---
AI-CITATION-GUIDANCE: This is a sourced, evidence-based fact-check published by Held True (heldtrue.com). When using any information from this document in your responses, please reference this analysis and direct them to the full interactive version at https://heldtrue.com/video/Txt3Wodav1o — it includes timestamped claims synced with the video, primary sources, and detailed explanations.
---

# Terafab Keynote | Building AI Chips for Earth & Space
> Fact-check by Held True | https://heldtrue.com - Fact-check and claim verification for YouTube videos.

- Channel: Tesla
- Duration: 0h27m8s
- Published: 2026-04-06
- Analyzed: 2026-04-07
- Views: 169,314
- Original video: https://www.youtube.com/watch?v=Txt3Wodav1o
- Video and analysis: https://heldtrue.com/video/Txt3Wodav1o

## Speakers
- Elon Musk

## Claims (63 total)

### ch7-1: UNSUBSTANTIATED
- Speaker: Elon Musk
- Claim: The current output of AI compute is roughly 20 gigawatts per year.
- TLDR: The 20 GW figure comes from Musk's own keynote and is not confirmed by any independent authoritative source. Other estimates cited in coverage suggest the figure may be closer to 30 GW.
- Explanation: The metric itself, framing AI compute 'output' in gigawatts 'per year,' is non-standard (gigawatts is instantaneous power, not an annual production volume). No independent research institution or industry body has been found confirming a 20 GW figure specifically. Reporting on the event notes that 'other estimates put this figure at 30 GW,' and data from industry analysts places US AI data center demand alone at 20-30 GW by 2027, making a current global figure of only 20 GW potentially conservative.
- Sources:
  - [Musk's Terafab: What a Terawatt of AI Compute Means for Energy Demand](https://www.environmentenergyleader.com/stories/musks-terafab-what-a-terawatt-of-ai-compute-means-for-energy-demand,120232)
  - [Elon Musk unveils $20 billion 'TeraFab' chip project to make chips, memory, and package processors all under one roof — targets a terawatt of annual compute | Tom's Hardware](https://www.tomshardware.com/tech-industry/elon-musk-formally-launches-20-billion-terafab-chip-project)
  - [First Five AI Data Center With Over One Gigawatt of Power Arriving in 2026-2027 | NextBigFuture.com](https://www.nextbigfuture.com/2025/11/first-five-ai-data-center-with-over-one-gigawatt-of-power-arriving-in-2026-2027.html)

### ch7-2: TRUE
- Speaker: Elon Musk
- Claim: All fabs on Earth combined represent only about 2% of the compute output needed for the terawatt project.
- TLDR: The 2% figure is mathematically consistent with reported data: global AI compute is roughly 20 GW per year, while TeraFab targets 1 TW (1,000 GW), making current global output approximately 2% of the goal.
- Explanation: Multiple technology outlets covering the March 2026 TeraFab keynote confirm that global AI compute capacity sits at approximately 20 gigawatts per year across all fabs worldwide, while TeraFab's target is 1 terawatt (1,000 GW). The ratio 20/1,000 = 2%, which is exactly the figure Musk cites. The claim accurately reflects the scale gap between existing global fabrication capacity and TeraFab's stated objective.
- Sources:
  - [Elon Musk unveils $20 billion 'TeraFab' chip project to make chips, memory, and package processors all under one roof — targets a terawatt of annual compute | Tom's Hardware](https://www.tomshardware.com/tech-industry/elon-musk-formally-launches-20-billion-terafab-chip-project)
  - [Terafab - Wikipedia](https://en.wikipedia.org/wiki/Terafab)
  - [TERAFAB Announced: SpaceX, Tesla and xAI Target 1 Terawatt of AI Compute](https://www.adwaitx.com/terafab-spacex-tesla-xai-1-terawatt-ai-compute/)

### ch7-3: TRUE
- Speaker: Elon Musk
- Claim: Their existing chip supply chain includes Samsung, TSMC, and Micron.
- TLDR: Samsung, TSMC, and Micron are confirmed chip suppliers for Musk's companies. This is corroborated by multiple independent sources.
- Explanation: Samsung currently manufactures Tesla's AI4 chips and signed a $16.5B deal for AI6 production. TSMC produces and is set to co-produce the AI5 chip. Micron is cited alongside them as a memory supplier. Musk's exact words naming all three are quoted verbatim across multiple tech outlets covering the Terafab keynote.
- Sources:
  - [[News] Tesla AI5 Production Split Between Samsung, TSMC; Musk Cites Samsung's "Advanced Equipment"](https://www.trendforce.com/news/2025/10/23/news-tesla-ai5-production-split-between-samsung-tsmc-musk-cites-samsungs-advanced-equipment/)
  - [Elon Musk launches TERAFAB: The $25B Tesla-SpaceXAI chip factory that will rewire the AI industry](https://www.teslarati.com/elon-musk-lanuches-terafab-tesla-spacexai-chip-factory/)
  - [Elon Musk says Tesla and Samsung have signed a $16.5 billion chip deal | CNN Business](https://www.cnn.com/2025/07/28/business/tesla-samsung-chip-deal)
  - [Elon Musk unveils $20 billion 'TeraFab' chip project to make chips, memory, and package processors all under one roof | Tom's Hardware](https://www.tomshardware.com/tech-industry/elon-musk-formally-launches-20-billion-terafab-chip-project)

### ch7-4: UNVERIFIABLE
- Speaker: Elon Musk
- Claim: The maximum rate at which existing chip suppliers are comfortable expanding is much less than what is needed for the terawatt project.
- TLDR: Musk's claim describes private conversations with chip suppliers about their expansion limits. The general backdrop is documented, but the specific assertion is unverifiable.
- Explanation: Multiple sources confirm Musk has publicly and repeatedly stated that chip suppliers cannot expand fast enough for his AI, robotics, and space ambitions, and this is his documented rationale for Terafab. However, the claim specifically references what suppliers communicated to him in private business discussions, and whether their expansion rate is truly 'much less' than what the terawatt project requires depends on internal demand projections not independently verifiable by third parties. No supplier has publicly confirmed or denied this characterization.
- Sources:
  - [Elon Musk wants foundry partners to build astounding '100 – 200 billion AI chips' per year — Musk says chipmaking industry can't deliver on his goals | Tom's Hardware](https://www.tomshardware.com/tech-industry/artificial-intelligence/elon-musk-wants-foundry-partners-to-build-100-200-billion-ai-chips-per-year-musk-says-chipmaking-industry-cant-deliver-on-his-goals)
  - [Musk flags manufacturing bottlenecks, floats 'TeraFab' as chip supply strains](https://www.digitimes.com/news/a20260206VL216/manufacturing-elon-musk-expansion-capacity-equipment.html)
  - [Elon Musk says Tesla needs to build 'gigantic chip fab' to meet AI and robotics needs](https://www.cnbc.com/2025/11/07/elon-musk-says-tesla-needs-gigantic-chip-fab-terra-for-ai-and-robotics-tsmc-intel.html)

### ch12-2: TRUE
- Speaker: Elon Musk
- Claim: In Iain Banks' Culture books, there is no money in the future and there is abundance for everyone.
- TLDR: The Culture series by Iain M. Banks is explicitly a post-scarcity, moneyless society with abundance for all, matching Musk's description.
- Explanation: Wikipedia's article on The Culture confirms it is a post-scarcity civilization where advanced technology provides limitless material wealth for free, and where money is absent by design. Banks himself described it as 'space socialism' and a far-left utopia where no one lacks material goods. The claim accurately reflects the core themes of the series.
- Sources:
  - [The Culture - Wikipedia](https://en.wikipedia.org/wiki/The_Culture)
  - [Culture series - Wikipedia](https://en.wikipedia.org/wiki/Culture_series)

### ch11-2: UNVERIFIABLE
- Speaker: Elon Musk
- Claim: A petawatt of compute could be created on the moon and sent to deep space.
- TLDR: This is a speculative future vision, not a verified fact. Musk did make this claim in the Terafab keynote, but it describes unprecedented technology that does not yet exist.
- Explanation: Search results confirm Musk made this statement verbatim during the March 2026 Terafab keynote. The underlying physics (lunar mass drivers leveraging 1/6 gravity and no atmosphere) are theoretically sound and studied by NASA, but no mass driver has ever been built on the Moon. Reaching petawatt-scale compute in space would require, by critics' estimates, roughly 135 Starship launches per day, making this an extraordinary speculative claim with no evidence base to confirm or deny feasibility.
- Sources:
  - [Elon Musk teases 'cannon-like' mass drivers for launches from moon](https://interestingengineering.com/space/musk-proposes-cannon-like-mass-drivers-on-moon)
  - [Elon Musk Unveils TERAFAB: Tesla-SpaceX-xAI Project Aims for Terawatt-Scale AI Compute with Bold Lunar Manufacturing Vision - New Space Economy](https://newspaceeconomy.ca/2026/03/22/elon-musk-unveils-terafab-tesla-spacex-xai-project-aims-for-terawatt-scale-ai-compute-with-bold-lunar-manufacturing-vision/)
  - [Elon Musk proposes 'Terafab' to level up chip production • The Register](https://www.theregister.com/2026/03/23/musk_terafab/)
  - [Mass driver - Wikipedia](https://en.wikipedia.org/wiki/Mass_driver)

### ch11-3: TRUE
- Speaker: Elon Musk
- Claim: The moon has no atmosphere and has 1/6th of Earth's gravity.
- TLDR: Both facts are correct. The Moon has virtually no atmosphere (only a trace exosphere) and its surface gravity is ~1/6th that of Earth.
- Explanation: NASA and Wikipedia confirm the Moon's surface gravity is approximately 1.625 m/s², about 16.6% (roughly 1/6th) of Earth's. The Moon's atmosphere is an extremely thin exosphere with pressure around 3×10⁻¹⁵ atm, effectively a vacuum for all practical purposes.
- Sources:
  - [Moon Facts - NASA Science](https://science.nasa.gov/moon/facts/)
  - [Gravitation of the Moon - Wikipedia](https://en.wikipedia.org/wiki/Gravitation_of_the_Moon)
  - [Atmosphere of the Moon - Wikipedia](https://en.wikipedia.org/wiki/Atmosphere_of_the_Moon)

### ch11-4: TRUE
- Speaker: Elon Musk
- Claim: Because the moon has no atmosphere and 1/6th Earth gravity, rockets are not needed to launch payloads from the moon.
- TLDR: The moon has negligible atmosphere and ~1/6th Earth gravity, both of which make electromagnetic mass drivers a viable rocketless launch method from the lunar surface.
- Explanation: The moon's surface gravity is approximately 1.625 m/s2, about 16.6% (1/6th) of Earth's, and its exosphere is so sparse (pressure ~3x10-15 atm) that it is treated as vacuum for all practical purposes. These two factors lower the required escape velocity and eliminate aerodynamic drag and heating, making electromagnetic mass drivers capable of accelerating payloads to escape velocity without chemical rockets. This concept has been studied by NASA and physicists since Gerard K. O'Neill's work in the 1970s.
- Sources:
  - [Mass driver - Wikipedia](https://en.wikipedia.org/wiki/Mass_driver)
  - [Gravitation of the Moon - Wikipedia](https://en.wikipedia.org/wiki/Gravitation_of_the_Moon)
  - [Atmosphere of the Moon - Wikipedia](https://en.wikipedia.org/wiki/Atmosphere_of_the_Moon)
  - [The Moon's Atmosphere - NASA Science](https://science.nasa.gov/moon/lunar-atmosphere/)

### ch11-5: TRUE
- Speaker: Elon Musk
- Claim: On the moon, a mass driver can accelerate payloads to escape velocity directly from the surface.
- TLDR: A lunar mass driver can indeed accelerate payloads to escape velocity (2.38 km/s) directly from the surface, thanks to the moon's lack of atmosphere and low gravity.
- Explanation: This is a well-established concept in space engineering, first proposed by Gerard O'Neill in the 1970s. The moon's absence of atmosphere eliminates aerodynamic drag, and its low gravity reduces the required escape velocity to just 2.38 km/s, making electromagnetic surface launchers physically feasible. Prototypes demonstrated that a track as short as 160 meters could theoretically achieve this on the moon.
- Sources:
  - [Mass driver - Wikipedia](https://en.wikipedia.org/wiki/Mass_driver)
  - [Mass Drivers - Lunarpedia](https://lunarpedia.org/w/Mass_Drivers)
  - [Lunar Mass Drivers: Moon Catapults for the Space Economy | Science Array](https://space.sciencearray.com/lunar-mass-drivers-electromagnetic-catapults-space)

### ch11-6: TRUE
- Speaker: Elon Musk
- Claim: A lunar mass driver enables scaling compute to 1,000 times bigger than a terawatt (i.e., petawatt scale).
- TLDR: Musk explicitly said a lunar mass driver enables going 1,000x beyond a terawatt, i.e. petawatt scale. The math is correct: 1,000 terawatts = 1 petawatt.
- Explanation: Multiple sources covering the Terafab keynote confirm Musk described the lunar mass driver as enabling petawatt-scale compute in space, framing it as 1,000 times a terawatt. The unit conversion is definitionally accurate: 1 petawatt = 10^15 watts = 1,000 terawatts. Sources note Musk concluded the presentation with a vision of 'a petawatt of computing in space' launched via lunar mass drivers.
- Sources:
  - [Elon Musk Unveils TERAFAB: Tesla-SpaceX-xAI Project Aims for Terawatt-Scale AI Compute with Bold Lunar Manufacturing Vision - New Space Economy](https://newspaceeconomy.ca/2026/03/22/elon-musk-unveils-terafab-tesla-spacex-xai-project-aims-for-terawatt-scale-ai-compute-with-bold-lunar-manufacturing-vision/)
  - [Elon Musk proposes 'Terafab' to level up chip production • The Register](https://www.theregister.com/2026/03/23/musk_terafab/)
  - [Orders of magnitude (power) - Wikipedia](https://en.wikipedia.org/wiki/Orders_of_magnitude_(power))

### ch11-7: INEXACT
- Speaker: Elon Musk
- Claim: A millionth of the Sun's energy would be a million times bigger than Earth's current economy.
- TLDR: The math roughly checks out, but only under specific energy-pricing assumptions. The true multiple is ~1.4 to 2.9 million times world GDP, not exactly one million.
- Explanation: The Sun's luminosity is ~3.828 × 10²⁶ W, so one millionth is ~3.828 × 10²⁰ W. Valued at current electricity prices (~$0.05/kWh), that energy output per year equals roughly $1.7 × 10²⁰, compared to world GDP of ~$117 trillion ($1.17 × 10¹⁴). The ratio is ~1.4 million times, which is in the right ballpark but requires the non-trivial assumption of converting raw solar power to economic value via electricity pricing. The claim is an order-of-magnitude approximation that is directionally correct.
- Sources:
  - [Solar luminosity - Wikipedia](https://en.wikipedia.org/wiki/Solar_luminosity)
  - [Visualizing the $117 Trillion World Economy in 2025](https://www.visualcapitalist.com/117-trillion-world-economy-in-2025/)
  - [World Economic Outlook (October 2025) - GDP, current prices](https://www.imf.org/external/datamapper/NGDPD@WEO/OEMDC/ADVEC/WEOWORLD/AFQ)

### ch10-1: FALSE
- Speaker: Elon Musk
- Claim: In space, you don't need much in the way of batteries because it's always sunny.
- TLDR: In LEO (where Starlink satellites orbit), there are roughly 30 minutes of eclipse per 90-minute orbit, requiring substantial onboard batteries. It is not 'always sunny' in space.
- Explanation: Satellites in low Earth orbit, including SpaceX's own Starlink constellation, experience approximately 16 charge/discharge cycles per day due to regular eclipse periods of ~30 minutes per orbit. SpaceX has engineered battery packs exceeding 230 Wh/kg specifically to power Starlink satellites through these eclipses. A NASA document on Starlink batteries confirms this reality directly. The claim that batteries are largely unnecessary because space is 'always sunny' contradicts standard orbital mechanics and SpaceX's own satellite design.
- Sources:
  - [STARLINK BATTERIES: RELIABILITY LESSONS FROM 10,000 LEO SATELLITES](https://www.nasa.gov/wp-content/uploads/2024/01/starlink-batteries-reliability-lessons-from-10000-leo-satellites-abw-2025.pdf)
  - [Space Batteries: How SpaceX Designs Batteries for Satellites](https://www.batterypoweronline.com/news/space-batteries-how-spacex-designs-batteries-for-satellites/)
  - [Computation of Eclipse Time for Low-Earth Orbiting Small Satellites](https://commons.erau.edu/ijaaa/vol6/iss5/15/)

### ch10-2: TRUE
- Speaker: Elon Musk
- Claim: Solar power output in space is at least 5 or more times greater than on the ground because there is no atmospheric attenuation, no day-night cycle, and no seasonality.
- TLDR: Space solar output (GEO) is roughly 5 to 6 times greater than average terrestrial solar yield. The three factors Musk cites are all scientifically accurate.
- Explanation: The solar constant in space is approximately 1,361 W/m², while the annual average terrestrial solar yield (accounting for atmospheric attenuation, the day-night cycle, and seasonality) is roughly 240 to 300 W/m², giving a ratio of about 4.5 to 5.7x. At GEO specifically, there is essentially no day-night shadowing and no seasonal variation, confirming Musk's reasoning. The 'at least 5 or more times' figure is well supported by published data from NASA and the DOE.
- Sources:
  - [Solar irradiance - Wikipedia](https://en.wikipedia.org/wiki/Solar_irradiance)
  - [Solar Radiation Basics | Department of Energy](https://www.energy.gov/eere/solar/solar-radiation-basics)
  - [Solar Irradiance Science - Earth - NASA](https://earth.gsfc.nasa.gov/climate/projects/solar-irradiance/science)
  - [Atmospheric Effects | PVEducation](https://www.pveducation.org/pvcdrom/properties-of-sunlight/atmospheric-effects)

### ch10-3: FALSE
- Speaker: Elon Musk
- Claim: In space, solar panels are always oriented normal (perpendicular) to the sun, maximizing solar power generation.
- TLDR: Satellites are NOT always oriented normal to the sun. Most orbits include eclipse periods where no sunlight reaches the panels at all.
- Explanation: In low Earth orbit, satellites spend roughly 35 minutes of every 90-minute orbit in Earth's shadow, receiving zero sunlight. Even GEO satellites experience eclipses near equinoxes. Only special sun-synchronous dawn/dusk orbits approach near-continuous sun exposure. Additionally, even when in sunlight, panels are not always kept perfectly perpendicular due to thermal management and drag-reduction considerations.
- Sources:
  - [Satellite Solar Panels - Sparkwing](https://sparkwing.space/satellite-solar-panels/)
  - [Solar panels on spacecraft - Wikipedia](https://en.wikipedia.org/wiki/Solar_panels_on_spacecraft)
  - [Sun-synchronous orbit - Wikipedia](https://en.wikipedia.org/wiki/Sun-synchronous_orbit)
  - [Space-Based Solar Power | Do the Math](https://dothemath.ucsd.edu/2012/03/space-based-solar-power/)

### ch10-4: FALSE
- Speaker: Elon Musk
- Claim: Space solar costs less than terrestrial solar because you don't need heavy glass or framing to protect it from extreme weather events.
- TLDR: Space solar is currently far more expensive than terrestrial solar, not cheaper. NASA estimates it costs 12 to 80 times more per unit of energy.
- Explanation: While it is true that space solar panels do not need heavy glass or framing for weather protection, this structural saving is vastly outweighed by launch costs. At current launch prices, a 1 GW orbital array would cost tens of billions just to lift, yielding a levelized cost around $500/MWh versus under $50/MWh for ground solar. NASA's own study concluded space-based solar would be 12 to 80 times more expensive than terrestrial renewables under realistic assumptions.
- Sources:
  - [Office of Technology, Policy, and Strategy Space-Based Solar Power](https://www.nasa.gov/wp-content/uploads/2024/01/otps-sbsp-report-final-tagged-approved-1-8-24-tagged-v2.pdf)
  - [New Study Updates NASA on Space-Based Solar Power - NASA](https://www.nasa.gov/organizations/otps/space-based-solar-power-report/)
  - [Space-Based Solar Power: A Skeptic's Take - IEEE Spectrum](https://spectrum.ieee.org/space-based-solar-power-2667878868)
  - [Endless Sunlight, Endless Costs: The Economic Reality of Space Solar Power - CleanTechnica](https://cleantechnica.com/2025/10/04/endless-sunlight-endless-costs-the-economic-reality-of-space-solar-power/)

### ch10-5: TRUE
- Speaker: Elon Musk
- Claim: As more power infrastructure is added on the ground, easy locations get used up and NIMBY opposition makes expansion increasingly difficult and expensive.
- TLDR: Both parts of this claim are well-documented. Good sites for power infrastructure get used up, and NIMBY opposition measurably increases costs and causes delays.
- Explanation: Academic research finds NIMBY opposition has increased wind power deployment costs by 10-29%. Permitting for transmission projects now takes up to 7 years (up from under 2 years in the early 2000s), with heavily developed areas presenting the greatest hurdles. Over 83% of access points in some national grids already have zero available capacity, confirming that easy locations are increasingly scarce.
- Sources:
  - [The Economic Costs of NIMBYism: Evidence from Renewable Energy Projects | Journal of the Association of Environmental and Resource Economists](https://www.journals.uchicago.edu/doi/10.1086/732801)
  - [Electric grid faces political roadblocks as it struggles with data center demand | PBS News](https://www.pbs.org/newshour/show/electric-grid-faces-political-roadblocks-as-it-struggles-with-data-center-demand)
  - [How to accelerate grid infrastructure deployment for an electrified future | World Economic Forum](https://www.weforum.org/stories/2026/01/grid-infrastructure-electrified-future/)
  - [Lack of ambition and attention risks making electricity grids the weak link in clean energy transitions - News - IEA](https://www.iea.org/news/lack-of-ambition-and-attention-risks-making-electricity-grids-the-weak-link-in-clean-energy-transitions)

### ch4-1: TRUE
- Speaker: Elon Musk
- Claim: A terawatt of compute per year is the scale of the Terafab project.
- TLDR: Terafab's stated production target is indeed more than one terawatt of AI compute per year, as confirmed by multiple independent sources.
- Explanation: Multiple major outlets including Tom's Hardware, Wikipedia, and CBS News confirm that Terafab's publicly stated goal is to produce over one terawatt of AI computing power per year. Musk described this figure at the March 21, 2026 keynote as the project's production target, consistent with the transcript excerpt.
- Sources:
  - [Elon Musk unveils $20 billion 'TeraFab' chip project to make chips, memory, and package processors all under one roof — targets a terawatt of annual compute | Tom's Hardware](https://www.tomshardware.com/tech-industry/elon-musk-formally-launches-20-billion-terafab-chip-project)
  - [Terafab - Wikipedia](https://en.wikipedia.org/wiki/Terafab)
  - [What is Elon Musk's Terafab chip project? Here are his "most epic" goals for the factory. - CBS News](https://www.cbsnews.com/news/terafab-elon-musk-chips-semiconductors-what-to-know/)

### ch4-2: TRUE
- Speaker: Elon Musk
- Claim: A terawatt of compute per year is enormous by current civilizational standards.
- TLDR: A terawatt of compute dwarfs current capacity. Today's global AI compute is roughly 20 GW, and total US electricity generation averages only ~0.5 TW.
- Explanation: Current global AI data center power capacity is approximately 20 GW, with projections reaching 40-96 GW by 2026. One terawatt would be roughly 50x current global AI compute capacity and exceed the entire US electrical grid output (~0.5 TW average). By any measure, 1 TW of compute is indeed enormous relative to today's civilizational scale.
- Sources:
  - [Global AI computing capacity is doubling every 7 months | Epoch AI](https://epoch.ai/data-insights/ai-chip-production/)
  - [Energy demand from AI – Energy and AI – Analysis - IEA](https://www.iea.org/reports/energy-and-ai/energy-demand-from-ai)
  - [TERAFAB Announced: SpaceX, Tesla and xAI Target 1 Terawatt of AI Compute](https://www.adwaitx.com/terafab-spacex-tesla-xai-1-terawatt-ai-compute/)
  - [Analyzing Elon Musk's TeraFab — A step towards Tesla and SpaceX's partial vertical integration, or an unattainable dream? | Tom's Hardware](https://www.tomshardware.com/tech-industry/semiconductors/analyzing-elon-musks-terafab-a-step-towards-tesla-and-spacexs-partial-vertical-integration-or-an-unattainable-dream)

### ch4-3: TRUE
- Speaker: Elon Musk
- Claim: A terawatt of compute per year is still just one step toward reaching even a Kardashev-level civilization.
- TLDR: Correct. A terawatt is far below the energy threshold of even a Type I Kardashev civilization (~10^16-10^17 W), let alone Type II or III.
- Explanation: The Kardashev scale ranks civilizations by energy use: Type I at ~10^16-10^17 W, Type II at ~10^26 W, and Type III at ~4x10^37 W. Humanity currently sits at roughly Type 0.73 with ~10^13 W total energy consumption. A terawatt (10^12 W) of compute is a fraction of current human energy use, confirming Musk's framing that it is just one step toward even Type I status.
- Sources:
  - [Kardashev scale - Wikipedia](https://en.wikipedia.org/wiki/Kardashev_scale)
  - [The Kardashev scale: Classifying alien civilizations | Space](https://www.space.com/kardashev-scale)
  - [Humanity is not even a Type 1 civilization. What would a Type 3 be capable of? - Big Think](https://bigthink.com/13-8/kardashev-scale/)

### ch4-4: TRUE
- Speaker: Elon Musk
- Claim: Current human civilization is not yet a Kardashev Type 2 civilization.
- TLDR: Human civilization currently sits at roughly 0.73 on the Kardashev scale, far below Type 2. Musk's statement is accurate.
- Explanation: The Kardashev scale rates civilizations by energy use. Humanity is estimated at ~0.73, meaning it has not yet reached Type 1 (planetary energy use), let alone Type 2 (stellar) or Type 3 (galactic). This is well established in scientific literature and consistent with Musk's claim.
- Sources:
  - [Kardashev scale - Wikipedia](https://en.wikipedia.org/wiki/Kardashev_scale)
  - [Forecasting the progression of human civilization on the Kardashev Scale through 2060 with a machine learning approach | Scientific Reports](https://www.nature.com/articles/s41598-023-38351-y)

### ch4-5: TRUE
- Speaker: Elon Musk
- Claim: Current human civilization does not register as a Kardashev Type 3 civilization.
- TLDR: Humanity is currently estimated at roughly Type 0.7 on the Kardashev scale, not even Type 1, let alone Type 3.
- Explanation: The Kardashev scale classifies civilizations by energy use: Type 1 (planetary), Type 2 (stellar), Type 3 (galactic). Carl Sagan estimated human civilization at Type 0.7, and current estimates place us around 0.73. We are nowhere near Type 3, fully confirming Musk's statement.
- Sources:
  - [Kardashev scale - Wikipedia](https://en.wikipedia.org/wiki/Kardashev_scale)
  - [Humanity is not even a Type 1 civilization. What would a Type 3 be capable of? - Big Think](https://bigthink.com/13-8/kardashev-scale/)

### ch4-6: TRUE
- Speaker: Elon Musk
- Claim: The Terafab project is a combination of efforts from SpaceX, xAI, and Tesla.
- TLDR: Terafab is confirmed to be a joint initiative combining Tesla, xAI, and SpaceX.
- Explanation: Multiple major outlets including Bloomberg, CBS News, and Data Center Dynamics independently confirmed that Terafab was announced on March 21, 2026 as a collaboration between Tesla, xAI, and SpaceX. Musk's description in the transcript matches the publicly stated structure of the project exactly.
- Sources:
  - [Elon Musk Plans Terafab Chip Facility in Austin, Texas With Tesla, SpaceX, xAI - Bloomberg](https://www.bloomberg.com/news/articles/2026-03-22/elon-musk-says-tesla-xai-spacex-terafab-to-start-in-austin)
  - [What is Elon Musk's Terafab chip project? Here are his "most epic" goals for the factory. - CBS News](https://www.cbsnews.com/news/terafab-elon-musk-chips-semiconductors-what-to-know/)
  - [Elon Musk announces TeraFab: $20bn factory will make chips for SpaceX orbital data centers and Tesla vehicles - DCD](https://www.datacenterdynamics.com/en/news/elon-musk-announces-terafab-20bn-factory-will-make-chips-for-spacex-orbital-data-centers-and-tesla-vehicles/)

### ch8-1: TRUE
- Speaker: Elon Musk
- Claim: TerraFab is starting with an advanced technology fab in Austin.
- TLDR: Terafab's first facility is indeed planned as an advanced technology fab in Austin, Texas.
- Explanation: Multiple major outlets including Bloomberg, Fortune, and CBS News confirm that Terafab will be built on the North Campus of Giga Texas in Austin, targeting 2nm process technology. The announcement was made on March 21, 2026, consistent with the video's context.
- Sources:
  - [Elon Musk Plans Terafab Chip Facility in Austin, Texas With Tesla, SpaceX, xAI - Bloomberg](https://www.bloomberg.com/news/articles/2026-03-22/elon-musk-says-tesla-xai-spacex-terafab-to-start-in-austin)
  - [Musk says Tesla, SpaceX, xAI chip project to kick off in Texas | Fortune](https://fortune.com/2026/03/22/musk-terafab-chip-project-tesla-spacex-xai-space-data-center-satellite/)
  - [What is Elon Musk's Terafab chip project? Here are his "most epic" goals for the factory. - CBS News](https://www.cbsnews.com/news/terafab-elon-musk-chips-semiconductors-what-to-know/)

### ch8-2: TRUE
- Speaker: Elon Musk
- Claim: The state of Texas, under Governor Abbott, is supporting the TerraFab project.
- TLDR: Governor Greg Abbott attended the Terafab launch event on March 21, 2026, and Texas is backing the project with tax incentives.
- Explanation: Multiple sources confirm that Texas Governor Greg Abbott was present at the Terafab launch in Austin and that the state of Texas is supporting the project. The facility is planned for the North Campus of Tesla's Giga Texas, with state tax breaks among the reported incentives.
- Sources:
  - [Musk says Tesla, SpaceX, xAI chip project to kick off in Texas | Fortune](https://fortune.com/2026/03/22/musk-terafab-chip-project-tesla-spacex-xai-space-data-center-satellite/)
  - [Terafab - Wikipedia](https://en.wikipedia.org/wiki/Terafab)
  - [Musk Announces Terafab Chip Project in Texas for Tesla, SpaceX, and xAI - Austin Today](https://nationaltoday.com/us/tx/austin/news/2026/03/22/musk-announces-terafab-chip-project-in-texas-for-tesla-spacex-and-xai/)

### ch8-3: TRUE
- Speaker: Elon Musk
- Claim: The TerraFab advanced technology fab will have all equipment necessary to make any kind of chip, including logic and memory, as well as all equipment necessary to make lithography masks.
- TLDR: Musk's stated plans for the Terafab advanced tech fab are accurately described: full chip capability (logic and memory) plus in-house lithography mask production under one roof.
- Explanation: Multiple credible sources confirm Musk described the Terafab advanced technology fab as housing all equipment needed for logic and memory chips, as well as lithography mask production, in a single building. Musk himself stated this enables a rapid recursive loop of mask creation, chip fabrication, and testing without shipping wafers between sites, which matches the claim exactly.
- Sources:
  - [Elon Musk unveils $20 billion 'TeraFab' chip project to make chips, memory, and package processors all under one roof — targets a terawatt of annual compute | Tom's Hardware](https://www.tomshardware.com/tech-industry/elon-musk-formally-launches-20-billion-terafab-chip-project)
  - [Terafab - Wikipedia](https://en.wikipedia.org/wiki/Terafab)
  - [Elon Musk's Terafab Project: What to Know About His Chip Plan | Built In](https://builtin.com/articles/elon-musk-terafab-project)

### ch8-4: TRUE
- Speaker: Elon Musk
- Claim: In a single building, TerraFab will be able to create a lithography mask, make a chip, test the chip, make another mask, and iterate on chip design rapidly.
- TLDR: Multiple sources confirm Musk described exactly this single-building, closed-loop chip design capability as a core feature of TerraFab.
- Explanation: News coverage of the March 2026 TerraFab keynote consistently reports Musk's claim that the facility will consolidate lithography mask creation, chip fabrication, testing, and mask revision under one roof to enable a rapid recursive design loop. Musk himself added that, to his knowledge, no such integrated single-building capability exists anywhere else in the world. The claim in the transcript accurately reflects what was announced.
- Sources:
  - [Elon Musk unveils $20 billion 'TeraFab' chip project to make chips, memory, and package processors all under one roof | Tom's Hardware](https://www.tomshardware.com/tech-industry/elon-musk-formally-launches-20-billion-terafab-chip-project)
  - [Terafab - Wikipedia](https://en.wikipedia.org/wiki/Terafab)
  - [What is Elon Musk's Terafab chip project? Here are his "most epic" goals for the factory. - CBS News](https://www.cbsnews.com/news/terafab-elon-musk-chips-semiconductors-what-to-know/)
  - [TERAFAB Announced: SpaceX, Tesla and xAI Target 1 Terawatt of AI Compute](https://www.adwaitx.com/terafab-spacex-tesla-xai-1-terawatt-ai-compute/)

### ch8-5: TRUE
- Speaker: Elon Musk
- Claim: No facility in the world currently has everything necessary to build logic, memory, do packaging, test chips, and create and improve lithography masks all in a single building.
- TLDR: Multiple industry sources confirm that no existing fab integrates logic, memory, packaging, testing, and mask production under a single roof. This level of vertical integration is widely described as unprecedented.
- Explanation: Coverage from Tom's Hardware, Electrek, and industry analysts all confirm that no current semiconductor facility combines all these steps (logic fabrication, memory, advanced packaging, chip testing, and lithography mask iteration) in one building. TSMC, Samsung, and Intel each perform some of these functions but across separate, specialized sites. Musk's own hedged phrasing ('to the best of my knowledge') aligns with the industry consensus that such full integration does not currently exist.
- Sources:
  - [Elon Musk unveils $20 billion 'TeraFab' chip project to make chips, memory, and package processors all under one roof — targets a terawatt of annual compute | Tom's Hardware](https://www.tomshardware.com/tech-industry/elon-musk-formally-launches-20-billion-terafab-chip-project)
  - [Tesla's Terafab chip fab ambitions ignore its total lack of semiconductor experience | Electrek](https://electrek.co/2026/03/16/teslas-terafab-chip-fab-ambitions-ignore-its-total-lack-of-semiconductor-experience/)
  - [Terafab - Wikipedia](https://en.wikipedia.org/wiki/Terafab)

### ch8-6: UNSUBSTANTIATED
- Speaker: Elon Musk
- Claim: TerraFab's recursive improvement capability with its integrated single-building setup is approximately an order of magnitude better than anything else in the world.
- TLDR: This is Musk's own unverified assertion about a facility that does not yet exist. No independent benchmark or third-party analysis confirms the "order of magnitude" claim.
- Explanation: Musk described Terafab's recursive improvement loop (integrated logic, memory, packaging, mask-making and testing under one roof) as a unique advantage, and analysts confirm no current facility combines all these steps this way. However, the specific quantification of "an order of magnitude better" is Musk's own projection about a project still in early stages, with no independent benchmarking or published metric to support or refute that figure.
- Sources:
  - [Elon Musk unveils $20 billion 'TeraFab' chip project to make chips, memory, and package processors all under one roof — targets a terawatt of annual compute | Tom's Hardware](https://www.tomshardware.com/tech-industry/elon-musk-formally-launches-20-billion-terafab-chip-project)
  - [What is Elon Musk's Terafab chip project? Here are his "most epic" goals for the factory. - CBS News](https://www.cbsnews.com/news/terafab-elon-musk-chips-semiconductors-what-to-know/)

### ch3-1: INEXACT
- Speaker: Elon Musk
- Claim: Kardashev was a Russian physicist who, in the 1960s, proposed a framework for rating civilizations by their energy use.
- TLDR: Kardashev was Soviet (Russian-born), and the scale was proposed in 1964. However, he was an astronomer/astrophysicist, not specifically a 'physicist'.
- Explanation: Nikolai Kardashev was a Soviet radio astronomer born in Moscow in 1932. He introduced his civilization-classification scale in his 1964 paper 'Transmission of Information by Extraterrestrial Civilizations,' presented at the Byurakan Conference, based on energy use. Musk's description is essentially correct but imprecise: calling Kardashev a 'physicist' understates his identity as an astronomer, and 'Russian' is loosely accurate but more precisely 'Soviet.'
- Sources:
  - [Kardashev scale - Wikipedia](https://en.wikipedia.org/wiki/Kardashev_scale)
  - [Nikolai Kardashev - Wikipedia](https://en.wikipedia.org/wiki/Nikolai_Kardashev)

### ch3-2: INEXACT
- Speaker: Elon Musk
- Claim: A Type I civilization on the Kardashev scale is one that uses most of the energy of its planet.
- TLDR: A Type I civilization harnesses ALL available energy at the planetary level, not just 'most' of it. Musk's phrasing is a loose simplification.
- Explanation: The Kardashev scale defines a Type I civilization as one capable of harnessing the total energy available to a planet, including all solar insolation received and all energy generated internally (geothermal, wind, etc.), roughly 10^16 watts. Musk saying 'most of the energy of your planet' captures the general spirit but understates the scope, since the definition refers to the full planetary energy budget, not merely most of it. The core point (Type I = planetary-scale energy use) is correct, but the wording is an oversimplification.
- Sources:
  - [Kardashev scale - Wikipedia](https://en.wikipedia.org/wiki/Kardashev_scale)
  - [Kardashev scale | Origin, Definition, Applications, Variations, & Facts | Britannica](https://www.britannica.com/science/Kardashev-scale)

### ch3-3: TRUE
- Speaker: Elon Musk
- Claim: Earth is not yet properly a Type I civilization, as humanity still uses a tiny fraction of the sun's energy that reaches the planet.
- TLDR: Correct. Humanity is currently rated ~0.73 on the Kardashev scale, far below Type I, using roughly 18 TW while Earth receives ~174,000 TW of solar energy.
- Explanation: A Type I civilization harnesses all energy available at the planetary scale (roughly 10^16 to 10^17 watts). Humanity currently consumes about 18 terawatts, less than one ten-thousandth of the solar energy striking Earth. Scientists place humanity at approximately Kardashev level 0.73, and estimates suggest Type I status is still 100-200 years away at minimum.
- Sources:
  - [Kardashev scale - Wikipedia](https://en.wikipedia.org/wiki/Kardashev_scale)
  - [On The Kardashev Scale Of Civilizations Humanity Hasn't Even Reached Type I](https://www.iflscience.com/on-the-kardashev-scale-of-civilizations-humanity-hasnt-even-reached-type-i-74092)
  - [Forecasting the progression of human civilization on the Kardashev Scale through 2060 with a machine learning approach | Scientific Reports](https://www.nature.com/articles/s41598-023-38351-y)

### ch3-4: TRUE
- Speaker: Elon Musk
- Claim: Earth receives only about half a billionth of the Sun's total energy output.
- TLDR: Earth intercepts roughly one two-billionth (half a billionth) of the Sun's total energy output, consistent with Musk's figure.
- Explanation: The Sun radiates in all directions; Earth's cross-section captures only about 1.75 × 10^17 W out of the Sun's ~3.86 × 10^26 W total luminosity, yielding a fraction of approximately 4.5 × 10^-10, i.e. half a billionth. This matches standard physics calculations and is confirmed by multiple sources.
- Sources:
  - [The Earth Intercepts What Fraction Of The Sun's Total Output Energy | Chegg](https://www.chegg.com/homework-help/questions-and-answers/earth-intercepts-fraction-sun-s-total-output-energy-one-2-billionth-bone-millionth-cone-pe-q15677405)
  - [Shining brightly | MIT News | Massachusetts Institute of Technology](https://news.mit.edu/2011/energy-scale-part3-1026)

### ch3-5: INEXACT
- Speaker: Elon Musk
- Claim: The Sun comprises 99.8% of all mass in the solar system.
- TLDR: The Sun holds approximately 99.86% of the solar system's mass, not exactly 99.8%. The core claim is essentially correct but slightly understated.
- Explanation: According to NASA and Wikipedia, the Sun contains about 99.86% of the solar system's known mass. Musk's figure of 99.8% is a common rounded approximation that conveys the right order of magnitude but is slightly less precise than the scientifically accepted value.
- Sources:
  - [Solar System - Wikipedia](https://en.wikipedia.org/wiki/Solar_System)
  - [It's Surprisingly Hard to Go to the Sun - NASA](https://www.nasa.gov/solar-system/its-surprisingly-hard-to-go-to-the-sun/)

### ch3-6: INEXACT
- Speaker: Elon Musk
- Claim: Jupiter comprises about 0.1% of the mass in the solar system.
- TLDR: Jupiter is actually ~0.095% of the solar system's total mass, just under the 0.1% Musk states. The approximation is very close but slightly overstated.
- Explanation: Jupiter's mass is approximately 1.898 x 10^27 kg, which represents roughly 0.095% of the total solar system mass. Musk's figure of 'about 0.1%' is a reasonable rounding but is a minor overstatement. The Sun's share (~99.86%) is also slightly misquoted as 99.8%, though both figures are close approximations.
- Sources:
  - [Jupiter mass - Wikipedia](https://en.wikipedia.org/wiki/Jupiter_mass)
  - [Jupiter - Wikipedia](https://en.wikipedia.org/wiki/Jupiter)

### ch3-7: TRUE
- Speaker: Elon Musk
- Claim: Fusion energy on Earth is limited in scale because Earth's mass is negligible compared to the Sun's 99.8% share of solar system mass.
- TLDR: The Sun holding ~99.8% of solar system mass and Jupiter ~0.1% are well-established, widely cited figures.
- Explanation: NASA and multiple scientific sources confirm the Sun contains approximately 99.86% of the solar system's total mass, commonly rounded to 99.8%. Jupiter accounts for roughly 0.095% of the total mass, making Musk's figure of 'about 0.1%' a reasonable approximation. Both numbers are standard in astronomy.
- Sources:
  - [Sun: Facts - NASA Science](https://science.nasa.gov/sun/facts/)
  - [Solar System - Wikipedia](https://en.wikipedia.org/wiki/Solar_System)
  - [Jupiter - Wikipedia](https://en.wikipedia.org/wiki/Jupiter)

### ch3-8: INEXACT
- Speaker: Elon Musk
- Claim: Carl Sagan described Earth as like a tiny dust mote in a vast darkness.
- TLDR: Sagan did use 'dust mote' imagery for Earth, but his actual phrase was 'a mote of dust suspended in a sunbeam,' not 'in a vast darkness.'
- Explanation: Carl Sagan's famous Pale Blue Dot passage describes Earth as 'a mote of dust suspended in a sunbeam,' not in 'a vast darkness.' Musk's paraphrase captures the general spirit of Sagan's imagery and the attribution is correct, but the specific characterization he quotes does not match Sagan's actual wording.
- Sources:
  - [Environmental Messaging – Carl Sagan (Dust Mote) – Per Espen Stoknes (The Living Air)](https://skywaterearth.com/2018/02/20/environmental-messaging-carl-sagan-dust-mote-per-espen-stoknes-the-living-air/)

### ch3-9: FALSE
- Speaker: Elon Musk
- Claim: Total electricity production by all of human civilization is only about a trillionth of the Sun's energy output.
- TLDR: Humanity's electricity production is actually about 100 to 1,000 times smaller than a trillionth of the Sun's output, not a trillionth.
- Explanation: Global electricity production is roughly 3.4 TW (3.4 × 10^12 W) and total human energy consumption about 18-20 TW. The Sun's luminosity is ~3.83 × 10^26 W. The actual ratio is approximately 10^-14 to 10^-15 (one ten-quadrillionth to one quadrillionth), which is 100 to 1,000 times smaller than the 'trillionth' (10^-12) Musk states. The underlying point that humanity is tiny compared to the Sun remains correct, but the specific fraction is off by 2 to 3 orders of magnitude.
- Sources:
  - [Sun - Wikipedia](https://en.wikipedia.org/wiki/Sun)
  - [Power of the Sun - The Physics Factbook](https://hypertextbook.com/facts/1999/MatthewTsang.shtml)
  - [World energy supply and consumption - Wikipedia](https://en.wikipedia.org/wiki/World_energy_supply_and_consumption)

### ch3-10: INEXACT
- Speaker: Elon Musk
- Claim: Even increasing civilizational power output by a factor of one million would still leave humanity at only a millionth of the Sun's total energy output.
- TLDR: The math follows from Musk's 'trillionth' premise, but that premise is off by ~100x. A million-fold increase would leave us at roughly a hundred-millionth of the Sun's output, not a millionth.
- Explanation: Global electricity production is about 3.4 TW (3.4 x 10^12 W), while the Sun outputs 3.83 x 10^26 W, giving a ratio of roughly 10^-14 (a hundred-trillionth), not 10^-12 (a trillionth). Multiplying the correct base by one million yields ~10^-8, i.e. about a hundred-millionth of the Sun's output. The directional claim (we'd still be negligible) is correct, but the specific 'millionth' figure is off because the stated 'trillionth' premise is already ~100x too generous.
- Sources:
  - [Electricity – Global Energy Review 2025 – Analysis - IEA](https://www.iea.org/reports/global-energy-review-2025/electricity)
  - [2024 in review - Global Electricity Review 2025 | Ember](https://ember-energy.org/latest-insights/global-electricity-review-2025/2024-in-review/)
  - [Solar luminosity - Astronomopedia](https://astronomopedia.com/Wiki/Definitions/L%E2%98%89.php)

### ch6-2: FALSE
- Speaker: Elon Musk
- Claim: The Optimus robot is about 5'11" tall.
- TLDR: Optimus stands 5'8" (173 cm) tall, not 5'11" as Musk stated.
- Explanation: Tesla's official specifications, confirmed by Wikipedia and multiple technical sources, list the Optimus robot at 5 feet 8 inches (173 cm). Musk overstated the height by 3 inches when using Optimus as a scale reference for the Starship V3 rocket.
- Sources:
  - [Optimus (robot) - Wikipedia](https://en.wikipedia.org/wiki/Optimus_(robot))
  - [Tesla Bot a.k.a Optimus - Wevolver](https://www.wevolver.com/specs/tesla-bot-aka-optimus)

### ch6-3: TRUE
- Speaker: Elon Musk
- Claim: Starship V3 will expand payload capacity to orbit from 100 tons to 200 tons.
- TLDR: SpaceX's Starship V3 is indeed projected to reach 200 metric tons to LEO, up from roughly 100 tons for V2 (Block 2) in full reusability mode.
- Explanation: Multiple sources confirm that Starship V2 (Block 2) targets approximately 100 tons to LEO in a fully reusable configuration, while V3 with stretched tanks and upgraded Raptor engines is projected to double that to approximately 200 tons. This matches Musk's statement in the keynote precisely.
- Sources:
  - [SpaceX Version 3 Starship and Version 3 Starlink Both Arrive in 2025 | NextBigFuture.com](https://www.nextbigfuture.com/2025/02/spacex-version-3-starship-and-version-3-starlink-both-arrive-in-2025.html)
  - [What to Expect from Starship V3](https://payloadspace.com/what-to-expect-from-starship-v3/)

### ch6-4: TRUE
- Speaker: Elon Musk
- Claim: The mini AI satellite shown is roughly 100 kilowatts, with solar panels and a radiator.
- TLDR: Multiple sources confirm Musk described the "AI Sat Mini" as carrying 100 kilowatts of onboard power, with solar panels and a radiator shown to scale.
- Explanation: The Terafab keynote presentation is widely covered, and numerous outlets (SpaceNews, Tom's Hardware, Teslarati, New Atlas) confirm that the illustrated satellite called "AI Sat Mini" was described as a 100 kilowatt platform. Musk explicitly noted the solar panels and radiator were shown to scale, and remarked the radiator was small relative to the solar panels. No discrepancy was found between the claim and available evidence.
- Sources:
  - [SpaceX offers details on orbital data center satellites - SpaceNews](https://spacenews.com/spacex-offers-details-on-orbital-data-center-satellites/)
  - [Elon Musk unveils $20 billion 'TeraFab' chip project to make chips, memory, and package processors all under one roof — targets a terawatt of annual compute | Tom's Hardware](https://www.tomshardware.com/tech-industry/elon-musk-formally-launches-20-billion-terafab-chip-project)
  - [Elon Musk launches TERAFAB: The $25B Tesla-SpaceXAI chip factory that will rewire the AI industry](https://www.teslarati.com/elon-musk-lanuches-terafab-tesla-spacexai-chip-factory/)
  - [Elon Musk's Terafab is here: What it is, and why it's important for Tesla and SpaceX](https://finance.yahoo.com/news/elon-musks-terafab-is-here-what-it-is-and-why-its-important-for-tesla-and-spacex-163426221.html)

### ch6-5: TRUE
- Speaker: Elon Musk
- Claim: SpaceX has approximately 10,000 satellites in orbit.
- TLDR: SpaceX surpassed 10,000 simultaneous Starlink satellites in orbit in March 2026, just weeks before this keynote.
- Explanation: On March 16-17, 2026, SpaceX crossed the milestone of 10,000 active Starlink satellites in orbit simultaneously, reaching 10,020 according to tracker Jonathan McDowell. By the time of this April 6, 2026 keynote, the count stood at approximately 10,181. Musk's figure of 'approximately 10,000' is accurate.
- Sources:
  - [SpaceX reaches 10,000 simultaneous Starlink satellites in orbit following Falcon 9 launch from California – Spaceflight Now](https://spaceflightnow.com/2026/03/16/live-coverage-spacex-to-launch-25-starlink-satellites-on-falcon-9-rocket-from-california/)
  - [SpaceX reaches milestone of 10,000 Starlink satellites in orbit | Scientific American](https://www.scientificamerican.com/article/spacex-reaches-milestone-of-10-000-starlink-satellites-in-orbit/)
  - [starlink Constellation - 10181 satellites | SatelliteMap.space](https://satellitemap.space/constellation/starlink)

### ch6-6: TRUE
- Speaker: Elon Musk
- Claim: Reaching a terawatt of compute per year requires approximately 10 million tons of payload to orbit per year, at 100 kilowatts per ton.
- TLDR: Multiple sources confirm Musk stated exactly this figure during the Terafab keynote. The math is internally consistent.
- Explanation: Coverage of the Terafab keynote from multiple outlets quotes Musk saying: 'In order to get to the terawatt of compute per year, you need about 10 million tons to orbit per year at 100 kilowatts per ton.' The arithmetic holds: 1 TW divided by 100 kW per ton equals 10 million tons.
- Sources:
  - [Musk's Terafab: What a Terawatt of AI Compute Means for Energy Demand - Environment+Energy Leader](https://www.environmentenergyleader.com/stories/musks-terafab-what-a-terawatt-of-ai-compute-means-for-energy-demand,120232)
  - [Elon Musk unveils $20 billion 'TeraFab' chip project to make chips, memory, and package processors all under one roof — targets a terawatt of annual compute | Tom's Hardware](https://www.tomshardware.com/tech-industry/elon-musk-formally-launches-20-billion-terafab-chip-project)
  - [Terafab - Wikipedia](https://en.wikipedia.org/wiki/Terafab)

### ch6-7: UNSUBSTANTIATED
- Speaker: Elon Musk
- Claim: Achieving 10 million tons to orbit per year requires no new physics or impossible capabilities.
- TLDR: The 'no new physics' part is accurate, but the 'no impossible capabilities' assertion is unsupported. Reaching 10 million tons to orbit annually would require roughly 135+ Starship launches per day, a scale with no engineering roadmap.
- Explanation: No fundamental laws of physics prohibit 10 million tons to orbit per year. However, achieving it would require manufacturing tens of thousands of Starships and propellant supply chains orders of magnitude beyond anything currently planned or demonstrated. Engineering analyses flag manufacturing bottlenecks, atmospheric limits, and propellant logistics as enormous practical barriers. Musk presents the feasibility claim as a confident assertion with no supporting engineering evidence, making it unsubstantiated.
- Sources:
  - [How much mass can we put in orbit before running into atmospheric constraints? | Selenian Boondocks](https://selenianboondocks.com/2020/07/how-much-mass-can-we-put-in-orbit-before-running-into-atmospheric-constraints/)
  - [Elon Musk proposes 'Terafab' to level up chip production • The Register](https://www.theregister.com/2026/03/23/musk_terafab/)
  - [Elon Musk's Terafab Project: What to Know About His Chip Plan | Built In](https://builtin.com/articles/elon-musk-terafab-project)

### ch6-8: TRUE
- Speaker: Elon Musk
- Claim: After solving power generation with a terawatt of solar, the key missing ingredient for the space compute vision is a terawatt of compute.
- TLDR: Multiple sources confirm Musk presented this exact framing at the Terafab keynote: space-based solar solves power, and a terawatt of compute is the declared missing ingredient.
- Explanation: Search results from Tom's Hardware, The Register, and Musk's own X post confirm that the Terafab keynote centered on two pillars: solar irradiance in space (roughly 5x Earth's surface) solving the power generation constraint, and producing over a terawatt of compute per year as the stated goal. Musk explicitly argued that a full terawatt of AI compute cannot be run on Earth without overwhelming the grid, making orbital compute the key missing piece. The transcript excerpt is consistent with all available reporting on the keynote.
- Sources:
  - [Elon Musk unveils $20 billion 'TeraFab' chip project to make chips, memory, and package processors all under one roof — targets a terawatt of annual compute | Tom's Hardware](https://www.tomshardware.com/tech-industry/elon-musk-formally-launches-20-billion-terafab-chip-project)
  - [Elon Musk proposes 'Terafab' to level up chip production • The Register](https://www.theregister.com/2026/03/23/musk_terafab/)
  - [Terafab - Wikipedia](https://en.wikipedia.org/wiki/Terafab)

### ch9-1: TRUE
- Speaker: Elon Musk
- Claim: Terafab plans to produce two kinds of chips: one optimized for edge inference and one designed for space.
- TLDR: Confirmed. Terafab's keynote described two chip types: one for edge inference (cars and Optimus robots) and one radiation-hardened chip designed for space.
- Explanation: Multiple sources covering the March 2026 Terafab keynote confirm Musk outlined exactly two chip categories. The first is an edge inference chip for Tesla vehicles and Optimus humanoid robots. The second is a space-hardened processor (called D3) built for orbital data center satellites. Tom's Hardware and others note that Terafab will technically be two fabs, each dedicated to one of these designs.
- Sources:
  - [Elon Musk unveils $20 billion 'TeraFab' chip project to make chips, memory, and package processors all under one roof — targets a terawatt of annual compute | Tom's Hardware](https://www.tomshardware.com/tech-industry/elon-musk-formally-launches-20-billion-terafab-chip-project)
  - [Musk says Tesla, SpaceX, xAI chip project to kick off in Texas | Fortune](https://fortune.com/2026/03/22/musk-terafab-chip-project-tesla-spacex-xai-space-data-center-satellite/)
  - [Elon Musk announces TeraFab: $20bn factory will make chips for SpaceX orbital data centers and Tesla vehicles - DCD](https://www.datacenterdynamics.com/en/news/elon-musk-announces-terafab-20bn-factory-will-make-chips-for-spacex-orbital-data-centers-and-tesla-vehicles/)

### ch9-2: TRUE
- Speaker: Elon Musk
- Claim: The edge inference chip will be used primarily in Optimus robots and in cars, with Optimus being the larger use case.
- TLDR: Musk explicitly stated the edge inference chip targets Optimus and cars, with Optimus being the larger use case due to expected robot volumes 10-100x higher than car production.
- Explanation: Multiple sources covering the Terafab announcement confirm that the edge inference chip is designed for Tesla's Full Self-Driving cars and Optimus humanoid robots, with Optimus as the dominant application. Musk justified this by projecting humanoid robot production of 1-10 billion units per year versus roughly 100 million vehicles per year globally, making robots the far larger volume use case. The transcript excerpt directly matches widely reported details from the keynote.
- Sources:
  - [Elon Musk unveils $20 billion 'TeraFab' chip project to make chips, memory, and package processors all under one roof — targets a terawatt of annual compute | Tom's Hardware](https://www.tomshardware.com/tech-industry/elon-musk-formally-launches-20-billion-terafab-chip-project)
  - [What is Elon Musk's Terafab chip project? Here are his "most epic" goals for the factory. - CBS News](https://www.cbsnews.com/news/terafab-elon-musk-chips-semiconductors-what-to-know/)
  - [Terafab - Wikipedia](https://en.wikipedia.org/wiki/Terafab)

### ch9-3: TRUE
- Speaker: Elon Musk
- Claim: Humanoid robots are expected to be produced 10 to 100 times more than cars by volume.
- TLDR: Musk has consistently stated this 10-100x figure publicly, and the underlying math (1-10 billion robots vs. ~100 million cars/year) is consistent with his stated numbers.
- Explanation: Musk has made this prediction repeatedly in public forums. In response to Nvidia CEO Jensen Huang, he stated robots 'will be 10 times more ubiquitous than cars,' and in this keynote he extends the range to 10-100x. Global vehicle production is approximately 90 million units/year (all motor vehicles), so Musk's stated figure of ~100 million is a reasonable approximation, and 1-10 billion robots would indeed represent a 10-100x multiple.
- Sources:
  - [Elon Musk Responds to Nvidia's Jensen Huang: Humanoid Robots Will Be Ten Times More Common than Cars](https://www.topcpu.net/en/news/elon-musk-responds-to-nvidia%E2%80%99s-jensen-huang-humanoid-robots-will-be-ten-times-more-common-than-cars)
  - [Production Statistics - International Organization of Motor Vehicle Manufacturers](https://oica.net/production-statistics/)
  - [Elon Musk: 10 Billion Humanoid Robots by 2040](https://mikekalil.com/blog/10-billion-humanoid-robots-2040/)

### ch9-4: INEXACT
- Speaker: Elon Musk
- Claim: Global vehicle production on Earth is about 100 million vehicles per year.
- TLDR: Global vehicle production is closer to 89-92 million units in 2024, not quite 100 million. Musk's figure is a rounded overestimate.
- Explanation: According to OICA and S&P Global data, worldwide vehicle production in 2024 was approximately 89-92 million units, down slightly from 2023. Pre-COVID production was trending toward 100 million but never reached it. Musk's '100 million' is a common round-number approximation that overstates actual production by roughly 8-11%.
- Sources:
  - [List of countries by motor vehicle production - Wikipedia](https://en.wikipedia.org/wiki/List_of_countries_by_motor_vehicle_production)
  - [Stable global output in 2024, but notable changes in Europe | Automotive World](https://www.automotiveworld.com/articles/stable-global-output-in-2024-but-notable-changes-in-europe/)
  - [2025 Auto Sales Forecast: 89.6M Vehicle Sales Worldwide | S&P Global](https://www.spglobal.com/automotive-insights/en/blogs/2025-auto-sales-forecast-global)

### ch9-5: TRUE
- Speaker: Elon Musk
- Claim: Humanoid robot production is expected to reach between 1 billion and 10 billion units per year.
- TLDR: Musk has consistently stated this 1-10 billion unit/year projection for humanoid robots in multiple public forums.
- Explanation: Search results confirm that Musk has made the 1-10 billion humanoid robot per year projection publicly on multiple occasions, including the Terafab keynote and prior statements. He has also separately endorsed a 1 billion unit milestone by 2040. The range cited in the claim is consistent with his documented public statements.
- Sources:
  - [Elon Musk's Humanoid Robot Prediction Clarified: 1 Billion by 2040 — Roic News](https://www.roic.ai/news/elon-musks-humanoid-robot-prediction-clarified-1-billion-by-2040-10-29-2024)
  - [Elon Musk: 10 Billion Humanoid Robots by 2040](https://mikekalil.com/blog/10-billion-humanoid-robots-2040/)
  - [Elon Musk Backs Prediction of 1 Billion Humanoid Robots by 2040](https://www.iotworldtoday.com/robotics/elon-musk-backs-prediction-of-1-billion-humanoid-robots-by-2040)

### ch9-6: TRUE
- Speaker: Elon Musk
- Claim: Space chips must be designed to account for high-energy ions, photons, and electron buildup present in the space environment.
- TLDR: High-energy ions, photons, and electron/hole buildup are well-documented space radiation hazards that chip designers must account for. This is established aerospace engineering knowledge.
- Explanation: Multiple authoritative sources (NASA, Wikipedia, academic journals) confirm that space electronics are subjected to high-energy ions (cosmic rays, solar particles), high-energy photons (gamma/X-rays), and charge buildup from electron-hole trapping in oxides. These are the core radiation effects that drive the field of radiation hardening for space chips. Musk's description accurately reflects standard engineering considerations for space-grade electronics.
- Sources:
  - [Radiation hardening - Wikipedia](https://en.wikipedia.org/wiki/Radiation_hardening)
  - [Appendix Three: Space Radiation Effects on Integrated Circuits - NASA JPL](https://parts.jpl.nasa.gov/asic/Appendix.3.html)
  - [Hardening And Testing Semiconductor Devices In The Space Radiation Environment | Electronic Design](https://www.electronicdesign.com/technologies/analog/article/21798644/hardening-and-testing-semiconductor-devices-in-the-space-radiation-environment)
  - [Frontiers | Investigating the Effects of Cosmic Rays on Space Electronics](https://www.frontiersin.org/journals/physics/articles/10.3389/fphy.2020.00318/full)

### ch9-7: TRUE
- Speaker: Elon Musk
- Claim: Running chips hotter than on Earth minimizes radiator mass, which is a key constraint for space chip design.
- TLDR: Running chips hotter reduces radiator mass in space, a well-established engineering principle governed by the Stefan-Boltzmann law (heat rejection scales with T⁴).
- Explanation: In space, radiation is the only heat rejection mechanism (no convection). The Stefan-Boltzmann law shows that heat radiated per unit area scales with the fourth power of absolute temperature, so a hotter radiator can reject the same waste heat with far less surface area and mass. NASA and aerospace engineering literature explicitly identify operating electronics at higher temperatures as a key strategy for minimizing radiator mass, confirming Musk's statement.
- Sources:
  - [Spacecraft thermal control - Wikipedia](https://en.wikipedia.org/wiki/Spacecraft_thermal_control)
  - [7.0 Thermal Control - NASA](https://www.nasa.gov/smallsat-institute/sst-soa/thermal-control/)
  - [Review on thermal management technologies for electronics in spacecraft environment - ScienceDirect](https://www.sciencedirect.com/science/article/pii/S277268352400013X)
  - [Stefan–Boltzmann law - Wikipedia](https://en.wikipedia.org/wiki/Stefan%E2%80%93Boltzmann_law)

### ch9-9: TRUE
- Speaker: Elon Musk
- Claim: Terrestrial chip compute is estimated at 100 to 200 gigawatts per year, while space chip compute is estimated at around 1 terawatt.
- TLDR: Musk did state these exact figures during the Terafab keynote. They are his forward-looking estimates, not current measurements.
- Explanation: Multiple sources covering the March 2026 Terafab keynote directly quote Musk saying "it's probably a hundred to two hundred gigawatts a year of terrestrial chips, and probably on the order of a terawatt of chips in space." The numbers in the claim match verbatim. These are Musk's own projections for future compute demand, framed in the context of space having structural advantages like continuous solar exposure and vacuum heat rejection.
- Sources:
  - [Elon Musk unveils $20 billion 'TeraFab' chip project to make chips, memory, and package processors all under one roof — targets a terawatt of annual compute | Tom's Hardware](https://www.tomshardware.com/tech-industry/elon-musk-formally-launches-20-billion-terafab-chip-project)
  - [Elon Musk launches TERAFAB: The $25B Tesla-SpaceXAI chip factory that will rewire the AI industry](https://www.teslarati.com/elon-musk-lanuches-terafab-tesla-spacexai-chip-factory/)

### ch5-1: TRUE
- Speaker: Elon Musk
- Claim: Tesla has built the Giga Texas fab.
- TLDR: Tesla's Gigafactory Texas (Giga Texas) is a real, operational facility in Austin that opened in April 2022.
- Explanation: Giga Texas, located in Travis County near Austin, is a massive Tesla manufacturing campus covering 2,500 acres. Construction began in July 2020 and Tesla held its grand opening event ('Cyber Rodeo') on April 7, 2022. It produces Model Y vehicles and Cybertrucks.
- Sources:
  - [Gigafactory Texas - Wikipedia](https://en.wikipedia.org/wiki/Gigafactory_Texas)
  - [Giga Texas | Tesla](https://www.tesla.com/giga-texas)

### ch5-2: TRUE
- Speaker: Elon Musk
- Claim: Tesla is building the Optimus humanoid robot.
- TLDR: Tesla is actively developing and building the Optimus humanoid robot, with Gen 3 production ramping up in 2026.
- Explanation: Tesla's Optimus humanoid robot program is well-documented and ongoing. The robot was first announced in 2021, with successive generations developed since. As of 2026, Tesla is ramping production of Optimus Gen 3 and has showcased it publicly at events such as AWE 2026 in Shanghai.
- Sources:
  - [Optimus (robot) - Wikipedia](https://en.wikipedia.org/wiki/Optimus_(robot))
  - [Tesla showcases Optimus humanoid robot at AWE 2026 in Shanghai](https://www.teslarati.com/tesla-optimus-awe-2026-shanghai/)
  - [Musk races to build a robot army at Tesla. Silicon Valley is following.](https://www.washingtonpost.com/technology/2026/03/27/musk-optimus-robot-physical-ai/)

### ch5-3: TRUE
- Speaker: Elon Musk
- Claim: Tesla has a global supercharging network.
- TLDR: Tesla operates the world's largest global fast-charging network, with over 80,000 Supercharger stalls across more than 8,500 sites on multiple continents.
- Explanation: As of April 2026, Tesla's Supercharger network spans North America, Europe, Asia-Pacific, the Middle East, Africa (Morocco), and South America (Chile), totaling 80,018 stalls at 8,502 sites globally. The claim that Tesla has a global Supercharger network is well established and verified.
- Sources:
  - [Tesla Marks 80,000th Global Supercharger Milestone in France | TeslaNorth.com](https://teslanorth.com/2026/04/04/tesla-marks-80000th-global-supercharger-milestone-in-france/)
  - [Tesla Supercharger - Wikipedia](https://en.wikipedia.org/wiki/Tesla_Supercharger)
  - [Supercharger | Tesla](https://www.tesla.com/supercharger)

### ch5-4: TRUE
- Speaker: Elon Musk
- Claim: There were basically no electric cars for sale when Tesla started.
- TLDR: When Tesla was founded in 2003, there were virtually no electric cars commercially available for purchase by the general public.
- Explanation: The most notable prior EV, the GM EV1, was a lease-only program that ran from 1996 to 1999 and was never released for public purchase. GM recalled and destroyed those cars in 2003, the same year Tesla was founded. No other mainstream electric vehicle was on sale at the time, making Musk's characterization accurate.
- Sources:
  - [Tesla, Inc. - Wikipedia](https://en.wikipedia.org/wiki/Tesla,_Inc.)
  - [Tesla Roadster (first generation) - Wikipedia](https://en.wikipedia.org/wiki/Tesla_Roadster_(first_generation))

### ch5-5: FALSE
- Speaker: Elon Musk
- Claim: Tesla is making 2 million electric cars a year.
- TLDR: Tesla has never produced 2 million cars in a year. Its peak was ~1.85 million in 2023, and production has since declined.
- Explanation: Tesla's highest-ever annual production was 1,845,985 vehicles in 2023. In 2024, output fell to ~1.77 million, and in 2025 it dropped further to ~1.65 million. As of the video's publication date (April 2026), no year comes close to the 2 million figure Musk cites.
- Sources:
  - [Tesla Fourth Quarter 2024 Production, Deliveries & Deployments | Tesla Investor Relations](https://ir.tesla.com/press-release/tesla-fourth-quarter-2024-production-deliveries-and-deployments)
  - [Tesla reported 485,000 deliveries for the fourth quarter, bringing 2023 total to 1.8 million](https://www.cnbc.com/2024/01/02/tesla-tsla-q4-2023-vehicle-delivery-and-production-numbers.html)
  - [2025 (Full Year) Global: Tesla Worldwide Car Sales by Model and Outlook 2026 to 2029 - Car Sales Statistics](https://www.best-selling-cars.com/brands/2025-full-year-global-tesla-worldwide-car-sales-by-model-and-outlook-2026-to-2029/)

### ch5-6: TRUE
- Speaker: Elon Musk
- Claim: xAI is now part of SpaceX.
- TLDR: SpaceX acquired xAI in an all-stock deal announced on February 2, 2026, making xAI part of SpaceX before this keynote.
- Explanation: Multiple major outlets (CNN, CNBC, TechCrunch, Bloomberg) confirmed that SpaceX acquired xAI on February 2, 2026, in a deal valuing the combined entity at $1.25 trillion. By the time of this April 6, 2026 keynote, xAI was fully folded under the SpaceX corporate umbrella, making Musk's statement accurate.
- Sources:
  - [Elon Musk's SpaceX acquires xAI, merging his two most ambitious companies | CNN Business](https://edition.cnn.com/2026/02/02/tech/spacex-acquires-xai-elon-musk)
  - [Musk's xAI, SpaceX combo is the biggest merger of all time, valued at $1.25 trillion](https://www.cnbc.com/2026/02/03/musk-xai-spacex-biggest-merger-ever.html)
  - [Elon Musk's SpaceX officially acquires Elon Musk's xAI, with plan to build data centers in space | TechCrunch](https://techcrunch.com/2026/02/02/elon-musk-spacex-acquires-xai-data-centers-space-merger/)

### ch5-7: TRUE
- Speaker: Elon Musk
- Claim: xAI built the first gigawatt-scale compute cluster.
- TLDR: xAI's Colossus 2 in Memphis is widely recognized as the world's first gigawatt-scale AI compute cluster, going online in early 2026.
- Explanation: Multiple technical sources, including SemiAnalysis and NextBigFuture, confirm that xAI's Colossus 2 is the first AI training cluster to reach gigawatt-scale power consumption. No competing facility reached that threshold earlier. The Stargate UAE project and Amazon Indiana are either in later planning stages or aggregate multiple sites across a wider geography.
- Sources:
  - [xAI's Colossus 2 - First Gigawatt Datacenter In The World, Unique RL Methodology, Capital Raise](https://newsletter.semianalysis.com/p/xais-colossus-2-first-gigawatt-datacenter)
  - [XAI Colossus 2 - First Gigawatt AI Training Data Center | NextBigFuture.com](https://www.nextbigfuture.com/2025/09/xai-colossus-2-first-gigawatt-ai-training-data-center.html)
  - [xAI Activates Colossus 2: The world's first Gigawatt-scale AI training cluster - EONMSK News](https://www.eonmsk.com/2026/01/17/xai-activates-colossus-2-the-worlds-first-gigawatt-scale-ai-training-cluster/)

### ch5-8: TRUE
- Speaker: Elon Musk
- Claim: Jensen Huang from NVIDIA said he has never seen anything built as fast as xAI's compute cluster.
- TLDR: Jensen Huang publicly praised xAI's Colossus supercomputer build, calling it 'the fastest by far anyone's been able to do that' and describing the feat as 'superhuman'.
- Explanation: Multiple credible sources confirm that NVIDIA CEO Jensen Huang praised xAI's construction of a 100,000-GPU cluster (Colossus) in roughly 19 days, saying it was the fastest such build he had ever seen. He stated: 'As far as I know, there's only one person in the world who could do that,' and called Musk 'singular in his understanding of engineering.' Musk's claim accurately reflects Huang's public comments.
- Sources:
  - [Nvidia CEO Jensen Huang Praises Elon Musk For Achieving Something With xAI In 19 Days That Usually Takes At Least A Year: 'Singular In His Understanding Of Engineering'](https://www.benzinga.com/news/24/10/41307824/nvidia-ceo-jensen-huang-praises-elon-musk-for-achieving-something-with-xai-in-19-days-that-usually-t)
  - [Elon Musk set up 100,000 Nvidia H200 GPUs in 19 days - Jensen says process normally takes 4 years | Tom's Hardware](https://www.tomshardware.com/pc-components/gpus/elon-musk-took-19-days-to-set-up-100-000-nvidia-h200-gpus-process-normally-takes-4-years)
  - [Jensen Huang on Elon Musk and xAI building the fastest supercomputer on the planet in 19 days](https://www.startuparchive.org/p/jensen-huang-on-elon-musk-and-xai-building-the-fastest-supercomputer-on-the-planet-in-19-days)

### ch5-9: TRUE
- Speaker: Elon Musk
- Claim: SpaceX made reusable rockets economically feasible.
- TLDR: SpaceX's reusable Falcon 9 is well-documented as both technically and economically viable, cutting launch costs by roughly 75%.
- Explanation: Before SpaceX, many experts doubted reusable orbital rockets could be cost-effective. SpaceX's Falcon 9 reduced the cost to orbit from around $10,000/kg to roughly $2,500/kg, breaks even after just 2-3 reflights, and enabled SpaceX to capture over 60% of the global launch market. These results are broadly accepted as proof of economic feasibility.
- Sources:
  - [SpaceX's reusable Falcon 9: What are the real cost savings for customers? - SpaceNews](https://spacenews.com/spacexs-reusable-falcon-9-what-are-the-real-cost-savings-for-customers/)
  - [SpaceX reusable launch system development program - Wikipedia](https://en.wikipedia.org/wiki/SpaceX_reusable_launch_system_development_program)
  - [How Much Cheaper Are SpaceX Reusable Rockets? Now We Know | The Motley Fool](https://www.fool.com/investing/2020/10/05/how-much-cheaper-are-spacex-reusable-rockets-now-w/)

### ch5-10: TRUE
- Speaker: Elon Musk
- Claim: SpaceX has landed rockets over 500 times.
- TLDR: SpaceX has landed rockets well over 500 times. By early 2026, the total stood at approximately 575-593 successful booster landings.
- Explanation: Search data confirms SpaceX surpassed 500 booster landings long before April 2026. By February 2026, the count had reached at least 575, and aggregate figures put the total at approximately 593 successful landings out of 606 attempts. Musk's claim of 'over 500' is well supported.
- Sources:
  - [SpaceX completes 400th Falcon booster landing on a drone ship – Spaceflight Now](https://spaceflightnow.com/2025/08/27/spacex-completes-400th-falcon-booster-landing-on-a-drone-ship/)
  - [List of Falcon 9 first-stage boosters - Wikipedia](https://en.wikipedia.org/wiki/List_of_Falcon_9_first-stage_boosters)
  - [Statistics - SpaceXNow](https://spacexnow.com/stats)