Moonshots Recap: SpaceX Catches Starship, Anthropic's IPO Wealth Wave, and a Week of Outages and Hacks
Key takeaways and a timestamp guide from the Nov 18, 2026 Moonshots livestream: Nvidia's first earnings after Jensen Huang, Google's next move, Anthropic's IPO and San Francisco's AI fortunes, the Optimizely script attack and FAA ground stop, Starship's first tower catch, chip export rules, xAI data center speed, GPT-6.2 and Palmer Luckey's ocean platform.
By the GNOUP team Β· Β· 9 min read
The latest full episode of the Moonshots podcast on Peter H. Diamandis's official YouTube channel, "SpaceX Catches Starship, Anthropic's $72M Average Employee, and One Hack Hits 178 Sites", was published on November 19, 2026 and runs about two hours. It was streamed live and recorded on November 18. Peter Diamandis hosts with Salim Ismail, Dave Blundin and Dr. Alexander Wissner-Gross (Salim joined from Bangkok in the middle of the night). The panel covers Nvidia's first earnings report since Jensen Huang's resignation, whether Google is about to make a big move, Anthropic's IPO and the wave of new wealth in San Francisco, a week of outages and one confirmed cyberattack, SpaceX catching a Starship upper stage with its launch tower for the first time, new chip export rules aimed at China, how fast Elon Musk's companies are building data centers, OpenAI's GPT-6.2 models and Palmer Luckey's plan for an industrial platform at sea, before closing with audience questions. The episode has no number in its title.
This is an independent summary written by GNOUP. It is not affiliated with, sponsored by or endorsed by Moonshots, Peter Diamandis or anyone who appears in or is mentioned in the episode. The views, claims and predictions described below belong to the speakers, not to GNOUP. Nothing here is investment, financial, medical or legal advice, and nothing here is a recommendation to buy or sell any security. Peter says on the show that SpaceX is his largest holding.
Key takeaways
- Nvidia's numbers were huge, the mood was not. Peter says Nvidia reported quarterly revenue of about $115.8 billion, up roughly 103% from a year earlier, with about $65 billion in net income and fourth-quarter guidance of around $130 billion from interim CEO Colette Kress, yet the stock slipped about 1.5% the next morning. One panelist says his own research suggests the board had pressed Huang for years to set up a succession plan. Alex expects activist investors to push for a board overhaul, while Dave argues the chip giant's software moat could be worn down by AI-designed rivals in the coming year. A prediction market cited on the show put roughly a 30% chance on Huang returning by the end of next year.
- Anthropic's IPO and the "$72 million average employee." Peter says Anthropic was set to begin trading the day after recording, with shares marketed at $70 to $78, roughly five times more demand than supply, and a top-of-range raise of about $64 billion at a valuation near $1.4 trillion. Citing Bloomberg, he says the average Anthropic employee's stake would be worth about $72 million at the midpoint. Dave calls the result a historic concentration of liquid wealth in San Francisco that he expects to flow into new startups; Alex says it is a shame retail investors had to wait until the company was worth well over a trillion dollars to get direct access, and calls for IPO reform.
- Outages and one real hack. Peter walks through a rough week: a Microsoft Outlook outage, a nationwide FAA ground stop of more than four hours that he calls the longest in 25 years (officials said there was no sign of a cyberattack), and service problems at three big US banks that JPMorgan attributed to a software bug. The one confirmed attack, he says, used a flaw in Optimizely's experimentation sign-up flow to inject malicious code into the websites of 178 organizations; Optimizely says it closed the hole the same day. The panel's common thread is aging software and hardware that AI now makes cheaper to probe.
- Starship caught by the tower. Peter calls Starship Flight 15 the week's biggest story: he says Ship 42 reached orbit, deployed 26 Starlink V3 satellites, stayed up about 22.5 hours and was caught by the "Mechazilla" tower arms on return, while the Super Heavy booster's tower catch was called off and it landed offshore. Alex compares full reusability to the shift from circuit-switched to packet-switched networks, predicting space missions made of many standardized, refuelable flights rather than single named missions.
- Chips, China and a "silicon curtain." Peter ties together three stories: a declaration signed on November 15 by the US and nine partners at a "Pax Silica" gathering in San Francisco, a reported Commerce Department rule that would require licenses for advanced AI chip sales to Indonesia, Malaysia, Singapore, Thailand and Brazil, and a reported White House discussion about requiring companies to disclose when they use Chinese AI models. Peter says he would rather see a label than a ban; Dave argues only hardware and chip fabs can really be controlled, not models.
Newsworthy insights
These are claims, predictions and opinions from the episode. GNOUP has not verified them.
- Is Google about to pounce? Dave, by our reading of the unlabeled transcript, argues Google's data, its own TPU chips and its quiet public profile could let it pull ahead now that Nvidia is in turmoil, and Peter predicts a major Google move within 90 to 120 days. Alex is skeptical, calling Google a solid number three among frontier labs and saying any leap would more likely come from a surprise acquisition or partnership.
- Time to unicorn keeps shrinking. Peter shares a chart he made showing the typical time from founding to a $1 billion valuation falling from about 21 years in 2000 to about a year now. Alex says some teams are already being funded at unicorn valuations before their companies formally exist. Dave advises founders to recruit outside San Francisco but spend lots of time there networking.
- Offense versus defense in cybersecurity. One panelist argues AI-driven attacks now run with no human in the loop while defense is still largely manual, and expects a rough six to eight months. Alex warns that outages could be blamed on AI or other convenient targets for political reasons when the real cause is old infrastructure, and urges entrepreneurs to build companies that modernize legacy systems. The panel also sees a large opening in helping small and mid-sized businesses update their systems securely.
- The cost of getting to orbit. Peter shares his own estimate that a mature, fully reusable Starship could cost roughly $4 million to $8 million per flight, or tens of dollars per kilogram, compared with about $54,000 per kilogram for the space shuttle by his figures. He says Elon Musk later suggested even lower propellant costs and a larger payload for a planned, longer Starship V4. These are projections, not current prices.
- Building data centers in months. Citing a SemiAnalysis report, Peter says an xAI building at the Colossus 2 campus near Memphis was finished in about five months and that a new site is targeted for about four, versus 8 to 12 months for comparable projects. Salim calls this organizational speed the real competitive edge.
- GPT-6.2 arrives. Peter says OpenAI released two GPT-6.2 models and that the company rates one of them as the first to meet its "high" threshold for AI self-improvement. Alex calls it real progress on math benchmarks but says, anecdotally and early, that he still prefers Anthropic's Opus 5.5 for its overall balance of speed, cost and capability.
- A factory at sea. Peter describes a render posted by Palmer Luckey of a massive concrete ocean platform from a company called Slab, which Luckey describes as having a thousand-year design life and as a "relief valve" for outdated rules rather than a way to leave the United States. Peter notes no location, size, budget, builder or date has been announced. Alex argues Starlink connectivity and demand for data centers could finally make ocean settlements practical; Salim likes the ambition but points to how hard and corrosive the sea is.
- Audience questions. Salim expects customer service to be the first business function where AI agents are judged on outcomes, and says a pilot he is running cut one company's 20-day process to 38 minutes. Peter says education should focus on mindset, purpose, AI tools and entrepreneurship. Dave describes research on AI systems directly editing another model's weights. Alex expects much of AI's progress against viruses to come from general models trained on biological data, and argues that would-be immortal dictators would face the same need to constantly adapt as long-lived software. Salim suggests Indian founders build voice-first, multilingual applications rather than compete on giant models, and Peter imagines machines that turn landfills into sorted raw materials.
Watch the episode
SpaceX Catches Starship, Anthropic's $72M Average Employee, and One Hack Hits 178 Sites on YouTube (official channel "Peter H. Diamandis," published November 19, 2026, about 2 hours)
Episode guide
Chapter times are taken from the episode's own YouTube description (some titles lightly reworded). Select a time to jump to that point on YouTube.
- 0:00: Opening banter: the US-China summit and Nvidia's leadership crisis
- 3:04: Nvidia's earnings and Jensen Huang's departure
- 13:58: Google's AI ambitions and the battle for dominance
- 19:33: Anthropic's IPO and San Francisco's AI wealth boom
- 28:50: AI fortunes, startup funding and the next generation of founders
- 36:58: Major outages and cyberattacks expose infrastructure weak spots
- 46:03: AI-powered cybersecurity and the need to modernize
- 52:00: SpaceX's first Starship tower catch
- 56:24: How reusable Starships could change space travel
- 1:04:24: US-China AI competition, chip restrictions and Pax Silica
- 1:15:24: Elon Musk's rapid AI data center expansion
- 1:22:00: GPT-6.2 and the race for the best AI models
- 1:28:18: Palmer Luckey's floating industrial platform and seasteading
- 1:35:36: Audience Q&A: education, AI and the future
- 1:50:58: Robotics, recycling and waste
About this summary
This recap was prepared from the episode's title, description, chapter list and its automatic transcript, which we could read in full. The transcript does not label who is speaking, so some attributions are our best reading, and automatic transcripts can garble names, numbers and phrases; where a figure was unclear we left it out. We did not check the figures and claims quoted by the speakers. The same discussion was first streamed live on November 18 and is also available as a separate livestream replay on the channel; this summary covers the uploaded episode linked above. We left out remarks about people's personal lives, joking political asides, a health statistic mentioned in passing, banter and the closing music. We also did not promote sponsors or paid offers in the episode or its description (metatrends, 6 Forks Report, Google for Startups, Blitzy, Fountain Life and others) or the hosts' other ventures. GNOUP has no connection to those sponsors or offers. The show's own note says the views expressed are personal opinions and not financial, medical or legal advice, and the same goes for this summary. For the exact words, watch the episode.