The rotation that didn't happen
Prices up and dominance up is not rotation. It is leveraged beta, and it unwinds together.
Thesis
The standard model of a crypto cycle says capital moves down the risk curve in sequence: Bitcoin first, then Ethereum, then large-cap alts, then the long tail. Bitcoin dominance falls as it happens. That falling dominance is the confirmation signal.
August 2026 delivered the price action without the signal. Altcoins rallied hard, in some cases harder than Bitcoin, while Bitcoin dominance rose to roughly 61% before easing to around 59%, near its highest level of the year.
Prices up and dominance up simultaneously does not describe rotation. It describes leveraged beta. Capital was not moving down the risk curve; it was entering the whole complex, and Bitcoin was still taking the larger share. The distinction determines whether you are trading a durable regime change or a correlated bounce that will unwind together.
The evidence for rotation
The bull case has real data behind it, and it should be stated fairly.
Breadth improved dramatically. CryptoQuant analyst Darkfost found that 56% of Binance-listed altcoins reclaimed their 200-day moving averages, a sharp reversal from months when 80–85% traded below that line.
Aggregate capitalisation expanded. TOTAL2, the altcoin market excluding Bitcoin, added roughly $215 billion between 19 and 22 August and pushed back above $1 trillion. Total crypto capitalisation recovered to around $2.75 trillion.
Relative performance favoured alts. Over the seven days to 25 August: XRP +43.7%, Ethereum +28.6%, Solana +25.6%, all beating Bitcoin's +22.6%. Chainlink added more than 30%. Zcash rose roughly 75% and Aave more than 60%.
The long tail participated. Bitcoin dominance opened September lower after being rejected at the 60% level, while the market capitalisation of altcoins outside the top ten rose more than 10% from the month’s start and cleared $200 billion.
Ethereum posted its best month in a long time. ETH gained 32.6% in August, closing 31 August at $2,467.65, and the ETH/BTC ratio moved toward 0.030 from a May low near 0.024–0.027.
Taken alone, that is a convincing rotation picture.
Why it does not confirm
Three counters, in order of importance.
One: dominance rose during the rally. In a genuine rotation, dominance falls because capital exits Bitcoin to fund positions further out the curve. Here it climbed to ~61%. Whatever bid lifted alts also lifted Bitcoin, proportionally more. That is a market-wide risk appetite change, not a rotation.
Two: relative performance is still poor. The altcoin season index sits near 30, meaning the majority of alternative assets continue to underperform Bitcoin over the measurement window. A handful of large weekly gains does not move that reading. Capital remains concentrated in Bitcoin.
Three: Ethereum's institutional bid never arrived. This is the most damaging point for the sequential model. Spot Ethereum ETFs saw outflows in most months of 2026, with May the worst at roughly $541 million. When money returned to Bitcoin during the summer lows, it largely skipped ETH entirely.
If the rotation model held, Ethereum would be the first stop after Bitcoin. It was not a stop at all for most of 2026.
Ethereum: current structure
ETH trades at $2,465 after peaking near $2,507 intraday on 8 September.
| Level | Role |
|---|---|
| $2,185–$2,203 | 50-day and 200-day EMA cluster |
| $2,334 | Monthly downside boundary |
| $2,394–$2,400 | 20-day EMA, primary support |
| $2,475 | Key on-chain support |
| $2,511–$2,546 | Immediate resistance |
| $2,626 | Next resistance |
| $2,723–$2,822 | Major supply zone |
Scenarios for September:
| Case | Range | Confirmation |
|---|---|---|
| Bearish | $2,180–$2,400 | Loses the 20-day EMA near $2,394, fails to reclaim $2,400, exposes the EMA cluster |
| Base | $2,400–$2,700 | Holds $2,394–$2,400, establishes above $2,510, momentum insufficient for $2,800 |
| Bullish | $2,700–$2,800 | Closes decisively above $2,510, holds $2,600, continues through $2,700 with spot demand |
The flow picture is deteriorating at the margin. US spot Ethereum ETFs attracted roughly $218.41 million last week against approximately $824.42 million the week before, a decline of about 73.5%. Daily prints in early September were uneven: $8.6m inflow on 1 September, $48.2m outflow on 2 September, then $141.4m on 3 September and $25.9m after.
Supply pressure is concrete. Lookonchain tracked a whale that had accumulated 167,855 ETH and appeared to sell the entire position over about five days, worth roughly $408 million. Partially offsetting this, about 116,000 ETH worth more than $300 million left exchanges, and spot ETH ETFs recorded $1.85 billion of inflows in August.
The picture is a market absorbing large discretionary supply with institutional demand that is positive but decelerating fast.
Where the real performance was
The interesting observation of 2026 is that the assets that worked were not the ones the rotation model points to. They had specific, identifiable catalysts.
Hyperliquid (HYPE) was the outlier of the year. It set an all-time high of $88.04 on 3 September with a market capitalisation around $21.8 billion, outperforming BTC, ETH and SOL over recent weeks. Critically, the protocol generates real revenue, roughly $2.8 million.
That is a cash-flow story, not a beta story, which is why it decoupled upward while ETH lagged. It also carries a dated, quantifiable risk: a token unlock of approximately 14.2 million tokens (~$1.2 billion) on 29 September, with roughly 47% going to insiders.
The privacy complex ran on a regulatory-access catalyst rather than sentiment. Zcash reached $838.78 (market cap $14.18bn) and Monero $536.77 (market cap $10.13bn). ZEC cleared $750 on 22 August ahead of a Grayscale ZCSH listing on NYSE Arca on 25 August that drew roughly $14.8 million.
Both winners share a property the rotation model ignores: an asset-specific catalyst independent of where capital sat on the risk curve. In a market without a genuine rotation, that is the only thing that generates differentiated returns.
Solana traded near $103.33 with a $60.44 billion capitalisation, having approved SIMD-0550 on 28 August. XRP sat at $1.32.
The confirmation framework
Rather than asserting a view, here are the falsifiable conditions I would require before treating rotation as confirmed.
| Condition | Threshold | Status |
|---|---|---|
| BTC dominance sustained below 60% | Was rejected at 60%, now ~59% | Partial |
| BTC dominance breaks below 58% | Not reached | Not met |
| TOTAL2 holds above $1tn | Above, achieved in August | Met |
| Altcoin season index above 50 | Currently ~30 | Not met |
| ETH/BTC sustained above 0.030 | Approaching, not held | Partial |
| ETH monthly close above $2,510 | Tested 8 September, not secured | Not met |
| ETH ETF weekly inflows re-accelerating | Down 73.5% week over week | Not met |
Three of seven partial or met. On that basis the correct classification is a Bitcoin-led beta rally with broad participation, not a rotation. The practical consequence is that altcoin positions in this tape carry Bitcoin’s direction plus their own drawdown profile, without the compensating outperformance a real rotation provides.
What I take from this
Dominance is the signal; price is the noise. Altcoin prices rose. The rotation thesis was still wrong. Reading the level and ignoring the ratio would have produced a confident and incorrect call.
Sequential rotation is a model, not a law. 2026 falsified it. Ethereum, the model’s mandatory second stop, saw ETF outflows in most months of the year while Bitcoin absorbed the institutional bid. Models built on prior-cycle sequencing need to survive the arrival of new capital structures. This one did not.
In a non-rotating market, only catalysts differentiate. HYPE had revenue. ZEC had an index listing. Everything else moved with beta. That is where research effort belongs when the tide is not doing the work.
Write down invalidation conditions before the trade, not after. The table above is more useful than any directional call, because it converts "is altseason here" from a sentiment question into seven observable series.
Data as of 10 September 2026. Sources: Glassnode, CryptoQuant, Lookonchain, Farside Investors, SoSoValue, CoinGecko, Cryptorank, KuCoin Research, CryptoTimes, TradingView, Yahoo Finance. Analysis is for research purposes and is not investment advice.