The loudest stocks this week, scored: Aug 17–21

By Maya Koeva · August 24, 2026 · 5 min read

A glossy chrome desk calendar with a glowing signal waveform running across its face, marking a week of mentions against scores.

The accounts Quantral tracks posted 1,758 mentions from 681 distinct authors over the week of August 17 to 21. The leaderboard below is the easy part: count the mentions, sort them, done. The column that takes the work is the score: who is doing the talking on each name, which way they lean, and whether they have ever been right before. This week the two columns disagreed almost top to bottom, and the gaps are the story.

The board

Score is the 7-day Quantral signal score as of Aug 24; trusted is the share of each name's mentions that came from authors with a graded track record. Price moves run Friday close to Friday close, August 14 to 21.

CompanyScoreMentionsTrustedWeek move
ModernaMRNA51165Bull 55% · Bear 21%24%+129.2%
NebiusNBIS81150Bull 70% · Bear 11%50%-21.1%
MicronMU72107Bull 49% · Bear 29%21%-0.5%
MetaMETA7672Bull 32% · Bear 50%19%-6.8%
SanDiskSNDK8369Bull 74% · Bear 12%28%-2.7%
RedditRDDT5966Bull 48% · Bear 45%15%-13.9%
MarvellMRVL7959Bull 61% · Bear 20%31%+6.8%
NikeNKE6351Bull 24% · Bear 57%16%+0.1%

The loudest name carries the lowest score

Moderna out-talked Nebius and Micron, names our set discusses all year, and its 51 is the weakest score on the board. The sequence explains both numbers. On Wednesday the accounts we track lit up with news of a positive Phase 3 readout for the mRNA cancer vaccine Moderna developed with Merck, and the stock closed that day up 177%. Before the news, Moderna does not appear in our week at all: zero mentions Monday, zero Tuesday. That is a coverage gap, not market silence. The accounts we track are concentrated in tech and trading rooms, and a biotech catalyst was not on their radar until it was on everyone's.

Mentions
196
PositiveNegativeChatterNoise
Moderna mentions across the accounts we track, by the app's rolling 24-hour windows. The spike bar covers Wednesday morning through Thursday pre-market: the crowd arrived with the 177% move, not before it.

We take no credit for the week's biggest winner. The score's contribution was narrower: it read the room that showed up. Of the 154 mentions on the spike bar, a quarter carried no direction at all, and the bullish share settled at 55%, low for a stock that had just tripled. A crowd that came to watch, not to call the next leg, earns a 51, and Thursday's 23.5% giveback is roughly what a 51 deserves.

The most credible room had its worst week

Nebius is Moderna's mirror. Second-loudest name, and the strongest crowd on the board: half its 150 mentions came from trusted accounts, the room leaning 70% bullish, the score at 81. The tape disagreed all week. Nebius fell every session, 21.1% in total, with nothing arriving to snap it back the way Microsoft's print did in the July drawdown.

Price
$219.13-21% over the window+0% off the low
Aug 14Aug 21
Nebius daily close, down 21% across five straight red sessions while the most credible room on the board stayed 70% bullish.

The honest ledger on this one: the stock ran from $188 at the last scorecard to $278 by mid-August, up 48%, so the room that held through July is still up 17% on the round trip. But this week the credible crowd was long and wrong, and the score followed its conviction rather than the tape. If the slide continues, this section gets harder to write, and we will write it anyway.

The middle of the board is one thesis

Micron and SanDisk are the memory trade our graded accounts have backed since July, the same names Jim Cramer endorsed on Monday. Both survived the memory selloff that greeted his segment and ended the week roughly flat, down 0.5% and 2.7%, with scores of 72 and 83. The third memory name from that post, Western Digital, stayed below this table's volume cut at 15 mentions, but it is the row to watch: our graded accounts doubt it, Cramer backs it, and it fell 14.3% from Monday's close. The first week of the live disagreement went to the graded room.

Marvell sits in the same AI-adjacent lane: 61% bullish, the second-most trusted crowd on the board at 31%, and a 6.8% gain into its earnings report this coming week. The print will grade that room in the next scorecard.

Three loud names leaning bearish

The bottom half of the board leans bearish. Meta drew 72 mentions leaning 50% bearish and fell 6.8%; the crowd's lean was right this week, and the score, which weights the thinner trusted subset, stayed warmer at 76. Reddit's room split almost evenly, 48% bull to 45% bear, while the stock gave back its S&P 500 inclusion pop, down 13.9%. The score reads a split that close as a 59, and the tape sided with the bears. Nike was the strangest room of the week: 57% bearish, the most bearish room on the board, and the stock closed the week three cents higher than it started. Sometimes the crowd argues and the tape abstains.

The best call missed the table

The sharpest read of the week came from a room too small to crack this leaderboard. Robinhood drew 28 mentions, 24 of them bullish, between Wednesday morning and Friday's open, citing the crypto run and the Clarity Act, while the stock sat flat. On Friday it jumped 13.7% to $108.13. The room formed first and the move came after, which is the order this product exists to surface. The caveat: only 14% of those mentions came from trusted accounts, so this was a right-and-early crowd more than a right-and-credible one. Its score of 82 now sits near the top of our rankings on a fraction of the board's volume.

The takeaway

Read the two columns against each other and the week sorts itself. Volume with no history behind it (Moderna) scored low and deserved to. Credibility without confirmation (Nebius) kept its high score through a losing week, and that bet is still open. And the week's cleanest call (Robinhood) never got loud enough to make the board at all. The leaderboard tells you where the attention went. The score exists because attention and signal are rarely the same thing.


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.