How far along are we in this crypto bear market? Fidelity’s Q3 Signal Report provides a set of indicators: the weighted NUPL has dropped to -0.01, BTC dominance has risen to 68%, and multiple metrics are nearing historical capitulation zones. Referring to the approximately 300-day bottoming periods in 2018 and 2022, the current 203-day correction may have already completed two-thirds. The report notes that October 2026 is a time window to watch, but does not constitute a bottom prediction.
Author: Fidelity Digital Assets Research Team
Compiled by Jiahuan, ChainCatcher
I. Market Overview

Chart: Overview Dashboard of Three Major Asset Signals
Weighted NUPL: BTC Alone Supports the Market
Weighted NUPL measures whether a market-cap-weighted portfolio of digital assets is, as a whole, in a state of unrealized profit or unrealized loss. Because BTC represents a significantly larger share of market capitalization than ETH and SOL, the current indicator is largely driven by BTC.
Currently, only BTC among the three assets remains in unrealized profit, while ETH and SOL are both in unrealized loss. After weighted calculation, the NUPL stands at -0.01, indicating that the market as a whole is slightly below the break-even line.
In other words, the remaining unrealized gains in the market are primarily concentrated in BTC, rather than reflecting broad-based improvements across various assets. BTC has acted as a stabilizer, partially offsetting the selling pressure from ETH and SOL, but not enough to bring the overall portfolio back into positive territory.
If the three assets have equal weights, the portfolio's condition is weaker due to the deeper losses in ETH and SOL. Therefore, despite the weighted NUPL having turned negative, the current market structure is slightly healthier than if losses were evenly distributed across all assets.
This further highlights BTC's role as a barometer in the digital asset market. ETH and SOL have continued to weaken relative to BTC, as investors show a clearer preference for the largest and most liquid asset, while remaining cautious toward other digital assets with historically higher volatility.
For investors, the current market resembles a search for a bottom during a repair phase, rather than a broad profit-taking stage at the end of a cycle. BTC’s relative strength has not spread to other assets, and market participation remains highly concentrated.
In the second quarter of 2026, BTC's NUPL further declined, and the weighted NUPL of the three assets dropped to -0.01. Before more assets re-enter positive unrealized profit territory, the market is more likely to remain range-bound or continue under pressure, rather than rapidly rebound into full expansion.

Chart: Weighted NUPL Score
Bitcoin's dominance has risen to 68%, but asset rotation has yet to emerge.
After a prolonged decline in the second half of 2025, the BTC dominance has begun to slowly rise again, and the long-term upward trend remains intact.
Historically, an increase in BTC dominance has often been accompanied by other digital assets underperforming BTC. This reflects a tendency for capital to flow toward the most liquid and mature market assets during periods of heightened uncertainty and valuation pressure.
The rising阶段性 low points of BTC dominance and the relatively stable upward slope indicate that this preference is not a short-term phenomenon.
The current dominance rate indicates that capital remains concentrated in BTC, with market risk appetite showing clear selectivity. Assets outside of BTC have limited participation, and the broader market has not yet regained widespread relative strength.
In the second quarter of 2026, BTC's dominance rose slightly from 67% to 68%, with minimal signs of capital rotating into other digital assets.
If the BTC dominance rate begins to decline or enters a sideways trend in Q3, it may indicate that risk appetite is recovering, other digital assets are regaining capital inflows, and could serve as an early signal of shifting market structure.

Chart: Asset Dominance Ratio and Total Market Capitalization of Digital Assets, Excluding Stablecoins
Asset performance: BTC, ETH, and SOL all declined
Over the past year, BTC has declined by approximately 45%, ETH by 37%, and SOL by 53%.
Year to date, three assets have performed weakly alike: BTC, ETH, and SOL have declined by 33%, 47%, and 41% respectively.
By the end of the second quarter, BTC had fallen below the 200-week moving average, with market sentiment at an extremely low level and the entire digital asset market weakening in tandem.
Unfavorable macroeconomic conditions, capital shifting toward AI-related investments and stock markets, combined with sustained weakening of market momentum, have collectively amplified this downturn. Currently, multiple indicators have approached or reached historical panic zones.
Spot ETP flows have also remained negative. Year-to-date in 2026, spot ETPs have seen a cumulative net outflow of $5.4 billion, with $4.9 billion occurring in the second quarter.
In just June, spot ETPs experienced net outflows of approximately $4.5 billion, the worst month since the launch of spot Bitcoin ETPs.
Between June 1 and 4, nearly $6 billion in cascading liquidations occurred in the market. Passive liquidations further amplified selling pressure and disrupted existing position structures.
The macroeconomic and geopolitical environment has also been a drag. Inflation remains elevated, energy prices continue to face pressure, and markets have significantly adjusted their expectations for monetary policy: at the beginning of the year, markets believed rates might not be cut until 2026; by the end of the second quarter, expectations had shifted toward the possibility of rate hikes by year-end.
Short-term volatility remains high, and forming a market bottom typically takes time.
However, the current valuation levels, along with the growing negative correlation between digital assets and traditional risk assets, may also provide attractive entry points for long-term investors, provided that the adoption trends of the underlying networks continue to strengthen.

Chart: 1-year rolling performance of BTC, ETH, SOL
II. Bitcoin
NUPL 0.09: Positive
At the end of the second quarter, BTC's NUPL was 0.09, falling within the "Hope—Fear" range, indicating that there is still a small amount of unrealized profit, but investor sentiment is becoming more cautious.
Some holders are still in profit, but the market has not yet reached a general consensus that the bottom has been established.
In Q2, BTC fell 14%, and the NUPL dropped from 0.21 to 0.09, a decrease of 0.12. This resembles a moderate contraction of holders' unrealized profits rather than a large-scale capitulation.
Based on current data, BTC's market price is approximately 10% above the network-wide composite cost basis, with investors holding a combined $108 billion in unrealized profits.
For most of April and May, BTC's NUPL was in the "Optimism–Anxiety" zone, as the market gradually came to believe that the bottom may have been reached.
Recent indicators have retreated back into the "hope-fear" range, indicating that market sentiment has shifted from sustained optimism to caution and uncertainty.
Historically, during BTC bear markets, the NUPL has fallen further into the surrender zone, so the current state should still be viewed conservatively.
Historical data shows that when BTC's NUPL is within the range of 0.09 ± 0.05, the median return over the subsequent year is 53%, with a three-year annualized compound growth rate of 69%, based on 73 occurrences.
The correlation coefficients between NUPL and future one-year and three-year returns are -0.26 and -0.80, respectively, indicating that lower NUPL values are often associated with higher long-term subsequent returns. This is why Fidelity considers a lower NUPL a positive signal.
However, historical relationships may weaken or become invalid, and must be evaluated in conjunction with the macroeconomic environment and overall market structure.

Chart: Bitcoin Net Unrealized Profit and Loss (NUPL)
Momentum signal: Negative
The current momentum indicators for BTC reflect that this downturn has formed a bearish impulse, with the price failing to sustain higher highs over the past quarter.
The signal was upgraded on April 21, 2026, when BTC was priced at $78,317, with both short-term volatility and price momentum rising above their medium-term averages.
However, this rally failed to sustain. On June 1, the signal turned negative again when BTC was at $66,540, indicating that momentum had faded and the price failed to hold.
In the second quarter, BTC's price fluctuated between $58,500 and $82,256, with particularly intense volatility in April and May.
This quarter, the momentum model produced one false positive: the positive volatility at the beginning of the second quarter was initially interpreted as potentially continuing, but it quickly reversed.
This is an unavoidable cost of trend-following models. Such models aim not to accurately capture every top and bottom, but to participate as much as possible after an upward trend emerges, while managing downside risk.
Since the signal turned negative on June 1, BTC has declined by approximately 10% and remains in a negative momentum zone.
BTC's short-term realized volatility is approximately 34% annualized, below the 40% medium-term volatility. For the momentum signal to turn positive again, either short-term volatility must moderately rebound, or medium-term volatility must decline further.
It is important to note that this indicator is not used to predict exact tops and bottoms, but rather to identify phases where price direction and volatility move in sync. Historically, such phases often correspond to market acceleration.
The current reading still points to caution, rather than a restoration of upward momentum.

Chart: Bitcoin Momentum Signal
Benchmark: Positive
BTC price has fallen more than 50% from its all-time high, but the network hash rate has only decreased by approximately 22% from its peak.
Miners are facing clear pressure, but the entire network continues to demonstrate strong resilience.
As a result, the Bitcoin Yardstick has approached its historical low, suggesting that BTC may currently be significantly undervalued relative to the computing power required to maintain and secure the network.
However, this cycle also has some differences, including lower price volatility and a more mature mining industry.
BTC price is a direct input variable for Yardstick. When price volatility decreases, the relative impact of hash rate on this ratio becomes more pronounced.
Meanwhile, mining companies have improved their ability to manage energy costs and operational efficiency, allowing them to more flexibly adjust operating rates, relocate equipment, or optimize power contracts based on profitability.
As a result, mining capacity can adjust more flexibly to price changes, and large deviations between price and network energy input are less likely to occur as they did in the past.
Falling prices and declining hash rate have jointly pushed Yardstick into the "undervalued" zone. Over the past 92 days, the indicator was below one standard deviation from the long-term average for 76 days, or approximately 83% of the time.
Historically, this area has typically corresponded to a market accumulation phase or a relative bottom.
In 2018, Yardstick remained in a similar range for 298 days; in 2022, it lasted 299 days before market sentiment gradually recovered.
This bear market has lasted approximately 203 days so far. For investors monitoring cyclical patterns, October 2026 may serve as a noteworthy time window to observe, though it does not guarantee the market will bottom out at that time.

Chart: Bitcoin Yardstick
Performance relative to gold: Negative
One of the core investment theses for BTC is its potential as a store of value.
Pricing in gold allows you to measure BTC's performance relative to another traditional physical safe-haven asset, rather than just observing its price changes against fiat currencies.
Recent price volatility has indeed weakened the store-of-value narrative for BTC in the short term, but it is not enough to prove that its long-term investment thesis has failed.
In the second quarter of 2026, despite BTC denominated in USD falling by 14%, its price relative to gold remained largely unchanged.
After approximately one year of consistent underperformance, investor preference between BTC and gold may be rebalancing.
Since February 28, BTC has risen 15% relative to gold. Amid ongoing geopolitical uncertainty, their relative performance has begun to stabilize.
The strong performance of gold over the past year has largely come from continuous purchases by central banks worldwide. Against this backdrop, the relative relationship between BTC and gold may be forming a more stable equilibrium.
Meanwhile, on-chain metrics still point to an accumulation phase, with some indicators even showing signs of capitulation.
For long-term allocators, BTC’s previous underperformance relative to gold may make its current valuation more attractive.

Chart: Bitcoin performance in USD versus gold
Hashrate: Negative
BTC's daily and 30-day average hash rates remain below the historical milestone of 1,000 EH/s, or 1 ZH/s, set in September 2025.
There are primarily two reasons behind this: first, AI and high-performance computing businesses are increasingly attracting demand for electricity, land, and data center resources; second, the bear market has compressed profit margins for Bitcoin mining.
Some mining companies may be redirecting power capacity, data center infrastructure, and new capital expenditures toward AI or high-performance computing businesses.
When BTC prices are low, AI computing power contracts often provide more stable and predictable income, making it economically rational for mining companies to reconfigure their infrastructure.
It should be noted that Bitcoin ASIC miners are highly specialized hardware and generally cannot be directly used for AI computing. Therefore, mining companies transitioning to AI are primarily repurposing their electricity, land, data centers, and cooling infrastructure, rather than directly converting existing mining rigs into AI servers.
Idle mining machines are also unlikely to be parked long-term. More commonly, the equipment is sold or relocated to areas with lower electricity prices, rather than permanently exiting the network.
Since BTC prices peaked in October 2025, hash rate has consistently declined. Meanwhile, mining difficulty has remained elevated for an extended period without adjusting downward in line with the price.
Price declines and compressed profit margins have led miners operating at the cost edge to gradually shut down.
Historically, deterioration in mining economics during bear markets has led to temporary declines in hash rate. However, competition from AI data centers and power contracts for infrastructure resources may cause this round of hash rate decline to last longer than the historical average.
In the second quarter, average daily hash rate increased by 8% quarter-over-quarter, but the 30-day average hash rate decreased by 6% year-over-year.
Since 2026, BTC's price has dropped more than 29%, while hash rate has declined by only about 12%, indicating that although miner economics are under pressure, the network as a whole remains somewhat resilient.

Chart: Bitcoin Hash Rate and Mining Difficulty
Three: Ethereum
NUPL -0.43: Positive
In the second quarter, ETH's NUPL continued to decline within the "capitulation" zone, further worsening market sentiment.
Within the quarter, the price of ETH fell by 25%, and the NUPL dropped from -0.12 to -0.43, a decrease of 0.31, indicating a significant expansion of investors' unrealized losses.
Based on current data, ETH is trading approximately 30% below the network-wide cost basis, with a total unrealized loss of around $87 billion.
On June 6, ETH's NUPL reached a temporary low of -0.46 and has not dipped below that level since.
Although the indicator may still decline to new lows, this level has currently held, which could be a noteworthy position for long-term investors.
Historically, when ETH's NUPL has been within the range of -0.43 ± 0.05, subsequent returns have typically been substantial.
Since 2018, the median future one-year return corresponding to similar readings has been 70%, with a three-year annualized compound growth rate of 133%, each time horizon comprising 90 observations.
The correlation coefficients between NUPL and future one-year and three-year returns are -0.13 and -0.81, respectively, indicating a stronger negative relationship with long-term subsequent returns.
This is also the basis for Fidelity regarding a lower NUPL as a positive signal.
However, historical patterns may weaken or become invalid, so it is still necessary to assess them in conjunction with broader market conditions.

Chart: Ethereum Net Unrealized Profit and Loss (NUPL)
Momentum signal: Negative
ETH's momentum signal turned positive on April 16, 2026, at a price of $2,350, with both volatility and price momentum above their respective medium-term averages.
However, ETH failed to sustain its upward momentum. On May 17, the signal turned negative again as the price retraced to $2,130.
In the second quarter, ETH fluctuated between $1,564 and $2,422, with particularly volatile movements in April and May.
Similar to BTC, ETH's momentum model also produced a false positive this quarter.
Since the signal turned negative on May 17, ETH has declined by approximately 25% and remains in a negative momentum zone.
Short-term realized volatility remains around 50% annualized, significantly lower than the medium-term volatility of 71%.
For the signal to return to positive, either short-term volatility must rise significantly, or medium-term volatility must decline substantially.
The current indicators reflect a simultaneous weakening of price and volatility, not a restoration of upward momentum.

Chart: Ethereum Momentum Signal
Use indicator: Neutral
In the second quarter, the fundamentals of Ethereum's base layer generally cooled, consistent with the continued decline in ETH price and falling volatility.
However, the total number of transactions still showed resilience, declining by only 5% during the quarter, with daily transaction volumes consistently remaining above 2 million.
Trading fees remain well below the historical average but are still sensitive to short-term demand fluctuations.
On April 22, Ethereum's median transaction fee rose to $0.42 before steadily declining to around $0.02 by the end of the quarter.
Active and new addresses declined significantly after reaching all-time highs in the previous quarter, dropping by 10% and 31%, respectively.
Despite weak price performance during the quarter, on-chain activity remains strongly correlated with asset prices.
In the second quarter, the proportion of addresses with real economic activity slightly increased, but a significant portion of addresses still did not make substantial contributions to Ethereum's revenue or security.
This trend may continue into 2026. The upcoming Glamsterdam upgrade, planned for later release, will focus on increasing base layer capacity, potentially further increasing block space supply.

Chart: Ethereum Usage Metrics
Stablecoin transfer volume: Positive
Driven by advancements in scaling and an improved regulatory environment, stablecoin transfer volumes on Ethereum have surpassed historical averages.
Over the past 12 months, the total volume of stablecoin transfers has consistently set new records, exceeding $20 trillion in total.
However, the growth rate has shown signs of slowing. In June, the daily average stablecoin transfer volume was about 9% lower than in the previous three months.
Over the past year, the stablecoin market has experienced rapid expansion, and growth rates over the next year may gradually return to more stable levels.
Notably, despite the overall decline in digital asset prices, stablecoin transfer volumes on Ethereum continue to grow.
This indicates that the real usage demand for stablecoins is gradually moving away from market sentiment and asset prices, and is increasingly being used for payments, settlements, and accessing on-chain USD globally, rather than solely serving speculative trading.
The average transfer cost of stablecoins has remained below $1 for three consecutive quarters, confirming the real-world effectiveness of previous scaling measures.

Chart: Total Ethereum stablecoin transfer volume
Network fee: Negative
Over the past year, fees on the Ethereum network have continued to decline.
In the second quarter, the rolling 12-month network fees decreased by 15% from $344 million to $294 million.
The pace of scaling at the protocol and infrastructure levels continues to outstrip the growth in block space demand. As developers refocus their efforts on base-layer scaling, this trend may be long-lasting.
The upcoming Glamsterdam upgrade is expected to further increase block space capacity, so Ethereum network fees may still face downward pressure over the next year.
Ethereum's fee levels themselves fluctuate significantly, making it difficult to determine a reliable long-term equilibrium value.
In the second quarter, Ethereum's daily network fees fluctuated between $145,000 and $2.75 million, averaging approximately $575,000 per day.
One key signal investors need to watch in the coming years is how core developers balance network growth with value capture.
As a technology platform still in development, Ethereum has previously prioritized user adoption, ecosystem expansion, and network utility over short-term revenue.
Unless developers and researchers invest more effort into improving value capture mechanisms, network fees and protocol revenues may continue to face pressure.

Chart: Ethereum Network Fees
Four, Solana
NUPL -0.72: Positive
In the second quarter, SOL's NUPL remained in the "capitulation" zone.
Within the quarter, the SOL price fell by 12%, and the NUPL decreased from -0.67 to -0.72, a drop of 0.05, further expanding unrealized losses.
Based on current data, SOL is trading approximately 41% below the network-wide cost basis, with a total unrealized loss of around $29 billion.
On June 6, SOL's NUPL rebounded significantly from a temporary low of -1.08, once again highlighting SOL's high volatility during this bear market.
The recovery from the low point may indicate that a significant portion of early holders have sold their positions, while new investors are beginning to take over at lower prices.
Historically, SOL's NUPL has rarely fallen within the range of -0.72 ± 0.05, but market performance following such instances has been strong.
Since the launch of the Solana network, similar readings have occurred 21 times. The median future one-year return is 542%.
Due to insufficient historical data, a reliable three-year return cannot be calculated at this time.
SOL's current NUPL has a correlation coefficient of -0.56 with return rates over the next year, indicating a relatively strong negative correlation.
However, Solana has a shorter operational history and limited sample size, so this historical relationship must be viewed with caution and may not be replicable in the future.

Chart: Solana Net Unrealized Profit and Loss (NUPL)
Momentum signal: Negative
The momentum signal for SOL has been negative since October 28, 2025, with price and volatility moving in tandem downward; the market environment remains cautious.
However, SOL’s short-term realized volatility has recently risen above its medium-term volatility, at approximately 63.5% and 61% annualized, respectively.
Historically, this condition has sometimes occurred before or after a momentum reversal. If the price strengthens in tandem, it will provide stronger support for the formation of a阶段性 bottom.
Notably, SOL did not generate false positive momentum signals this quarter, unlike BTC and ETH.
For most of the second quarter, SOL traded between $63 and $97, starting the quarter at approximately $81.
According to SOL's own model parameters, even though the price rose temporarily to $97, the momentum was insufficient to turn the signal positive, after which the price made a new lower low.
On October 28, 2025, when the signal first turned negative, SOL's price was approximately $194. Since then, the price has dropped by about 60%, and the correction is not yet fully complete.
The current indicators are closer to an attempt to stabilize rather than a restoration of upward momentum.

Chart: Solana Momentum Signal
Use indicator: Positive
Despite the bear market, Solana's fundamentals have shown strong resilience. While asset prices continue to weaken, on-chain activity demand has not declined proportionally.
The number of monthly transactions continues to rise, increasing 1% quarter-over-quarter in the second quarter, with potential to set a new all-time high.
However, the growth rate of users has been slower than the growth in network activity, indicating that the average number of transactions per user is increasing.
In the second quarter, monthly active addresses and new addresses decreased by 15% and 7%, respectively, while addresses with actual economic activity declined by only 4%.
On-chain activity with economic significance remains relatively stable in the current market environment, presenting a certain contrast to Ethereum's trend.
Another advantage of Solana is its stable transaction fees.
Throughout the quarter, Solana's median transaction fee remained consistently below 0.1 cent, with minimal fluctuations, providing strong cost predictability for users and investors.

Chart: Solana Usage Metrics
Stablecoin transfer volume: Positive
Stablecoin transfer volumes on Solana have historically been volatile, but the long-term upward trend remains intact since a significant increase over a year ago.
Currently, Solana's daily average stablecoin transfer volume has stabilized above $8.4 billion, representing a 17% month-over-month increase.
The average stablecoin transfer amount on Solana is smaller compared to Ethereum, reflecting differences in user demographics and usage patterns between the two networks.
Over the past 12 months, Solana has processed over $2.6 trillion in stablecoin transfers.
While SOL's price experienced a significant decline, stablecoin trading volume and other on-chain activity remained stable.
Similar to Ethereum, a significant portion of stablecoin demand on Solana exhibits strong retention and is largely unaffected by short-term market fluctuations.
If on-chain activity continues to grow, Solana’s stablecoin ecosystem is also expected to expand accordingly.
In the second quarter, the total size of the stablecoin market decreased by approximately 1.3%, but stablecoin supply on Solana increased by about 3%, adding around $300 million.

Chart: Total Solana stablecoin transfer volume
Network fee: Neutral
Solana's network fees continue to decline, but signs are emerging that they are gradually finding an equilibrium level.
In the second quarter, the rolling 12-month network fees decreased by 18% to $221 million, with daily average fee revenue of approximately $390,000.
If the second-quarter network fees are annualized, the annualized revenue is approximately $141 million, and the gap with the rolling 12-month data has significantly narrowed.
Meanwhile, on-chain activity on Solana continues to grow, and several Solana Improvement Proposals (SIPs) are beginning to place greater emphasis on enhancing value capture for SOL holders.
These advancements further solidify Solana’s position as a technology platform capable of generating protocol revenue, with SOL at the core of its economic system.
According to Fidelity's assessment, Solana's network fee income may have approached a cyclical low.
If on-chain activity continues to grow and proposals related to value capture are gradually implemented, Solana’s fee revenue is expected to begin recovering within the next year.
Appendix: Indicator Method Description
Kinetic signal
The momentum signal assesses the current momentum state of a digital asset by comprehensively evaluating price trends and changes in volatility.
The model examines the relationship between short-term price movements and long-term trends, while also determining whether current volatility is expanding or contracting relative to recent benchmarks.
These two dimensions will be combined into a single momentum category to identify phases where price and volatility move in the same direction, diverge, or are undergoing a transition.
The model's lookback window and other parameters are selected through an optimization process to achieve relatively clear and stable distinctions across different market conditions.
However, this indicator is only used to describe the current market condition and does not constitute price predictions, investment advice, or trading signals.
Yardstick
The Bitcoin Yardstick, also known as the Hash Rate Yardstick, can be roughly understood as the price-to-earnings ratio metric for the Bitcoin network.
Traditional P/E ratio divides stock price or market capitalization by corporate earnings, while Yardstick divides Bitcoin’s total market capitalization by the network’s total hash rate to measure Bitcoin’s market value relative to network security investment.
The lower the ratio, the cheaper BTC is relative to the computing power required to secure the network, similar to how a lower price-to-earnings ratio is typically interpreted as a lower stock valuation.
However, hash rate does not equate to corporate profitability, so Yardstick can only serve as a relative valuation framework and cannot be directly equated with a stock’s price-to-earnings ratio.
NUPL
Over one- and three-year time horizons, the NUPL relationship with future returns is one of the strongest on-chain indicator relationships identified in Fidelity's research.
However, the network histories of Ethereum and Solana are significantly shorter than that of Bitcoin, resulting in fewer available observational samples and thus lower reliability of historical relationships.
Theoretically, when a network's realized market cap exceeds twice its total market cap, the NUPL may fall below -1.0.
In the early days of BTC, ETH, and SOL, large amounts of supply were moved or distributed without public market prices.
For example, early portions of BTC were transferred before the market price was established; ETH and SOL experienced early token issuance, presales, pre-mining, seed funding rounds, and foundation allocations.
These early allocations can affect the realized price, sometimes being accounted for at a cost higher than subsequent public market prices.
When the market price falls below the realized price, the network-wide cumulative unrealized loss may exceed the current total market cap, driving the NUPL below -1.0.
As the network matures and on-chain transaction history accumulates, realized market cap will increasingly reflect actual market transactions rather than early allocation events.
Therefore, the reference value of NUPL typically increases as the network matures.



