# Both megacaps fell on earnings. The crowd we track abandoned Tesla and doubled down on Alphabet. > Two days ago both names carried a warm but thin-credibility bullish lean into their prints. Both then fell, Tesla 14.5% and Alphabet 7.1%, so the lean did not pay. What split was the response: Tesla's room flipped bearish, while Alphabet drew nearly triple the trusted share and stayed bullish, leaning into the drop. And the Fed decides Wednesday. By Maya Koeva · 2026-07-24  Two days ago, on the eve of Wednesday's prints, we [looked at Tesla and Alphabet](/blog/tesla-alphabet-pre-earnings-chatter) and found the accounts we track doing something a little counterintuitive: they warmed into the catalyst. Both names got louder over the week, both turned net bullish, both scores climbed. But we were skeptical, and said so. The credibility behind that lean was thin, only about one in ten mentions came from an account with a track record, so it read as expectation building into an event, not the credible conviction the [score](/learn/what-is-a-credibility-score) is built to reward. The tell, we wrote, would come afterward: whether the trusted accounts that sat it out started showing up. Here is what happened. On the tape, the warm lean was wrong on both. And then the accounts we track split about as cleanly as they can. ## The setup they walked in with Here is where the two sat going in, the numbers from Tuesday's piece, over the seven days into the print: | Company | Mentions | Bullish | Bearish | Trusted | | --- | ---: | ---: | ---: | ---: | | Alphabet (GOOGL) | 72 | 50% | 31% | 10% | | Tesla (TSLA) | 32 | 44% | 38% | 13% | Two modest, positive-leaning rooms heading into a binary event, near-identical in the one thing we said mattered most: barely one in ten voices carried a track record. On that measure you could not tell them apart. ## Both prints disappointed Then they reported, and both fell. Tesla dropped **14.5%** the next session, from 374.01 into the print to 319.69 at Thursday's close. Alphabet dropped **7.1%**, from 341.91 to 317.69. Two down prints, Tesla roughly twice as hard. On the tape, the warm bullish lean the crowd we track walked in with did not pay on either name, which is exactly what the thin credibility behind it was warning about. That is the part that matters going in: the pre-print lean was low-information on both. What happened next is where the two names stop looking alike. ## Tesla: the room abandoned it Tesla's crowd turned with the tape. The 44% bullish, 38% bearish lean it walked in with flipped to **32% bullish, 47% bearish** across the 75 directional mentions in the two days since the close, already more than double the 32 the name drew across the entire prior week. Its 24-hour score, the one that reacts fastest, has dropped to 54. The trusted share ticked up only slightly, to 16%, so this was less a case of credible voices arriving than of the whole room souring at once as the stock fell. *[Chart: Tesla mentions on Quantral, July 15 to 24: a light, two-sided base through the week, then a red-heavy spike the day the print landed. The green pre-print lean is the small stuff on the left.]* ## Alphabet: the credible crowd doubled down Alphabet fell too, but its room did the opposite of Tesla's. It drew **162 directional mentions** in the two days since the close, more than its entire prior week's 72, and it held its tilt: **52% bullish to 29% bearish**, essentially where it went in, even as the stock dropped 7.1%. What changed is who was talking. The trusted share jumped from 10% to **26%**, nearly triple the credibility it carried going in. Its 24-hour score reads 72 and its 7-day score has climbed from 70 to 78. *[Chart: Alphabet mentions on Quantral, July 15 to 24: a quiet base, then a wall of green-led volume on the print, far bigger than anything the week before. The last bar reflects the read as of Friday morning.]* This is the follow-up we said to watch for, and it is more interesting than a simple win or loss. The trusted accounts that sat out the pre-print lean showed up in force, and they showed up leaning into the drop, not away from it. Where Tesla's credible voices sold with the tape, Alphabet's are taking the other side of it. ## What actually told them apart Going in, these two looked like the same setup: a modest bullish room on thin credibility, and both fell on the print. The pre-print lean, the part a mention leaderboard would have shown you, was wrong on both and told you nothing. What separated them came after the result was known, in what the credible crowd did with a drop. Tesla's abandoned the name. Alphabet's leaned harder into it. That is the whole reason the score weights credibility and grades over months rather than days. It is not a next-day price call, both of these names would have burned you on the lean this week. It is a read on who is committing and how much of a track record they carry, built to be settled over a quarter, not a session. Right now it is flagging a real disagreement on Alphabet: the credible crowd we track on one side, Thursday's tape on the other. That is not a prediction of a bounce. It is a question, and the next few months are what answer it. The useful signal was never the warmth going in. It is whether the accounts that keep being right stick with a name when the tape turns against it, the way they are on Alphabet and pointedly are not on Tesla. ## One layer down: the read-through to AI infrastructure There is a second read in the data, underneath the two headline names, and it helps explain why Alphabet's credible room refused to sell the drop. The accounts we track came out of Wednesday's prints leaning harder into the names that build the AI layer the megacaps spend on. It is the read-through this crowd tends to make: a megacap pouring money into AI is rough on its own margins, one reason a stock can fall on a heavy-spending quarter, but it is a tailwind for the picks-and-shovels names selling into that spend. And that is where the credible rooms turned more positive, not less, in the two days since. | Company | Positive, week into the print | Positive, since the print | 7-day score | | --- | ---: | ---: | ---: | | Nvidia (NVDA) | 69% | 77% | 73, up from 68 | | Nebius (NBIS) | 73% | 83% | 84, holding near its high | Nvidia's room got both louder and more one-sided after the prints, and its 7-day score rose from 68 to 73. Nebius, already the loudest and highest-conviction name on our board when we [mapped the earnings week](/blog/earnings-week-crowd-looking-elsewhere), turned even more one-sided, 83% positive from 73%, with its trusted share climbing to 49%. Be honest about the limits: this did not lift the whole group evenly. The smaller optics names the crowd liked going in, Lumentum and Aehr among them, drew only a handful of mentions each in the two days and are too thin to read into, and not all of them held their scores. But on the two AI-infrastructure names this crowd cares most about, the same prints that knocked the megacaps down read as a positive. That is a piece of why Alphabet's own credible room leaned into the drop instead of running from it: the accounts with a track record were reading the spending, not just the stock. ## And the Fed is next The calendar is not done. The Fed decides Wednesday, July 29, the next catalyst stacked on the back of a rough earnings week for the megacaps. We do not grade the Fed, we grade the crowd, and after the decision the question will be the same one that separated these two names: not how the room leaned going in, but which accounts, with which track records, stay with their read once the [after-hours](/learn/what-is-after-hours-trading) tape moves against it. The first reaction is always the loudest. The signal worth reading usually shows up a beat later. --- *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.*