Investment Thesis
The December 2025 Fusaka upgrade has broken ETH tokenomics by expanding the gas limit in a way that catastrophically underestimated fee elasticity β causing a ~90% fee collapse that has flooded the network with address poisoning fraud rather than institutional adoption.
Culper's proprietary chain-wide SQL analysis of every ETH transaction from January 2025 through February 2026 reveals that 95% of new wallet growth and over 50% of transaction growth is attributable to address poisoning attackers, not genuine users. Bulls like Tom Lee cite these "parabolic" activity metrics as evidence of strengthening fundamentals; they are in reality evidence of systemic on-chain fraud made economically viable by Fusaka's blockspace glut β which reduced the cost of a poisoning attack from $2.10 to $0.03. The same fee collapse has impaired validator economics, driven staking yields below treasuries, tipped Ethereum into inflation above 0.8%, and created a reflexive loop that Ethereum's own creator appears to be betting against: Vitalik sold 19,326 ETH in February 2026 β 16% more than he pre-announced.
Core Allegations
Protocol Failure
Fusaka Wildly Miscalculated Fee Elasticity
Vitalik and the Pump the Gas initiative estimated Fusaka's gas limit increase from 45 to 60 million would reduce fees 10β30%. In reality, gas prices collapsed ~90%. Culper argues elasticity estimates were based on pre-EIP-1559 data from 2018 and never updated to reflect today's L2-dominant ecosystem, where over 85% of activity occurs off mainnet.
On-Chain Fraud
Address Poisoning Explosion β It's Not Institutions, It's Fraud
Culper's SQL heuristic analysis across all ETH and major stablecoin transactions found that 95% of new wallet growth since Fusaka is from poison/dust senders. In February 2026, 22% of all Ethereum mainnet transactions were address poisoning attacks β up from under 10% pre-Fusaka. Annualized losses are estimated at $348M+, versus $83.8M over two full prior years per CMU researchers.
Validator Economics
Validator Fees Collapse; Staking Yields Fall Below Treasuries
Post-Fusaka, blocks are filled with negligible-value spam. Legitimate transactions no longer need to competitively tip validators, causing median tips per gas to fall precipitously. Staking yields now sit below U.S. treasuries while ETH has collapsed in price, undermining the core premise that staking offers strong yields in a deflationary asset with no real alternatives.
Reflexive Risk
ETH Now Inflationary β Reflexive Tokenomics Deterioration
Lower priority fees reduce ETH burn, tipping the protocol from deflationary to inflationary (>0.8%). This leads to lower staking demand, which reduces network security, which further erodes institutional confidence β a reflexive loop experts warned could emerge if blockspace expanded faster than underlying demand. Glamsterdam's planned 200M gas target warrants scrutiny as likely accelerating this cycle.
Misleading Narrative
ETH Bulls Fundamentally Misread the Activity Metrics
Tom Lee (Fundstrat CEO, BMNR Chairman) cited active addresses up 117% year-over-year and 80% network usage growth in six months as proof ETH is not in a death spiral. Lee explicitly said "utility is going up." Culper's analysis shows this activity is driven almost entirely by poisoning attackers β not institutional adoption from UBS, Standard Chartered, or Fidelity as Lee suggests.
Competitive Displacement
Solana and L2s Continuing to Erode ETH Mainnet Share
As of July 2025, over 85% of Ethereum transactions occur on L2s. Solana overtook Ethereum in daily transactions in July 2024. In 2025, Solana grew developer count 29% versus 6% for Ethereum. Stripe, Visa, Citigroup, Deutsche Bank, and others selected Solana for key initiatives. Culper draws the parallel to Netscape and Nokia β platforms that dominated but failed to capture value when successors arrived.
Key Data
Cost Per Poisoning Attack β Before & After Fusaka
Source: Culper Estimates. Assumes 28,000 gas per attack per Guan et al. methodology.
Key implication: When Fusaka crossed the 60M gas threshold, the cost to attack an ETH wallet fell 70Γ compared to mid-2025 β making ETH more attractive to poison than BSC, which previously saw 252M poisoning attempts vs. just 17M for ETH over the same period per CMU data.
Address Poisoning as % of Total Ethereum Transactions
Source: Culper on-chain analysis. ETH + major stablecoins (USDT, USDC, DAI). Jan 2025 β Feb 2026.
Key Chronology β The Fusaka Failure Timeline
2013 β 2015
Ethereum Proposed, Crowdfunded & Launched
Vitalik Buterin proposed Ethereum in 2013; it was crowdfunded in 2014 and launched in 2015. The ETH token is required for all network activity β users pay gas fees in ETH, and validators stake ETH to secure the network.
2021
EIP-1559 Introduces Base Fee Model; L2s Emerge
Ethereum moved to a base fee and priority tip model with EIP-1559. Major L2s including Optimism and Arbitrum emerged, eventually siphoning the overwhelming majority of activity from the Ethereum mainnet. Gas limit held at 30M from 2021 through early 2025.
Feb β Jul 2025
Incremental Gas Limit Increases to 45 Million
Gas limit raised from 30M to 36M in February 2025, then to 45M in July 2025. By July 2025, over 85% of Ethereum transactions occurred on L2s. Solana overtook Ethereum in daily transactions for the first time. Active validator count peaked. Cost per poisoning attempt: $0.11.
December 3, 2025
Fusaka Upgrade β Gas Limit Expanded to 60 Million
The Fusaka upgrade raised the gas limit to 60 million, explicitly to scale the L1. Pump the Gas (PTG) β praised by the Ethereum Foundation for "helping mobilize the effort" β estimated fees would decline 10β30%. In reality, gas prices collapsed ~90%, crossing a cost threshold that made ETH wallets far more attractive to poison than BSC.
January 30, 2026
Vitalik Announces "Mild Austerity," Pre-Discloses 16,384 ETH Sale
Vitalik Buterin announced the Ethereum Foundation would be "entering a period of mild austerity" and pre-disclosed a personal sale of 16,384 ETH. CMU researchers had by this point already detected $50.5M in wallet poisoning losses from December 3, 2025 to January 28, 2026 alone.
February 2026
Vitalik Sells 19,326 ETH β 16% More Than Pre-Announced
On-chain tracking revealed Vitalik sold 19,326 ETH in February 2026 β 16% more than pre-announced. Crypto influencer Sillytuna lost $24M to a poisoning scam, bringing reported post-Fusaka losses to at least $87M in three months β an order of magnitude above prior CMU benchmark levels.
March 5, 2026
Culper Publishes Short Report; Glamsterdam Upgrade Looms
Culper publishes its chain-wide analysis. Ethereum has targeted a gas limit of 100M+ in 2026, with Glamsterdam targeting 200M. Culper believes further expansions will accelerate ETH's decline, push yields even lower, and increase ETH inflation. The staking entry queue stands at 3.35 million ETH, driven primarily by ETH treasury companies like BMNR.