Neocloud stocks are heating up - but only three of six
13 August 2026 · 6 min read
The short version: the neocloud group's median market sentiment score went from 6.5 on 29 July to 72 on 12 August, and on that last session three neoclouds took three of the seven largest one-day moves on the 34-name board. That is a real warming and the numbers below are our own. But the segment now carries the widest internal spread of any of the six - 50 points between its best and worst name - and two of the six are still falling. "Neoclouds are heating up" is true of half a sector.
All figures are as at the close on 12 August 2026, the last complete session. The segment page carries the live version.
The warming, in one column
Median market sentiment score across the six listed neoclouds, by session, against the median for all 34 names:
Two things are worth taking from that column, and they pull in opposite directions.
The first is how deep the hole was. On 29 July, five of the six neoclouds scored in single digits - Nebius 5, IREN 6, Applied Digital 6, Cipher 7, CoreWeave 9 - while the board as a whole sat at 14. That is a segment-wide capitulation, and it is the base every percentage in this article is measured from.
The second is that the line is not a climb. It is 6.5, then 74.5, then back to 26.5 for two sessions, then 72. This group has been violent in both directions for three weeks, and a reading taken on any single day of it would have supported an entirely different headline.
The session itself
On 12 August the whole board rose - median move +3.85% - so neoclouds were not moving alone. They were, however, moving hardest. Of the seven biggest one-day moves on the board, three were neoclouds:
- Nebius (NBIS) - +34.1%, the largest move on the board, sentiment 44 → 86
- CoreWeave (CRWV) - +19.3%, second largest, sentiment 59 → 82
- IREN - +9.9%, seventh, sentiment 53 → 76
- Applied Digital (APLD) - +4.9%, sentiment 47 → 68
- Cipher Mining (CIFR) - +3.8%, sentiment 35 → 36
- TeraWulf (WULF) - +2.6%, sentiment 33 → 40
Nebius's 42-point gain was the largest single-session sentiment move anywhere on the board that day. Cipher's single point was the smallest of the six, and six names elsewhere on the board moved less still.
The part most write-ups will miss: it is not a sector move
The neocloud segment now spans 86 at the top and 36 at the bottom - a 50-point spread, the widest of the six segments we track. Equipment and EDA is next at 42; the tightest two, optical and networking and power and systems, span 24 apiece.
Concretely: TeraWulf and Cipher Mining rank 31st and 32nd of 34 names. Only Western Digital and Cadence score lower. Both are still down over the five sessions to 12 August - TeraWulf -4.9%, Cipher -4.5% - in a week when the cap-weighted complex rose 2.6%.
The distinction that explains this is one the sector's own structure suggests: Nebius and CoreWeave sit at the cloud layer, selling accelerated compute; TeraWulf and Cipher sit closer to the power-and-campus layer, converting sites first developed for bitcoin mining. IREN straddles both, and scores between them at 76. We are not claiming the score knows that - it reads price and volume and nothing else. We are pointing out that where it draws its line and where the business models divide happen to be the same line.
What actually moved the score
The score is six weighted components, all derived from price and volume. For Nebius on 12 August, three of them sat within a hundredth of their ceiling:
- Momentum 12.00 of a possible 12.00
- Trend 9.60 of 9.60
- Relative strength 9.59 of 9.60
- Volume 7.06 of 7.20
- Drawdown -0.50
Nothing in the formula clips - each component is compressed through tanh, so it approaches its bound without ever reaching it - but a name sitting this close to three ceilings is telling you something the score cannot: above this point it stops discriminating. If Nebius rose another 20% tomorrow, those three numbers would barely move. CoreWeave's are effectively identical (12.00, 9.60, 9.60). The score can say both are extremely strong. It cannot say which is stronger.
Everything here is a rebound, not a breakout
This is the figure that should temper the whole piece. Not one of the six is near its 52-week high. The group averages 65% of it:
- Nebius 86% of its 52-week high
- CoreWeave 70%
- Applied Digital 61%
- Cipher Mining 59%
- TeraWulf 58%
- IREN 57%
The drawdown component is negative for all six, IREN's at -4.65 against a bound of 4.80 - close to as negative as that input goes. A stock can be the strongest name on a board by momentum and trend while remaining 43% below where it traded a year ago. Both are true of IREN right now, and only one of them makes a headline.
The attention arrived first, and we cannot tell you that means anything
Our X cashtag counts run one complete New York day behind the price data, so the most recent day is 11 August - the session before the move. On that day the neocloud group took 32.4% of the board's 23,330 posts, up from 19.3% the day before. CoreWeave went from 601 posts to 3,448 in a single day. Nebius went from 1,545 to 2,440.
The attention split the same way the prices did, incidentally: while CoreWeave was multiplying by five and a half, Cipher fell from 427 posts to 274 and TeraWulf from 156 to 145.
It would be easy, and wrong, to turn that into a leading indicator. We tested exactly this question across the whole board and found post volume has a rank correlation of -0.00 with the direction of the following day's move - the working is here. One well-timed instance is not evidence against thirty-four names measured over weeks; it is what a coin flip looks like when it comes up heads. A mention count measures how much a name is being discussed and nothing else. It carries no direction.
The limits of this
Stated plainly, because they matter more than the headline does:
- This is one session. Momentum is 25% of the score and trend another 20%, which is why a single day can move a reading 42 points. A score that swings that far in a session is describing a session.
- The score has no view on why. It reads price and volume. It does not know what was announced, filed or reported on 12 August, and we are not going to pair a real number with an invented reason for it. Most financial headlines about a move are written after it and about it.
- Three weeks is not a trend. Sentiment history on this site began on 7 August 2026 and earlier sessions are reconstructed from daily bars, so the series above is short by design.
- None of this is a forecast or a recommendation. The segment that led for one session is not thereby the segment that leads for the next.
What would make this a trend rather than a day
Three things are worth watching on the neocloud segment page over the coming fortnight, and each is checkable against our own data:
- Does the spread close or widen? If TeraWulf and Cipher climb toward the leaders, the sector is re-rating. If the gap grows, the market is repricing the cloud layer and leaving the power-and-campus layer where it is. Those are different events wearing the same headline.
- Does the drawdown component turn positive? That is the input that separates a rebound from a breakout, and today it is negative for every one of the six.
- Does the trend component hold when momentum fades? Momentum decays within days. Trend does not. A group that keeps its trend score after the initial move is behaving differently from one that gives it all back - as this group did between 3 and 6 August.
The daily board carries all 34 names with the same score, rebuilt every session.
Frequently asked questions
Are neocloud stocks heating up?
Partly. The group's median market sentiment score rose from 6.5 on 29 July to 72 on 12 August 2026, and Nebius, CoreWeave and IREN took three of the seven largest one-day moves on our 34-name board on that session. But TeraWulf and Cipher Mining scored 40 and 36, ranking 31st and 32nd of 34, and both were still down over the prior five sessions. The segment carries a 50-point internal spread, the widest of the six we track.
Which neocloud stock has the highest sentiment score?
As at the close on 12 August 2026, Nebius (NBIS) at 86, followed by CoreWeave (CRWV) at 82 and IREN at 76. Applied Digital was 68, TeraWulf 40 and Cipher Mining 36. Scores are recomputed every session and will have moved since - the neocloud segment page carries the live figures.
Why did Nebius stock jump on 12 August 2026?
Our sentiment score reads price and volume only, so it can tell you that Nebius rose 34.1% on the day, that this was the largest move on the board, and that its momentum, trend and relative-strength components all reached the top of their ranges. It has no view on what caused the move, and we do not pair a real number with a reason we cannot evidence.
Are neocloud stocks near their highs?
No. As at 12 August 2026 the six averaged 65% of their 52-week highs - Nebius 86%, CoreWeave 70%, Applied Digital 61%, Cipher Mining 59%, TeraWulf 58% and IREN 57%. The drawdown component of the sentiment score was negative for all six, which is what distinguishes a rebound from a breakout.
Do rising X mentions predict a neocloud rally?
Our own measurement says no. Neocloud cashtag posts jumped from 19.3% to 32.4% of the board's total on 11 August, the day before the move, which looks predictive in isolation. Measured across all 34 names over weeks, post volume has a rank correlation of -0.00 with the direction of the following day's move. A mention count measures how much a name is discussed, not which way it goes.
Which neocloud stocks does ChipSentiment track?
Six: Nebius (NBIS), CoreWeave (CRWV), IREN, Applied Digital (APLD), TeraWulf (WULF) and Cipher Mining (CIFR). The first three sit closer to the cloud layer, selling accelerated compute; TeraWulf and Cipher sit closer to the power-and-campus layer, generally converting sites originally developed for bitcoin mining.
Sources
- ChipSentiment - neocloud segment - the six constituents, live prices and cap-weighted sentiment history
- ChipSentiment - live board - prices, one-day and five-day moves for all 34 names, quoted as at the close on 12 August 2026
- ChipSentiment - X mentions board - daily cashtag post counts, counted by this site against the X API; latest complete day 11 August 2026
- ChipSentiment - how the sentiment score is calculated - the six components, their weights and their bounds
- ChipSentiment - do X mentions predict stock moves? - the -0.00 rank correlation between post volume and next-day direction
Figures are taken from the public filings and the reporting linked above.