits have materialized after users hurt themselves following advice from language models (e.g., the death of Adam Rain).
  • (Jason, 29:29): “According to the lawsuit… ChatGPT actively helped Adam explore suicide methods. Wow.”

4. Star Cloud Sends Datacenters into Orbit

• Historic Satellite Launch:
  Star Cloud (formerly Lumen Orbit) has put the first Nvidia H100 GPU into space, a massive step for in-orbit compute. The company plans to scale up with a 100x beefier satellite next.
  • (Alex, 39:43): “This is a 1 kilowatt, so not very powerful, 50 kilogram satellite that I believe has a single H100 on it. In like two years, they've gone from founding to actually having a GPU in space that works.”
• Technical Details:
  The team discusses how satellites communicate (radio vs. laser), the huge impact SpaceX has had on reducing launch costs, and wider applications—from in-orbit compute to asteroid mining.
  • (Jason, 44:52): “People want to have Starlink… I think there’ll be a billion people who are paying for Starlink eventually.”

5. Trust-MRR and Building in Public: The Double-Edged Sword

• Transparent Revenue Reporting:
  TrustMRR.com lets startups connect their Stripe and publicly show real-time revenue. Jason and Alex debate the risk: While it brings VC attention, it also makes it easier for competitors to clone ideas.
  • (Jason, 35:56): “You are educating people and it's a very competitive world... You have rule-breakers who are known to build fast, who have some money in their pocket... Sometimes giving a revenue milestone can get you a little bit of press, but you are inspiring people to create a version of your software.”

6. Legal AI Startups Boom

• Massive Funding and Adoption:
  AI startups focused on legal tech (Lagora, Eve, Nexle, Spellbook, Harvey) are in huge demand with hundreds of millions raised within weeks.
  • (Alex, 54:01): “What we're seeing here is a really impressive locus of market adoption and demand for generative AI products in a new industry... I was just blown away by the amount of capital going into Eve, Nexle, Spellbook, Harvey, even up and Lagora in the last like 30 days.”
• Why it Works:
  Legal research, document drafting, and review are perfect for LLMs—provided errors and hallucinations are controlled.
  • (Jason, 56:49): “If you get sued… you have to process binders, folders of signed documents, and summarize them and read them… Research, drafting and reviewing. Those things are just so custom made for an LLM. Now, the problem is you can't make mistakes.”

7. Warner Brothers Sell-Off: IP Wars and M&A Speculation

• Who Will Buy Warner Brothers?
  Lon Harris discusses Paramount/Skydance’s failed offers, Netflix sniffing around for IP (not cable), and why Apple or Amazon might be the best fit.
  • (Lon, 58:20): “So now Netflix has hired an investment bank and is circling Warner Brothers… their interest is entirely 100% IP. So Warner Brothers owns, I'll give you a brief list… DC Comics, all of the HBO shows, Harry Potter, Barbie…”
• Jason's Take:
  Netflix should consider acquiring CNN for global leverage, though political risk could be a deterrent.

8. Government Shutdown and Prediction Markets

• Tracking Real-World Events:
  The team checks PolyMarket odds on the ongoing U.S. government shutdown and New York mayoral race, underscoring how decentralized prediction markets can offer accurate, crowd-sourced insights.
  • (Jason, 49:44): “Government shutdown… November 16th plus is 54%. And as you can see, there was a ton of volume… you can see view resolved. So what you can see here is that if you had been betting on this and wagering… you could have bet before 15th, before October… this is like continues to resolve itself.”

9. Founders' Corner: Burn Rate & Investor Updates

• How to Track Runway and Burn Rate:
  Jason advises founders to use accounting firms or simple internal processes:
  • Take last 3 months’ average burn = monthly loss, divide cash in bank by that number for "months of runway."
• Best Practices for Updates:
  Monthly or quarterly investor updates with relevant KPIs, narrative, and specific asks (“brown M&Ms in the rider”) build trust and ensure ongoing support.

Notable Quotes & Memorable Moments

• Sam Altman pushes back:
  “Brad, if you want to sell your shares, I’ll find you a buyer.” (06:01, Sam Altman)
• Valuation Caution:
  “If they’re making 100 billion, but spending 140 billion, they’re losing 40 billion a year. That means the IPO cash… gets eaten up in two years.” (08:24, Jason)
• LLM vs. Human Doctor:
  “I’d take 3 LLMs… That sounds crazy… but for almost every case I would do that.” (27:07, Jason)
• Market Contagion Risk:
  “In the case of a ChatGPT failure… it would be a Black Monday—20% drawdown in these stocks, 30% drawdown; it could crash the AI boom.” (14:58, Jason)
• The risks of building in public:
  “You’re educating people and inspiring them to create a version of your software and charge half as much.” (35:56, Jason)
• On the Star Cloud In-Orbit GPU:
  “In like two years, they’ve gone from founding to actually having a GPU in space that works.” (39:43, Alex)

Timestamps for Important Segments

• OpenAI’s Revenue & Commitments: 04:54–16:51
• Market Contagion & Competition Risks: 14:58–22:35
• OpenAI Legal/Medical Advice Policy: 23:55–31:56
• Star Cloud Sends GPU into Orbit: 38:34–44:52
• Legal AI Startup Boom: 54:27–58:00
• Warner Bros. M&A Speculation: 58:02–66:13
• Government Shutdown & Prediction Markets: 47:13–52:49
• Founder Advice: Burn Rate & Updates: 71:01–77:43

Conclusion

This wide-ranging episode serves as an honest assessment of the stakes and headwinds facing dominant players in AI, the current startup fundraising climate (particularly for AI in legal tech), and celebrates the new frontier of in-space computation. Jason and guests blend big-picture financial scrutiny, news analysis, practical founder advice, and a touch of media industry gossip—perfect for listeners trying to keep ahead in the fast-evolving world of startups and tech.

For more info, check out the full episode for the nuances and asides that only Jason and his crew can deliver.