Risk-on crypto stayed constructive despite a weak macro tape: BTC held around 82K support, September remained green, ETF flows stayed strong, and altcoins outperformed hard, with SOL singled out for volume, tokenized equities, app revenue, and AI-agent activity. The video tied that crypto bid to a broader thesis that AI agents need fast, cheap, final settlement rails, which favors SOL over BTC and ETH, while CEXs, banks, insurers, brokers, and distribution businesses face fee compression and capital flight. Outside crypto, the notes turned to AI and compute: open models were framed as dominant, orbital data centers as the next compute frontier, and SpaceX/Tesla/Elon as the clearest physical-world AI/infra lever. Macro and stocks were weaker: consumer confidence, fear-and-greed, and AI equities softened, while diesel, oil, and a possible October Fed hike were used to argue inflation pressure remains sticky.
Wintermute had about $126M in shorts on HYPE, led by $50M in ETH shorts, $11M in SOL shorts, and $10M in HYPE shorts.
US consumer confidence fell 6.7% to 81.9, the lowest level since 2013; the stock fear-and-greed index sat at 32.
BTC was down about 4% over seven days to roughly 82–83K, while ETH was also down 4%, SOL was flat, BNB down 6%, XRP down 1%, ZEC was pausing, and HYPE fell about 8%.
BTC stayed up 6% for September despite the weekly drop, and over the last 30 days it was up 6.6% to 12%; ETH was up 8.7%, BNB up 8%, SOL up 12%, ZEC up 64%, XRP up 8%–8.7%, and HYPE up 3.2%.
Crypto fear and greed moved from extreme greed last week to 73, while BTC ETF flows were described as tapering slightly after the highest ETF inflows since October 2025.
BTC was framed as likely to beat the S&P 500 for the next 3 years, with rotating an S&P 500 401k into IBIT presented as a way to outperform the index.
Mike Novogratz/Galaxy was cited as bullish on BTC reaching 100K before year-end, with the timeframe described as about 12 weeks; 82K was presented as the key support zone, with bears calling for 79K or lower if it breaks.
Cardone’s goal shifted from owning 100,000 apartments to owning 10,000 BTC.
Spot SOL ETFs stayed very green after their biggest day ever two days earlier, and the notes flagged a possible rotation out of ETH ETFs into SOL.
SOL was said to be crushing centralized exchanges in volume: 1.147B daily volume, up 51.45%, versus Binance at 386M and Coinbase at 100M, with Gate.io and Kraken called small; DEX usage was preferred because fees are cheaper and transactions are faster.
Tokenized equities were led by SOL at 32% of volume, followed by Robinhood at 29%, Binance at 25%, Base at 12%, and ETH at 1%; app revenue by chain was put at SOL $5.42B, more than 3x ETH, more than 3x HYPE, and nearly 8x Robinhood chain.
Market cap per daily active user was cited as $550K for ADA, $520K for ETH, nearly $8M fully diluted for XRP, $47K for BNB, $255K for AVAX, $11K for SOL, $34K for SUI, and $9,593 for TRX, with TRX called the winner.
AI agents were framed as needing finality and therefore avoiding BTC and ETH because both are too slow and too expensive; Nansen was cited showing X42 transactions up 3x in a week, with SOL handling about 70%–75% of them.
The AI-agent thesis extended to finance and labor: agents were said to be able to pull excess profits from distribution, insurance, brokers, and agents, and a $100K Bank of America deposit at 0.01% was used to show an agent moving cash to 6% or 12% instantly.
Banks were said to be worried about the Clarity Act, but deposits were called cooked either way.
For October 1, the notes listed SOL price targets of $126 from Deepseek, $125.50 from Gemini Flash, and $5.73.50 for "Luna for GBT"; SOL was said to be around $119 and as high as $121.60 intraday.
Over the last seven days, AI stocks were weak: AAPL down 3%, GOOGL down nearly 5.3%, TSLA down 6%, AVGO down nearly 2%, while AMD and MU were only slightly higher and AMZN was red or flat.
AMD reportedly bought a company described as 'Lufay or something' focused on robots, movement, and self-driving tech; SpaceX was said to be training Grock on physical-world data from SpaceX and Tesla, giving it a huge advantage.
SpaceX’s latest Starship launch was described as equivalent to 20 Falcon 9 launches, with about $200M in launch cost for kilograms that would otherwise cost about $5M–$10M to loft; Starlink v3 was said to be covered by 26 satellites worth 444 AI2 satellites, with the next launch in 2 weeks.
Orbital AI/orbital data centers were said to 'smoke' terrestrial data centers, with the tipping point around 2030 and stronger growth in 2028–2029; regulation and permits were cited as the reason SpaceX cannot scale terrestrial data centers, making space compute a 'nuclear arms race.'
TSLA was said to have opened an M&A office, and Tesla 2026 sales per million people were listed with Norway leading at 3,000 new Teslas per million, while the USA, South Korea, Iceland, Australia, and Denmark were all above 1,000 per million.
Open-source models were said to use above 80% of all tokens, with the trend expected to continue; closed frontier models were reserved for special projects, and open models were said to run on a laptop or Mac Studio.
A rate hike for October was framed as 70% probable because diesel and oil were very high, with diesel at an all-time high and driving prices across the economy; a Belgian student finding 49 gold bars worth $10M and turning them in closed the segment.
The final story questioned whether he kept one gold bar for himself, calling 49 a suspicious number.