SpaceX stock sentiment: the score flipped, the float doubled, the price followed
By Maya Koeva · August 17, 2026 · 5 min read
Ten days ago we published a post about SpaceX that ended on a deadline: both sides now have a date. Quantral's score on SPCX had spent a month at 19, then flipped to 77 in the three days after the company's first earnings report, while the raw crowd stayed nearly two to one bearish. That morning, roughly 912 million pre-IPO shares came off lockup, enough to double the public float. If buyers absorbed the unlock, the flip was early and right. If they did not, the score would follow the credible accounts back down.
Buyers absorbed the unlock. SPCX rose 15.8% on lockup day itself. On August 10 it closed at $138.74, its first close above the $135 first-day price since July 15. On Friday it finished at $140.00, up 21.8% from the $114.92 the stock was worth when that post went out, and 29.3% off its August 5 low close. The flip held, and the tape came to it.
What the unlock did
The bear case was not silly. Every share that had wanted out since June and could not get out was free to leave, into a stock that had fallen 13.6% on its first earnings report three sessions earlier. Instead, the selling met a bid. The stock jumped to $133.11 on unlock day, cleared its IPO price the following Monday, touched $146.15 on Wednesday, 35% above the August 5 close, and settled at $140.00 by Friday.
The stock has round-tripped: Friday's $140.00 is within a dollar of the $139.14 close the day our first SpaceX autopsy went out. A price chart alone says nothing happened in those five weeks. The score went 19, then 77, and read both halves of the round trip correctly. That is the difference between watching the tape and watching the conversation.
The room that called it
The score follows the accounts with graded track records, and their calls on SPCX split cleanly by calendar month. From accounts scored above 0.5 credibility:
| Month | Graded calls, bullish : bearish | What the price did |
|---|---|---|
| June | 35 : 70 | $135 first-day close, $211.39 top, then the slide begins |
| July | 8 : 70 | $139.14 on July 13 falls to $108.37 by July 31 |
| August | 40 : 12 | $108.27 low close recovers to $140.00 by August 14 |
Bearish through the slide that cut the stock nearly in half, bullish through a 29% recovery. Right both times, and the flip was on the record before the price moved: the heaviest cluster of graded bullish calls landed in the twenty-four hours around the August 4 earnings report, at closes between $125.33 and $108.27. The calls placed before the print were down 13.6% a session later. The room did not flip back: over the three sessions after the print, its calls ran 10 bullish to 4 bearish, into a stock that had cracked.
Some of the names are familiar from our Nebius receipts table. On SpaceX, they spent July on the other side:
| Account | Credibility | July calls on SPCX | August calls on SPCX |
|---|---|---|---|
 @PhotonBull | 0.64 | 7 bearish, 0 bullish | 8 bullish, 1 bearish |
 @CKCapitalxx | 0.80 | 2 bearish, 0 bullish | 2 bullish, 0 bearish |
 @epictrades1 | 0.53 | 2 bearish, 2 bullish | 3 bullish, 1 bearish |
 @ThematicTrader | 0.58 | no calls | 2 bullish, 0 bearish |
@PhotonBull has 299 graded calls behind that 0.64, @CKCapitalxx has 725, @ThematicTrader 956. They spent July telling us the slide had further to go, were right, and then changed their minds at the print. The July autopsy said it plainly: if the credible accounts change their minds, the number will follow them up, and that is the design.
The crowd never came around
The raw mention count is what makes this a story about the score rather than about SpaceX: it stayed bearish the whole way. The room flipped hard after the first leg down in June, nearly four to one bearish in the week of June 22, and never turned back. July ran more than two to one bearish. August, through a 29% recovery: still 122 bearish subject mentions against 87 bullish.
Since the unlock, the split is starker when you sort it by platform. Of the bullish directional mentions, 20 of 28 came from X, spread across 13 accounts. Of the bearish ones, 23 of 24 came from Reddit. And the credibility gradient that pointed bearish in July now points the other way: in the week before our July post, the bears averaged 0.48 credibility to the bulls' 0.43; since the unlock, the bulls average 0.50 to the bears' 0.44. Same lens, opposite reading, because the accounts underneath it moved.
A raw mention counter looking at SPCX in August would have reported a stock the crowd still hated, and it would have missed the only thing that mattered: who had switched sides. That is the whole argument for weighing mentions instead of counting them.
What this proves, and what it does not
The caveats are worth stating precisely because the test went our way. Six sessions is six sessions. SPCX still sits 33.8% below its June 16 top of $211.39, and a stock that absorbed one unlock can still choke on the next headline. The score did not predict the recovery, and we will keep saying so: it read the conversation, the conversation flipped at the print, and this time the price agreed within a week. The July version of that same mechanism got its answer over the two and a half weeks of decline that followed.
The arc shows the thing we built the score to do. A month of 19 while a famous stock fell, a flip to 77 the week the argued case changed, and a crowd count that would have kept you two to one bearish through a 29% recovery. Twice now the number has stood against the loudest read on this stock, and twice the price ended up on its side.
As of this morning the 7-day score on SPCX reads 76. The dated snapshot of this conversation lives on our SpaceX stock sentiment page; the live version lives in the app.
See the live reading on SpaceX in Quantral: the score, every call behind it, and the track record of the account that made it.
Mention counts, sentiment splits, and credibility figures cover the accounts Quantral tracks, bucketed by rolling 24-hour windows, with monthly splits by calendar month over May to August 2026. Scores are the current seven-day reading as of August 17, 2026, or the values published in the posts linked above, not a stored history. Prices are closing prices via public markets through the August 14 close. Positions attributed to accounts are their stated views, not Quantral's. Quantral surfaces signals and context from public sources to support your own research. Nothing here is financial advice or a recommendation to buy or sell. Past signals are not indicative of future results.