The market wasn’t paying for promises this morning. It asked for proof of funds, and the AI complex couldn’t produce any. The premarket bid is gone, the gap faded at the open, and TSLA and META now report after the close with the sector already on the back foot.
The catalyst for the premarket bid was SMCI, which raised its revenue and margin outlook before the bell, got paid premarket, and gave it back at the open. That tells you what this market is willing to reward right now: guidance it can underwrite, and even then only until the opening bell. Not the opinion piece you find on Medium but the audited, non-refundable one that comes with a receipt, and apparently even that’s not enough today.
The open answered half the question. A genuine re-rating of AI earnings power doesn’t evaporate in the first few minutes of trading. A positioning squeeze does.
And the damage underneath was already there. Over the past two weeks NVDA is down about 6%, SMCI is down about 22%, and META is sitting roughly 11% below its three-month high. One green premarket was never going to repair two weeks of technical damage. That’s like deciding your liver is fine because you drank one cup of green tea after a week-long bender.
Zoom out on SMCI and the picture gets even uglier. It is still about 29% below its three-month high, in a fresh bearish transition, with a weekly expected move near 9%, roughly $29 to $35. The options market was pricing a volatile week, not a trend reversal, and so far it’s been right.
NVDA is the sibling who never showed up to the party. It closed yesterday near $216, mid-range, in its own fresh bearish transition, with a narrow weekly expected move of about 4%, roughly $205 to $223. If SMCI’s guidance had been lifting the whole complex, NVDA would have shown it premarket. It didn’t. The biggest name in AI wasn’t buying the story, and that was the tell.
TSLA is the exception. But then again, Musk makes flying cars now. It’s the one name still in a working uptrend, closing yesterday near $374 and only about 4% below its three-month high, near the top of its weekly expected move of roughly $328 to $381.
The market had already paid for a decent week going into the print. Holding into tonight means asking for the encore before the band has played the set.
Deliveries are already known: 486,532. The headline number cannot surprise, so margins, demand commentary, and the growth outlook are the real variables tonight. As for the SPCX merger chatter making the rounds, it’s unconfirmed speculation. Treat it as infotainment, not a catalyst.
META comes in from the other side: about 11% off its high, with the rally finally cashing out and leaving the bag to the stragglers. Tonight’s watch items are ad revenue growth against AI capex.
This tape has been punishing megacap AI spend that has not shown returns, so META needs the ad engine to visibly fund the buildout. If the ads don’t pay for the GPUs, the story ends with a margin call and the obligatory hazing on social media and ZeroEdge.
And there you have it: the theme running through all four charts. Investors are getting tired of the narrative and want to know how guidance and spending discipline will get them paid. Headline beats alone are not convincing anyone anymore. When the bar is already set above the number, a beat is just meeting expectations with a better story.
The first half of the kill signal already printed: the gap faded. What finishes it is megacap reports that meet expectations and still get sold. What rescues it is TSLA and META pairing beats with credible guidance. Until then, the gap fade is the only evidence on the table, and it points one way.
If you want this kind of read – trend state, expected move, and where price sits inside it – turned into live signals instead of homework, SPY Alpha covers equities and index options. You can check it out here.
Cheers,
Michael
