$SMCI An unaudited preliminary report sent the contract from 32.5 to 38. Everyone’s shouting that demand for AI servers has exploded, so let me pour a little cold water on that: this is a preliminary report, not an earnings report.
First, why did it move? After the U.S. market closed yesterday, Supermicro released preliminary first-quarter figures: revenue is expected to be $18.5 billion to $19.5 billion, versus its previous guidance of $14.5 billion to $15.5 billion; gross margin is expected to be 12% to 12.5%, versus previous guidance of 10.4% to 10.8%. Coming right after last week’s sell-off in AI data center stocks, those figures gave the market some relief, and the stock rose more than 10% after hours. Two things to note: the figures are unaudited, and the official earnings report and conference call aren’t until November 3.
Now let’s look at how the Gate perpetual contract moved. At 04:00 today, a roughly 1.44 million U trade sent it straight from 32.56 to 36.89. Normally, hourly volume is only around 10,000 to a few tens of thousands of U. It then steadily ground its way up to 38.11 at 10:00. It’s now around 36.76, up about 17% over 24 hours.
What I’m really watching is positioning:
1. The 04:00 move was a 13% spike, but open interest only increased from around 1.13 million U to 1.19 million U, then fell back to 1.12 million U the next hour. It’s now around 1.06 million U. The new money didn’t stick around. This was a repricing, not an influx of capital.
2. The account long/short ratio was between 2.4 and 2.9 before the spike, with 70% of accounts holding longs. After the spike, it kept falling and is now around 0.73, having dipped as low as 0.58. In plain English: longs who’d been holding on took the chance to get out as prices recovered, and some even flipped to shorts. The large-trader position long/short ratio has stayed around 1.0, so big money hasn’t shown its hand.
3. At 16:00, a roughly 190,000 U sell-off knocked the price from 37.61 to 36.31. It was the largest red candle since the rally, and open interest fell at the same time, from around 1.09 million U to 1.06 million U. Longs are exiting.
My view: don’t chase it above 38. The stretch from 32.5 to 36 was created by a single candle, leaving a vacuum with no turnover, so a retracement is quite possible. But I wouldn’t short here either: nearly 60% of accounts are already short, and the funding rate is zero. A crowded short trade isn’t a good spot to enter.
When U.S. markets open tonight, watch 36.3. If it breaks below that level on rising volume while open interest continues to climb, that would suggest new shorts are driving the sell-off and the preliminary-results rally is fading. First watch 35.2 to 34.9, the closing range from October 14–15.
Here’s what would prove me wrong: an hourly close back above 38.11, with open interest rising above 1.15 million U. That would mean buyers are willing to chase this preliminary report. I’d admit I was wrong and look for 39.2 to 39.7 above.
One more reminder: open interest in this contract on Gate is only around 1 million U. It’s a small market, and one trade can cause a sharp wick, so don’t size your position too heavily.
And while I’m at it, I’ll own up to a mistake. Yesterday I said $CBRS could be a short opportunity if its rebound to 122–125 failed. My invalidation criteria were a high-volume open and a close back above 125, with open interest falling. At 22:00, a roughly 1.74 million U candle surged to 126.2 and closed at 125.8, while open interest fell from around 4.49 million U to 4.36 million U. Every condition was met, so by my own rules, I was wrong. But the price couldn’t hold there; at 03:00 it was back to 118.4, and it’s now around 122, with more than 60% of accounts still long. The lesson: when your invalidation level is hit, that’s a stop-loss signal, not a reason to reverse your position. If I’d flipped and chased the long that day, I could have taken a hit just the same.
This is just market commentary, not a recommendation.
First, why did it move? After the U.S. market closed yesterday, Supermicro released preliminary first-quarter figures: revenue is expected to be $18.5 billion to $19.5 billion, versus its previous guidance of $14.5 billion to $15.5 billion; gross margin is expected to be 12% to 12.5%, versus previous guidance of 10.4% to 10.8%. Coming right after last week’s sell-off in AI data center stocks, those figures gave the market some relief, and the stock rose more than 10% after hours. Two things to note: the figures are unaudited, and the official earnings report and conference call aren’t until November 3.
Now let’s look at how the Gate perpetual contract moved. At 04:00 today, a roughly 1.44 million U trade sent it straight from 32.56 to 36.89. Normally, hourly volume is only around 10,000 to a few tens of thousands of U. It then steadily ground its way up to 38.11 at 10:00. It’s now around 36.76, up about 17% over 24 hours.
What I’m really watching is positioning:
1. The 04:00 move was a 13% spike, but open interest only increased from around 1.13 million U to 1.19 million U, then fell back to 1.12 million U the next hour. It’s now around 1.06 million U. The new money didn’t stick around. This was a repricing, not an influx of capital.
2. The account long/short ratio was between 2.4 and 2.9 before the spike, with 70% of accounts holding longs. After the spike, it kept falling and is now around 0.73, having dipped as low as 0.58. In plain English: longs who’d been holding on took the chance to get out as prices recovered, and some even flipped to shorts. The large-trader position long/short ratio has stayed around 1.0, so big money hasn’t shown its hand.
3. At 16:00, a roughly 190,000 U sell-off knocked the price from 37.61 to 36.31. It was the largest red candle since the rally, and open interest fell at the same time, from around 1.09 million U to 1.06 million U. Longs are exiting.
My view: don’t chase it above 38. The stretch from 32.5 to 36 was created by a single candle, leaving a vacuum with no turnover, so a retracement is quite possible. But I wouldn’t short here either: nearly 60% of accounts are already short, and the funding rate is zero. A crowded short trade isn’t a good spot to enter.
When U.S. markets open tonight, watch 36.3. If it breaks below that level on rising volume while open interest continues to climb, that would suggest new shorts are driving the sell-off and the preliminary-results rally is fading. First watch 35.2 to 34.9, the closing range from October 14–15.
Here’s what would prove me wrong: an hourly close back above 38.11, with open interest rising above 1.15 million U. That would mean buyers are willing to chase this preliminary report. I’d admit I was wrong and look for 39.2 to 39.7 above.
One more reminder: open interest in this contract on Gate is only around 1 million U. It’s a small market, and one trade can cause a sharp wick, so don’t size your position too heavily.
And while I’m at it, I’ll own up to a mistake. Yesterday I said $CBRS could be a short opportunity if its rebound to 122–125 failed. My invalidation criteria were a high-volume open and a close back above 125, with open interest falling. At 22:00, a roughly 1.74 million U candle surged to 126.2 and closed at 125.8, while open interest fell from around 4.49 million U to 4.36 million U. Every condition was met, so by my own rules, I was wrong. But the price couldn’t hold there; at 03:00 it was back to 118.4, and it’s now around 122, with more than 60% of accounts still long. The lesson: when your invalidation level is hit, that’s a stop-loss signal, not a reason to reverse your position. If I’d flipped and chased the long that day, I could have taken a hit just the same.
This is just market commentary, not a recommendation.