After nailing the long breakout and shorting the top, $SPCX has now pulled back to the exact level of the breakout ~$160.
This is where I wanna be long again and the vols are all elevated.
I like the $175/$200 call spreads again but this time, expiring no 11/13. They are about $4.00 per combo, which makes this a 5:1 payout, real juice in the trade.
Now, if we pull back further, I will sell puts to finance this trade. Right now I'm just paying premium.
Recall the market is a little creaky after the new ATH with Nasdaq, but we are in a trader's market and $SPCX is a retail favorite.
I'd rather be long after this textbook pullback than flat.
Sau khi bắt trúng cú bứt phá tăng và bán khống ở đỉnh, $SPCX giờ đã điều chỉnh trở lại đúng mức breakout khoảng ~$160.
Đây là nơi tôi muốn mua lại và biến động hàm ý đều đang ở mức cao.
Tôi lại thích spread quyền chọn mua $175/$200, nhưng lần này đáo hạn vào 11/13. Chúng khoảng $4.00 cho mỗi combo, tức là tỷ lệ chi trả 5:1, lợi thế rất rõ trong giao dịch này.
Giờ nếu giá điều chỉnh sâu hơn nữa, tôi sẽ bán quyền chọn mua để tài trợ cho giao dịch này. Hiện tại tôi chỉ đang trả phí quyền chọn.
Nhớ rằng thị trường đang hơi chông chênh sau đỉnh cao mới mọi thời đại của Nasdaq, nhưng đây là thị trường dành cho trader và $SPCX là mã được nhà đầu tư nhỏ lẻ ưa chuộng.
Tôi thà giữ vị thế long sau cú pullback đúng sách giáo khoa này còn hơn đứng ngoài.
Trong tuần cuối của tháng 9, các mô hình DeepSeek đã xử lý nhiều token hơn trên OpenRouter so với các mô hình của OpenAI, Google, Anthropic và xAI cộng lại.
Morgan Stanley: AI CDS basket trades ~45bp wider than CDX IG
MS says buying CDS protection "remains our preferred way of playing the AI story in the credit derivatives market."
The reason is not just a bearish default call.
MS expects "increasing needs for non-economic hedging to manage counterparty exposure," even in a benign scenario where AI investments are profitable and continue.
That is an important distinction.
The hedge bid can grow even if the capex cycle keeps working.
More AI financing creates more counterparty exposure, and that creates more demand for protection.
SPXXAI March 85% put costs 132bps vs 101bps for SPX equivalent
The 6th percentile on a 2-year lookback is the basis for that cheapness.
Downside protection on the S&P excluding AI stocks, which has fallen roughly 7% from its late-August peak while the cap-weighted index sits near all-time highs, you pay a roughly 30% premium over the equivalent SPX put.
The premium exists because SPXXAI is the index that is actually moving lower.
The cap-weighted SPX, dominated by the AI names, is suppressing implied vol at the index level even as the average constituent is well off its highs.
That’s roughly a 30% premium for protection on the part of the market that's actually falling.
Citadel Securities: token price fell more than half but Azure AI token consumption is up ~4.5x
Effective access per dollar is actually worse, not better.
This is as a demonstration of why cheaper intelligence need not mean a smaller compute bill.
Falling token prices are being offset by more intensive usage and broader deployment.
What is good for adoption can simultaneously be challenging for model producers' pricing power.
Non-technical Wall Street analysts continue to be short the right tail on compute and inference demand. It’s at least an order of magnitude higher than they expect, yet the hyperscalers know this and it’s why they are investing over $1T in 2027 for datacenter buildouts.
@morganstanly has revised AI capex estimates "sharply higher in every year," with 2027 ballooning from $687bn to $1.384tn and 2028 from $763bn to $1.547tn versus the Nov-2025 numbers.
GOOGL at $400bn and AMZN at $350bn drive the 2028 total, and the revision is the story: the base case got twice as large in under a year.