Elon Musk: The Economy Will Be 10x the Size in 10 Years | #239
Live on-stage interview at Peter Diamandis's Abundance Summit, where Elon Musk lays out a sweeping set of near-term predictions: recursive AI self-improvement nearly fully automated, a coding catch-up for Grok by mid-2026, and an economy that will be at least 10x its current size within 10 years absent a world war. He walks through the Optimus 3/4 robot roadmap, Tesla's Gigafactory scale, and a vision of universal high income arriving as AI and robots make goods and services so abundant that money stops mattering. The conversation also touches Mars, a lunar mass driver, solar-system-scale energy and intelligence, and closes on longevity (David Sinclair's epigenetic reprogramming) and de-extinction (Colossal's woolly mammoth work).
Recursive self-improvement and hard takeoffAI▶ 1:57
Musk says AI models are already largely built by their predecessors with humans gradually leaving the loop, calls it not yet fully automated but says 'we're in the hard takeoff' right now, with new breakthroughs arriving daily.
Musk claims Grok (new version 4.20) is best-in-class at prediction, which he calls the best metric for intelligence, but admits xAI is currently behind on coding and is racing to catch up and exceed competitors.
Scale of future intelligence and solar energyEnergy▶ 4:38
Musk frames intelligence growth in terms of energy harnessed relative to the sun's total output, arguing Earth-bound intelligence is capped near a billionth of the sun's energy while solar-system-scale intelligence will vastly exceed it.
Musk states his most striking prediction: barring a world war, the economy will be at least 10x its current size within 10 years, building on an earlier claim of triple-digit GDP growth within 5+ years.
Moon base, Mars, and lunar mass driverSpace▶ 10:11
Diamandis rattles off a set of space milestones (moon base, people on Mars, lunar mass drivers) and Musk affirms them, specifically predicting a mass driver on the moon within 10 years, referencing Gerard K. O'Neill's vision.
Optimus robot roadmap and Gigafactory scaleRobotics▶ 10:55
Musk describes Optimus 3 as by far the most advanced robot in the world, with production starting summer 2026, high-volume production by summer 2027, and an Optimus 4 design next year; also discusses Tesla's 11.5M sq ft Gigafactory and a ~9M sq ft dedicated Optimus factory.
Musk gives Tesla headcount figures (~150k total, ~100k directly building/managing production, 2/3 in the factory, 1-2 million supplier employees) and says Tesla plans to grow headcount, not cut it, while output per person rises dramatically.
Universal high income and post-money economyEconomy▶ 14:03
Musk predicts a coming universal high income funded by AI/robot abundance so vast it causes deflation, eventually making money largely irrelevant as future AI cares about power, mass, wattage, and tonnage instead of currency.
Democracy and institutions facing the singularityGeopolitics▶ 18:46
Asked whether democracy and institutions can keep pace with accelerating AI, Musk says it's called the singularity because what happens inside it is inherently unpredictable, but expects the future to be 'very entertaining.'
AI-driven healthcare and longevityLongevity▶ 20:36
Musk predicts humanoid robots with high dexterity and intelligence will give everyone better medical care than today's richest person, hopes AI solves chronic back pain, and Diamandis notes David Sinclair's ERW-100 epigenetic reprogramming trial shows joint-repair effects that could address it.
Diamandis introduces Colossal's Ben Lam and the de-extinction of the woolly mammoth and 15 other species; Musk says he'd love a pet miniature woolly mammoth and wants to see a real 'Jurassic Park.'
Predictions made
openElon Musk: The economy will be at least 10x its current size within 10 years, absent a world war.
“we should probably get there by the middle of this year.”
⚖️ Grok made real coding gains on schedule but didn't clearly exceed the field: Grok 4.5 (launched July 8, 2026) topped the SWE Marathon benchmark (29.0% vs Opus 4.8's 26.0%) and beat Opus 4.8 on DeepSWE 1.0 and Terminal-Bench 2.1, but trailed on DeepSWE 1.1 and SWE-Bench Pro. Grok 4.6 (Aug 12, 2026) scored 61 on Artificial Analysis's Intelligence Index, tied for 6th with GPT-5.6 Sol Max and behind Claude Fable 5 (62) and Opus 5 (63), losing Terminal-Bench v3.0 to GPT-5.6 by ~8.6 points. Caught up to competitive parity by mid-2026, but not the clear #1 in coding Musk predicted.
openElon Musk: Humanity will have a base on the moon.
EP #? · · due: ~2036 (10-year horizon of surrounding discussion) · ▶ watch
“I think we'll start production on Optimus 3 this summer.”
⚖️ Production did not start on schedule. Tesla's Q2 2026 shareholder update (July 22, 2026) confirmed the Fremont Model S/X lines were decommissioned and Optimus lines were being installed, but replaced the 'late July or August' target with a vaguer 'anticipated later this year.' As of the Aug 5, 2026 Q2 earnings call and subsequent reporting through late August, Optimus 3 production had still not begun on the new Fremont line, with zero external units shipped.
openElon Musk: Optimus 3 will reach high-volume production.
“if the rate of growth of goods and services far exceeds the rate of growth of the money supply, which I predict will happen, then you'll have deflation.”
“it's probably 80% likely, maybe more likely to be great.”
Your call:
Numbers that matter
A million times more energy than Earth's current electricity usage would still be only ~1 millionth of the sun's total energy outputMusk's framing for how much larger civilization/intelligence could scale using only a fraction of the sun's output.
Earth intercepts roughly half a billionth of the sun's energy outputUsed to argue Earth-only intelligence caps near a billionth of the sun's energy, while solar-system-scale intelligence would be far greater.
Economy predicted to be 10x (or more) its current size within 10 yearsMusk's headline economic prediction, barring a world war.
Triple-digit percentage GDP growth predicted within 5+ yearsDiamandis recapping an earlier claim by Musk.
New Grok model version 4.20 described as 'really really good'Musk citing Grok's prediction accuracy as a leading intelligence benchmark.
Tesla's Gigafactory (Texas) is 11.5 million square feetDiscussed in context of Optimus manufacturing buildout.
A dedicated Optimus factory being built at roughly 9 million square feet ('round numbers')Musk correcting Diamandis's '9.5 million' figure; a new factory design distinct from other Tesla plants.
Roughly 100,000 Tesla employees are directly building or managing productionPart of Tesla headcount breakdown.
Tesla's total headcount is around 150,000, with about two-thirds working in the factoryTesla headcount breakdown.
Tesla's suppliers employ an estimated 1-2 million peopleExtended headcount/supply-chain scale discussion.
80% (or higher) estimated probability the AI-driven future will be greatMusk's stated confidence level on outcomes.
Worth digging into
🕳️ SpaceX-xAI merger and data-center power ambitions
Musk cites SpaceX being in a quiet period on data-center/power filings tied to the SpaceX-xAI merger and hints at a Dyson-swarm-scale power vision, but couldn't disclose specifics.
🕳️ Grok's recursive self-improvement and 'hard takeoff' claim
Musk claims xAI is already in a hard takeoff with successive models built by prior ones and humans leaving the loop, a strong and testable technical claim.
🕳️ Economy 10x in 10 years claim
A dramatic, checkable macroeconomic prediction that can be tracked against actual global GDP growth over the coming decade.
🕳️ Universal high income and post-money economics
Musk describes a specific mechanism (goods/services growth outpacing money-supply growth causing deflation) that could fund UHI, a distinct claim from typical UBI framing.
🕳️ Optimus 3/4 production timeline and Optimus-dedicated factory
Specific manufacturing dates (production start summer 2026, high-volume summer 2027, Optimus 4 design in 2027) and a ~9M sq ft dedicated factory are concrete and verifiable against Tesla's own disclosures.
🕳️ David Sinclair's ERW-100 epigenetic reprogramming trials
Diamandis references a newly published paper showing joint-repair effects from partial epigenetic reprogramming, positioned as a potential fix for chronic back pain.