Uploaded September 2025 | Updated September 2026, 22 hours ago
Tesla AI5 will be produced by TSMC starting in 2026. There will be Estimated volumes of 10-20% of relevant TSMC wafers, yielding 50-100 chips per wafer. AI6, will use Samsung's facilities in Texas, potentially allowing 50-80% (or near 100%) dedication, enabling 5x or more chips than TSMC.
Redlich's analysis involves estimating tokens (AI processing units) for tasks, but Wong focuses on "unlimited high-value" applications like fully automated AI coding, comparing it to robotaxis: Currently, it's like driver-assist (human coding with AI help), but full automation would create a "generalized system" with 100% accuracy through testing, eliminating probabilistic errors (referencing Jason Calacanis and others). This could generate unlimited value by making super low errors tested and validated software generated by AI.
Broader Implications: Chips in Everything That Moves. Every future car will need 1-4 inference chips (Elon specified four per vehicle), extending to bots, tractors, trucks, and anything that moves (echoing Nvidia's Jensen Huang). Not all need Tesla's $5k chips, but inference will be ubiquitous.
]
Energy equals intelligence equals wealth.
Physics fundamentals: Compute requires energy (reversible vs. non-reversible computing, with the latter generating heat), bound by unbreakable laws.Futuristic Vision: Unlimited Intelligence and Consumption. Elon's vision of capturing solar energy from Earth, Mars, and the Sun for computing, creating unlimited intelligence. AI won't just serve humans; advanced bots will consume resources like electricity and chips, mirroring human needs.
Pre-Industrial Revolution wealth was measured in horses/cattle/servants. Future equivalents are cars (horses) and bots (servants). As nanotechnology advances, physical objects become commoditized ("like potatoes"), but power and intelligence remain key.
Software becomes free, like today's open-source AI.
Chip Costs, Competition, and Vehicle Integration
TSLA hit $396. Fed rate cut (25bps Wednesday) could drive gains, as past cuts spurred $100 and $250 moves. With no recession, market rises expected. Catalysts: FOMO Monday, strong China sales Tuesday (potentially 4k+ Y Ls, blowing 430k Q3 consensus), leading to 15-20% upside. Resistances at 440/488/500; host predicts 440 next week, all-time highs by Oct 2. Wong sees potential for 600-700 by Oct/Nov on deliveries, FSD v14, Cybercab, energy. Without earnings growth or macro headwinds, pullbacks to 450-500 possible. Recent Tesla purchases signal demand rebound.Recap and Bullish Valuation (46:31–53:22)Nod to bulls like Redlich, Wong, Cern Basher. Tesla's potential: #1 in robots (biggest product ever), robotaxis (biggest revenue/profit), inference chips (#1 maker, possibly world's top product), motors/actuators (billions/year), stationary storage, electricity sales, transformers, trucking. Valuation (per Basher exercise): $3k/share now (just robotaxis/Optimus); with extras, much higher. If 1M Cybercabs next year + unsupervised FSD, could exceed 3k. Winners like Tesla keep winning (e.g., like GE, Microsoft, Apple expansions). The conversation ends on an explosive, mind-blowing note.The discussion is forward-looking, blending speculation with data, and highlights Tesla's interconnected ecosystem as unbeatable.
Follow Brian Wang on NextBigFuture.com
Follow Brian Wang on X.com/@NextBigFuture
This channel brings you the most important science and technology breakthroughs and analysis at least once per week.
Tesla AI5 will be produced by TSMC starting in 2026. There will be Estimated volumes of 10-20% of relevant TSMC wafers, yielding 50-100 chips per wafer. AI6, will use Samsung's facilities in Texas, potentially allowing 50-80% (or near 100%) dedication, enabling 5x or more chips than TSMC.
Redlich's analysis involves estimating tokens (AI processing units) for tasks, but Wong focuses on "unlimited high-value" applications like fully automated AI coding, comparing it to robotaxis: Currently, it's like driver-assist (human coding with AI help), but full automation would create a "generalized system" with 100% accuracy through testing, eliminating probabilistic errors (referencing Jason Calacanis and others). This could generate unlimited value by making super low errors tested and validated software generated by AI.
Broader Implications: Chips in Everything That Moves. Every future car will need 1-4 inference chips (Elon specified four per vehicle), extending to bots, tractors, trucks, and anything that moves (echoing Nvidia's Jensen Huang). Not all need Tesla's $5k chips, but inference will be ubiquitous.
]
Energy equals intelligence equals wealth.
Physics fundamentals: Compute requires energy (reversible vs. non-reversible computing, with the latter generating heat), bound by unbreakable laws.Futuristic Vision: Unlimited Intelligence and Consumption. Elon's vision of capturing solar energy from Earth, Mars, and the Sun for computing, creating unlimited intelligence. AI won't just serve humans; advanced bots will consume resources like electricity and chips, mirroring human needs.
Pre-Industrial Revolution wealth was measured in horses/cattle/servants. Future equivalents are cars (horses) and bots (servants). As nanotechnology advances, physical objects become commoditized ("like potatoes"), but power and intelligence remain key.
Software becomes free, like today's open-source AI.
Chip Costs, Competition, and Vehicle Integration
TSLA hit $396. Fed rate cut (25bps Wednesday) could drive gains, as past cuts spurred $100 and $250 moves. With no recession, market rises expected. Catalysts: FOMO Monday, strong China sales Tuesday (potentially 4k+ Y Ls, blowing 430k Q3 consensus), leading to 15-20% upside. Resistances at 440/488/500; host predicts 440 next week, all-time highs by Oct 2. Wong sees potential for 600-700 by Oct/Nov on deliveries, FSD v14, Cybercab, energy. Without earnings growth or macro headwinds, pullbacks to 450-500 possible. Recent Tesla purchases signal demand rebound.Recap and Bullish Valuation (46:31–53:22)Nod to bulls like Redlich, Wong, Cern Basher. Tesla's potential: #1 in robots (biggest product ever), robotaxis (biggest revenue/profit), inference chips (#1 maker, possibly world's top product), motors/actuators (billions/year), stationary storage, electricity sales, transformers, trucking. Valuation (per Basher exercise): $3k/share now (just robotaxis/Optimus); with extras, much higher. If 1M Cybercabs next year + unsupervised FSD, could exceed 3k. Winners like Tesla keep winning (e.g., like GE, Microsoft, Apple expansions). The conversation ends on an explosive, mind-blowing note.The discussion is forward-looking, blending speculation with data, and highlights Tesla's interconnected ecosystem as unbeatable.
Follow Brian Wang on NextBigFuture.com
Follow Brian Wang on X.com/@NextBigFuture
This channel brings you the most important science and technology breakthroughs and analysis at least once per week.
