The host believes the stock market is building a bubble right now and a big crash is coming; when it happens, the bond market will rally sharply.
The S&P 500 (SPY) remains in an overall uptrend, so the current drop is just a healthy short-term correction. The host expects the index to fall further to around 7,000–7,200 (about a 10% correction) before forming a higher low and launching a slingshot rally, making it a buying opportunity.
The host says gold is continuing to decline because market sentiment around gold remains excessively bullish; there is likely more downside ahead, and a bottom will not form until that sentiment flips.
The host maintains a bearish outlook on gold, arguing that under a Kevin Warsh Fed there will be no money printing, so gold peaked on February 2 and enters a lost period until the money printer is turned on again or Warsh is replaced.
The host said that if Discord had IPO'd he would have been dumping on it, and its stock would probably have gone straight down; he is glad it has not IPO'd.
The host noted that CoreWeave is finally starting to find an intraday bottom, which could support a recovery after last week's sell-off.
Host says he would take the opposite side of the rally in CoreWeave CDS, believing the market's implied probability of default is overstated. With potential Nvidia funding and remaining financing options, default fears should recede and CDS prices should fall.
Memory is still in massive shortage now, but the host thinks DRAM prices will start declining next year, which is actually good for the broader industry even if China-related factors do not help stabilize prices.
The host believes yields have already peaked and really need to come down; if yields instead rise back above his peak call, the AI trade worsens as other IPOs are delayed.
Declining model sizes and stabilizing memory prices make the host more optimistic about the broader buildout using AMD and Nvidia chips, benefiting AMD.
The host pushes back on the bullish extrapolation that AI agents will create massive CPU demand. He argues that because CPU time is shared, a single chip can support many agents, so actual CPU needs and AI capex expectations are far lower than the market assumes.
Host argues CoreWeave is not the black swan; although it faces near-term liquidity issues and a shift away from training, the market is overpricing its default risk. Nvidia, a 10% shareholder, could lend the roughly $7 billion needed, the stock is still above its IPO price, and equity dilution is possible, so he views the decline as a buy-the-dip opportunity rather than the start of a systemic collapse.
The host argues CoreWeave's financial engine 'sucks'; the company is $8 billion short and cannot pay its bills, so Elon Musk's reaction is no surprise and the situation is not good.
NBIS also dropped sharply, failing to hold the annual line and even breaking yesterday's low, showing extreme weakness.
CoreWeave's three founders have cashed out $33.1 billion since its IPO and now hold less than 20% combined; the company has essentially been sold off. The stock fell back to the prior 6,065 support zone and is being hammered repeatedly; the options linked to Jensen Huang's earlier 'buy' call will likely go to zero on the November expiry unless there is a sudden surge in the remaining weeks of October. The company's endgame is either bankruptcy or being acquired.
The host believes strong iPhone Duo demand and likely solid gross margins are great news for AAPL, and that the new CEO's focus on hardware rather than chasing AI models is the right, durable strategy.
The host believes neoclouds are screwed and his bearish thesis is playing out; these companies are likely to fail over the next 6 to 12 months.
System data looks quite bearish; the 706.58 area is unlikely to hold, and a break would test 700 and potentially lower. The overall bias is bearish from this week, but violent emotional rebounds mean no heavy short; keep 20-30% long base and add a tolerable short hedge.
The peaking of AI high growth is the recent main theme; if the peak is expected in late 2027–2028, the market will react about a year earlier. Current growth is high but could be revised down, with all AI semiconductors and semiconductor manufacturing potentially adjusting downward; the AI high-growth trade has begun to unwind starting with the storage crash.
Netflix is not closely tied to AI; although it has ample cash flow and low capex, user growth and price hikes are hard, leaving no high-growth story, and the market is gradually no longer valuing it as a high-growth tech stock.