Ethereum post-Pectra supply dynamics: myths versus on-chain facts

The Pectra upgrade went live on May 7, 2025, without the price spikes or narrative frenzy that usually accompany major protocol changes. Instead of a speculative rally, the network experienced a quiet technical shift that fundamentally altered how Layer 2 rollups interact with the base layer. The most significant change was EIP-7691, which increased the target and maximum blobs per block from 3 to 6 and 9, respectively. This expansion raised daily data capacity to about 8.15GB, up from the previous 5.5GB limit. While this sounds like a win for scalability, it immediately broke the "ultrasound money" narrative that many investors had pinned their hopes on. The assumption that higher usage would automatically lead to greater ETH burns proved wrong in this specific context. By increasing the supply of available data space, the upgrade reduced the scarcity of the blob market, causing prices to collapse rather than surge.

The burn rate paradox

I find the post-Pectra burn data particularly damning for the simple deflationary thesis. In the five full days following the upgrade, rollups purchased an average of 25,600 blobs daily, a 20.8% increase over the pre-upgrade baseline of 21,200. Despite this rise in volume, the average number of blobs per block remained 33% below the new target of six. Because demand did not saturate the new capacity, blob object fees dropped to nearly zero. The median cost per blob object fell to $0.00000000035, meaning rollups paid less than one-thousandth of a penny daily for data availability. This represents a near 100% decline in blob object fees compared to the $16,250 daily average seen in the 60 days leading into Pectra. Consequently, the total ETH burned from rollup data posting activity plummeted by 71%, dropping from 11.22 ETH per day to just 3.26 ETH. The only reason daily burn rates didn’t fall even further is that the Layer 1 base fee spiked more than 650% week-over-week, forcing rollups to pay higher execution costs. Without that unexpected L1 fee spike, the deflationary impact of Pectra would have been even more severe.

Exchange balances and staking locks

While the burn mechanism weakened, other supply metrics tell a different story. Ethereum balances on centralized exchanges hit a decade low of 8.7% last Thursday, the lowest level since the network launched in mid-2015. This figure declined 43% since early July, coinciding with accelerated digital asset treasury purchasing activity. By comparison, Bitcoin exchange balances stood at 14.7%, highlighting a distinct difference in holder behavior and custody preferences. The gap suggests that ETH holders are moving coins into destinations that do not typically sell, such as staking protocols, restaking platforms, and long-term custody solutions. As of early 2026, over 36 million ETH, roughly 30% of total supply, is locked in staking contracts. These validators commit coins to secure the network and earn yield, removing them from the tradable pool. With annual inflation hovering around 0.18% in 2025, the circulating supply is effectively shrinking relative to demand. You might assume that lower burn rates would offset these tight exchange balances, but the data shows that withdrawal from exchanges is outpacing the reduction in fee-driven destruction. Does the current level of staked ETH represent a permanent removal from circulation, or is it merely a temporary lock-up waiting for better prices? The answer remains unclear as the network continues to evolve.