Ethereum Wedged Between $1,850 and $1,955 With Mainnet Fees Down 99%

Ethereum (ETH-USD) traded $1,922.21 at 4:54 a.m. ET Wednesday, up 1.25%, and held $1,920.67 into the U.S. morning for a 1.02% gain. Market capitalization sits near $233 billion, second only to Bitcoin and roughly $50 billion ahead of Tether at $183 billion. BNB trails at $82 billion, XRP near $63 billion, Solana at $44 billion.

The month has been a grind rather than a trend. ETH opened August at $1,867.23 with a $225.34 billion capitalization, touched $1,923.13 on August 10, fell back to $1,887.60 by August 13, printed a session as narrow as 0.9% between $1,859.80 and $1,876.20, and has now returned to the top of that band. Total range for the month covers roughly 3.4%.

The number that frames everything: ETH sits approximately 61% below its all-time high of $4,951.66, set August 24, 2025. That is a deeper drawdown than Bitcoin's 49% and the widest divergence between price and network fundamentals in the asset's history.

Wednesday's bid came from the same source that lifted Bitcoin and gold — the 10-year Treasury eased to 4.70% and the 30-year retreated from a multi-decade high of 5.338%, while the Dollar Index fell toward 99.00. Solana led the majors at $77.41, up 1.52%. Bitcoin held $64,821.82, up 1.03%. XRP sat at $1.00. The CoinDesk 20 gained 0.79%.

Two things are scheduled and both matter. The July FOMC minutes land at 2:00 p.m. ET. And the Plataberget public testnet forks Glamsterdam on August 20 — tomorrow — beginning the first live test of the upgrade designed to fix the exact problem that has kept ETH at $1,920 while the network processed more activity than at any point in its history.

The ETH/BTC ratio sits at 0.0296. That is the number the entire structural debate reduces to.

The Range Is $1,850 to $1,955 and Both Ends Are Defended

The technical picture is a compressed moving-average ribbon with price wedged inside it, which is the textbook definition of a market in balance.

The 7-day moving average sits at $1,877.73, the 14-day at $1,889.69, the 30-day at $1,858.44. On the exponential side, the 20-day EMA reads $1,868 and the 50-day sits near $1,850. The 100-day EMA at $1,924 is directly overhead and functions as the first genuine trend hurdle. Price is currently pressed against it.

The 20-day and 50-day EMAs had compressed to just $17.60 apart during the quietest stretch of the month before widening to $20.90 — the first evidence that the two-week coil has begun resolving upward. That widening, paired with a MACD histogram that halved to -5.40 and a MACD line turning back up from 16.7 to 20.1, describes momentum reversing rather than momentum running.

Support is layered and specific. The 50-day EMA near $1,850 aligns with the 23.6% Fibonacci retracement at $1,850.21, making that the single most defended level on the chart. Beneath it, $1,842 is the next marker, then $1,810.85 as the deeper structural floor, then the $1,800 psychological level.

Resistance stacks tightly. The $1,900 round figure has already been cleared. The 100-day EMA at $1,924 is the current fight. Above it, the swing high near $1,955 is the objective, then $2,009 as the first major uptrend resistance, then $2,055.47 and $2,133.

Projected weekly range runs $1,850 to $1,975.

Momentum readings split by lookback, which is the honest read of a market with no direction. The 14-day RSI reads 42.42 in one measurement and 50.39 in another, with a post-breakout reading of 56.01 — neutral by every framing. The weekly RSI sits at 62.00. The Money Flow Index at 38.57 shows capital still leaning outward without reaching the sub-30 territory that marks capitulation.

Across 17 technical indicators, 3 call buy and 9 call sell on the short-term view. Neutral, coiled, and waiting.

The Value-Capture Problem That Killed the Burn

The reason ETH trades at $1,920 while network activity sits at record levels is a single mechanical failure, and it has a number attached.

EIP-4844 made Layer 2 the default execution environment. Blob transactions cut L2 costs by 90% to 99%. The practical result: mainnet became a settlement layer and Layer 2 became where users actually transact. That was the intended design. It also broke the economics.

When mainnet fees ran near $30 million a day, the EIP-1559 burn was substantial and ETH became scarcer as the network grew — the deflationary thesis that underpinned every valuation model written between 2021 and 2024. When fees fall to single-digit millions because users migrated to rollups paying a fraction of mainnet costs, the burn withers and that thesis breaks entirely.

Ethereum fees have dropped roughly 99% from their peak. The network processes more transactions than ever. The base asset captures almost none of the value.

Standard Chartered's Geoff Kendrick put a figure on it: Base alone removed approximately $50 billion from ETH's market capitalization. That estimate drove the bank to cut its year-end target from $10,000 to $4,000 in March 2025, warning of a structural decline caused by rollups siphoning fee revenue from mainnet.

Vitalik Buterin publicly questioned the pace of the rollup-only approach in early 2026, noting that only a couple of major Layer 2s had reached meaningful decentralization. That admission triggered the strategic reorganization: the Ethereum Foundation restructured its roadmap into three tracks — Scale, Improve UX, and Harden the L1 — an explicit acknowledgment that the base layer needed to reclaim economic relevance.

21Shares frames the current state precisely: ETH issuance is modestly positive during low-fee periods, and price now depends on whether network activity translates into sustainable economic throughput. Ethereum can win on adoption while ETH lags as an asset. That is the gap the market is pricing, and it is why positive ETF flows coexist with a sub-$2,000 price.

Every argument for higher ETH runs through fixing this. Every argument against assumes it stays broken.

Glamsterdam Forks on Plataberget Tomorrow

The Ethereum Foundation launched the Plataberget public testnet as the early testing ground for Glamsterdam, with the fork on that network set for August 20. It is permissionless — validators, developers and anyone else can join and test systems before the changes move to longer-running networks like Sepolia and Hoodi.

The upgrade bundles roughly ten EIPs, with two headline proposals defining it. EIP-7732 enshrines Proposer-Builder Separation, letting validators safely outsource block building and bringing builder auctions on-chain — cutting reliance on a handful of opaque relay companies and making MEV transparent. EIP-7928 introduces Block-Level Access Lists, mapping transaction dependencies upfront to enable parallel execution and faster syncs.

The scaling numbers are the headline. The block gas limit is designed to rise from approximately 60 million to 200 million, roughly tripling mainnet capacity, with throughput targeting around 10,000 transactions per second. ETH transfers become up to 71% cheaper. Liquid staking token withdrawal queues process four times faster. Smart contract size limits increase.

The threshold that matters for on-chain economics: strategies requiring under roughly $15 per transaction become viable again on mainnet versus $50-plus today. Active LP rebalancing, tight-margin looping and frequent position management all move back into range.

EIP-8061 carries a trade-off worth naming. Because stake moves in and out faster, the weak subjectivity period — how long a node can remain offline before requiring a recent trusted checkpoint to safely rejoin — roughly halves from about 15.7 days to about 7 days.

Glamsterdam is a portmanteau of Gloas, the consensus-layer component, and Amsterdam, the execution layer. It succeeds Fusaka, which activated in December 2025.

The timeline is the risk. Mainnet activation targets Q4 2026 per the Foundation's roadmap, with the feature set locked and testing underway. Ethereum upgrades run late — early 2027 remains realistic. And the fork itself does not raise the gas limit. Validators vote it up gradually after activation, meaning fee relief depends on behavior rather than code.

Hegotá follows with native privacy and enforced transaction inclusion lists, targeted anywhere from late 2026 to 2027.

ETF Flows Turned Positive but the Streak Is Short

U.S. spot Ethereum ETFs pulled in $71.468 million on August 18, driven largely by BlackRock's ETHA. That is a constructive print in a year where the flow record has been considerably worse than Bitcoin's.

The comparison matters. Ethereum endured a 17-day spot ETF outflow streak totaling roughly $708 million that only ended on June 9 — a stretch with no equivalent in the Bitcoin complex. On August 10, ETH funds recorded $14.6 million in net outflows, a fourth consecutive withdrawal day, while Bitcoin and Solana ETFs took in capital on the same session. Early August saw $12.3 million of single-day outflows and $30.4 million across seven days.

The counterweight is that when flows turn, they turn hard. Ethereum recently outpaced Bitcoin with $365 million in ETF inflows over a comparable window. A single session earlier this month produced $134 million. Across 2025 the category absorbed $9.8 billion in net creations, which materially changed Ethereum's ownership structure.

ETHA alone holds roughly $11 billion in ETH. BlackRock's ETHA saw $99.2 million in net inflows across five days ending July 24.

The structural development is staking. Grayscale's Ethereum Staking ETF launched on NYSE Arca on April 6, 2026, the first product to formalize staked asset redemption mechanics at scale. BlackRock's staked ETH product pulled $100 million on its first day of trading. Anticipation drove a 19-day inflow streak into the category.

Fidelity has proposed adding staking and quarterly payouts to its nearly $900 million ether fund, which would convert a passive wrapper into a yield instrument. That is the institutional story Ethereum lacked: a regulated, exchange-traded vehicle distributing 3% to 4% on-chain yield.

The open question is whether staking products draw net new capital or cannibalize existing ETH ETF demand. Nobody has answered it. The flow data is too short and too two-sided.

Watch the daily net figure. It is the clearest real-time gauge of institutional conviction against retail fear.

A Third of All ETH Is Staked and the Exit Queue Hit Zero

The supply picture is the strongest part of the bull case, and the numbers have moved to extremes.

Approximately 33.33% of ETH supply was staked as of late July — an all-time high — with more recent readings at 34% and roughly 40 million ETH locked. Over one million validators secure the network against more than $70 billion in collateral. Staking yields run 3% to 4% annually.

Exchange reserves hit a record low of 14.5 million ETH in June, with a separate reading at 14.9 million. That is the lowest since 2016 on some measures. Supply available for immediate sale has been shrinking continuously while ETF wrappers absorbed 2025's $9.8 billion.

The validator exit queue fell to zero for the first time since September 2025. Unstaking requests now process immediately, eliminating a backlog that once held more than 2.6 million ETH. That backlog was a persistent overhang — a visible queue of coins waiting to become sellable. It is gone, which reflects validator confidence and removes a source of forward selling pressure.

Stablecoins on Ethereum crossed $158 billion, keeping the network the dominant settlement layer for tokenized dollars regardless of where execution happens.

The bear reading of the same data: a high concentration of staked ETH among large liquid staking providers creates exit-queue pressure if one needs to unwind, and centralization risk grows as their share does. Six researchers including Justin Drake have proposed Tapered Issuance Burn, under which validator rewards decrease as the staking ratio rises, reaching zero issuance reward around a 50% staking ratio, phased over eighteen months. Nothing has been decided — it was presented as a candidate for a future upgrade.

That proposal would reduce ETH inflation further while cutting the yield that staking ETFs are being marketed on. Both effects are real and they point opposite directions.

What the Prediction Markets Actually Price

Positioning data shows a market that has priced out the upside entirely while assigning meaningful probability to the downside.

Reaching $2,000 at any point in August prices at 46%. Above $3,500 in 2026 prices at 19%. Below $1,500 in 2026 prices at 54% — the highest-probability directional outcome on the board. Below $1,250 in August prices at 2%.

Near-term contracts are tighter. ETH at $1,810 or above at 5:00 p.m. ET today prices at 99% across roughly 25,400 contracts. The $1,890 to $1,929.99 range at the same expiry prices at 55%, which puts the current print squarely inside the modal outcome.

Longer-dated Polymarket data assigns an 84.5% probability of reaching $2,000 by the end of 2026 and a 68.5% probability that $1,800 holds as support.

Read together: the market expects the range to hold through today, expects $2,000 eventually, and assigns better-than-even odds to a sub-$1,500 print at some point this year. Those positions are not contradictory — they describe an asset expected to chop violently within a wide band rather than trend.

Sell-side dispersion is the widest of any major asset. Citi carries $3,175 as a base case against $1,198 in a bear scenario. Standard Chartered holds $7,500 long-term with a $4,000 year-end figure after revising more times than any other major institution. Fundstrat runs $4,500 internally. Arthur Hayes views $4,000-plus as achievable. J.P. Morgan is bearish without publishing a target. VanEck declined to forecast for 2026 entirely. Galaxy Digital released no dedicated ETH forecast.

Several firms revised targets by more than 60% within months. When the spread between published estimates runs from $1,198 to $7,500 on the same asset in the same year, the honest conclusion is that nobody has a model that works.

Scenario weightings from one framework: 45% base case at $2,500 to $4,500 if Glamsterdam delivers and ETF outflows moderate, 20% bull case at $4,500 to $12,000.

The Aave Concentration Nobody Is Pricing

The tail risk sitting underneath ETH is in DeFi leverage, and the numbers are specific enough to model.

Roughly 9% of Aave positions carry approximately half the protocol's total debt, and that concentrated cohort runs near 90% loan-to-value against Ethereum-linked collateral.

Ninety percent LTV means those positions liquidate on a move of roughly 10% in the collateral. ETH at $1,920 puts the danger zone near $1,730 — below the $1,810.85 structural floor but well within the range prediction markets assign a 54% probability of reaching this year.

Concentration is what turns a normal drawdown into a cascade. When half of a protocol's debt sits in a small cohort at maximum leverage, liquidations do not arrive gradually. They arrive together, they hit the same collateral asset, and the resulting spot selling drives further liquidations. That mechanism has produced every major DeFi unwind since 2020.

The mitigating factor is that liquid staking has absorbed a large share of the collateral base, and Glamsterdam's four-times-faster LST withdrawal queues would improve the speed at which that collateral can be unwound in an orderly fashion. That improvement arrives in Q4 at the earliest.

The broader leverage picture across crypto is calmer than it has been. Bitcoin's 30-day realized volatility collapsed to 27.2% against a long-run average near 80%, and Ethereum's compression mirrors it — a 0.9% daily range on some recent sessions with average 30-day volatility around 2.39%.

That calm is the setup, not the outcome. Compressed volatility with concentrated leverage beneath it is precisely the configuration that produces outsized moves when a macro catalyst forces price outside the band.

The macro catalyst is scheduled for 2:00 p.m. ET.

The Fed Minutes Are the Only Live Variable Today

The July 28–29 FOMC minutes publish at 2:00 p.m. ET, and Ethereum's correlation to real rates has been the dominant price input all year.

The committee held the federal funds rate at 3.50%–3.75% on a 9–3 vote — the fifth consecutive meeting without a change, with three regional presidents dissenting in favor of a 25 basis point hike. That was the first time since September 2016 that three policymakers aligned on a single directional dissent. Markets currently assign roughly 33% probability to a September hike, down from about 44% a week earlier.

Chair Kevin Warsh has withdrawn forward guidance entirely, which turns the minutes into the primary window on committee thinking. If the document reveals hawkish sentiment extending well beyond the three named dissenters, September odds reprice toward 45%, the 30-year returns to 5.338%, and ETH loses the $1,850 shelf.

The path that lifted ETH into the $1,920s was the reverse: July CPI easing to 3.4% headline with core at 2.5%, PPI flat against expectations for a 0.2% rise, retail sales contracting 0.6%, and payrolls showing unexpected job losses. Those four prints pulled the dollar to 99.00 and the 10-year to 4.70%.

Ethereum has been described as a leveraged bet on the exact conditions that turned hostile in 2026: rising rates, risk-off macro, and a shift in institutional preference toward Bitcoin. That framing explains the 61% drawdown better than any on-chain metric does.

The structural disadvantage against Bitcoin remains unaddressed. Bitcoin has Strategy and its peers holding more than 840,000 BTC as price-insensitive balance-sheet buyers. Ethereum has no equivalent corporate treasury floor providing a bid under selloffs. SharpLink Gaming's planned $200 million ETH deployment on Linea, partnering with ether.fi, EigenCloud and Anchorage Digital Bank, is the closest analogue and is roughly 1/250th the scale.

Beyond today: July PCE on August 26, Jackson Hole from August 27 to 29 with Warsh's first keynote as chair on August 28, and the September 15–16 FOMC carrying a fresh dot plot.

The Amazon-in-2001 Argument and Its Limits

The strongest bull framing compares Ethereum to Amazon in 2001 — stock crashed, business metrics improved. The comparison holds on the fundamentals and breaks on the mechanism.

The fundamentals genuinely improved. Exchange reserves at their lowest since 2016. A third of supply staked at an all-time high. Stablecoins above $158 billion. The largest DeFi TVL by a wide margin. More than one million validators securing over $70 billion in collateral. The largest developer ecosystem in the sector. Every one of those metrics moved the right direction while price fell 61%.

The break is that Amazon's business metrics translated into Amazon's revenue. Ethereum's network metrics increasingly translate into Layer 2 revenue. Base, Arbitrum and Optimism captured the users. The base asset captured the security budget and little else.

That is the specific thing Glamsterdam is engineered to reverse, which is why the Q4 timeline carries more weight than any other item on the calendar. Tripling block capacity and cutting transfer costs 71% narrows the cost gap with rollups enough that mainnet becomes competitive for activity that migrated away. If validators then vote the gas limit upward, fee revenue returns to the base layer and the burn reactivates.

If they do not, or if the upgrade slips into 2027, the structural decline thesis holds and Citi's $1,198 bear case becomes the reference rather than the tail.

Ecosystem governance adds a second variable. The Ethereum Foundation's restructuring and layoffs left a development funding gap, and five former EF researchers launched EthLabs with major ecosystem backing to fill part of it. Whether that transition preserves research momentum is unresolved and unquantifiable.

Competitors are claiming transaction share. Solana at $44 billion market capitalization outperformed ETH on the day and has done so repeatedly through 2026. Developer attention has followed.

The 2030 range across published forecasts spans $360 in VanEck's bear scenario to $12,000 at the high end. That distribution is the entire argument in one line.

The Forecast: Levels, Triggers and the Verdict

The base case is that $1,850 holds and ETH resolves the 100-day EMA at $1,924 on the minutes rather than before them.

The bull sequence requires three steps. Hold above $1,900, take the 100-day EMA at $1,924, then run the swing high at $1,955. That ladder covers 1.8% from spot — a single session executes it on a dovish surprise. Above $1,955 the structure opens toward $2,009, the level that must close over to confirm continuation, then $2,055.47 and $2,133. The 200-day SMA is projected to reach $1,992.55 by September 11, which places a rising ceiling directly beneath $2,009 and makes that zone the defining battle of the next month.

The bear sequence is shorter. Losing $1,900 puts the 20-day EMA at $1,868 in play, then the $1,850 to $1,850.21 confluence of the 50-day EMA and the 23.6% Fibonacci retracement. Losing $1,850 opens $1,842 and then $1,810.85, with the $1,800 psychological level beneath. Below $1,800 the Aave liquidation cluster near $1,730 becomes the operative risk rather than a theoretical one, and prediction markets assign 54% probability to a sub-$1,500 print at some point in 2026.

The trigger set is defined. Bullish: minutes that isolate the three dissenters, sustained ETF creations building on the $71.468 million August 18 print, and a clean Plataberget fork tomorrow that keeps Glamsterdam on the Q4 schedule. Bearish: hawkish minutes repricing September toward 45%, a resumption of the outflow pattern that produced the $708 million 17-day streak, or any signal that Glamsterdam slips into 2027.

The verdict: Ethereum is the only major asset where fundamentals and price have argued opposite cases for a full year, and the reason is arithmetic rather than sentiment. Fees fell 99%, the burn died, and the base asset stopped capturing value from a network that keeps growing. Exchange reserves at 14.5 million, 34% of supply staked, an exit queue at zero and $158 billion of stablecoins are real — and none of them accrue to ETH while rollups take the revenue.

Base case targets $1,955 on a supportive minutes release, with $2,009 as the extension. Failure at $1,850 targets $1,810.85 first and $1,730 as the leverage-driven floor. The structural re-rating does not happen before Q4 and it does not happen at all unless validators vote the gas limit up after Glamsterdam activates. Everything until then is a rate trade on a $233 billion asset priced 61% below its high.

That's TradingNEWS