I Stockpiled $600 of ETH for Gas — the Merge Turned It Into a Bet I Never Chose

At 10:12 on a Tuesday night in August 2022, I moved $600 into a wallet labeled, with the confidence of a man about to be taught a lesson, "gas reserve." The plan was tidy: Ethereum fees were spiking and swinging, so I would buy six months of transaction costs up front at a price I could tolerate, park the ETH, and spend it down as needed. A prepaid toll card. A hedge against my own impatience during fee spikes. For five months it worked exactly as designed, and I want you to hold that in mind, cuz the ending is not about the plan failing. The plan worked. The ground moved.

The stakes were slight — $600 in a portfolio with more zeros elsewhere — but the lesson scaled far past its size, which is the only reason Im still telling this story four years later.

the logic of prepaid gas

The reasoning was sound, and I will defend it in its own time frame. My on-chain activity was predictable: a few token moves, some DeFi maintenance, occasional mints. Fees in mid-2022 were historically low, and every prior spike had punished me into transacting at the worst hour cuz I needed the transaction now. Prebuying converted an unpredictable expense into a fixed one, like filling a propane tank in September. I even wrote the logic in a note to myself in the wallet's label field, which future-me would read with the exact expression you are imagining.

five good months

From August 2022 through January 2023, the reserve behaved. I spent it down in measured draughts, refilled once, and my effective fee cost was locked while friends complained about gwei spikes. Total saved versus my panicked-transaction baseline: mayb $90. Modest, real, and I cited it in two conversations with the smugness those savings did not quite earn.

what the Merge actually did

Then, in September 2022, Ethereum switched its entire consensus mechanism, and over the following months the network's fee market changed character far more than I had modeled. EIP-1559 had already made base fees adaptive; post-Merge, with blob space arriving later and activity migrating to layer 2s, my main-chain usage simply stopped being the way anyone transacted. The transactions I had been prepaying for moved to rollups costing cents. My reserve was no longer a prepaid toll card for a road I drove. It was just $600 of ETH, held for its own sake, exposed to a price I had never intended to have an opinion about.

the position I never chose

This is the part that recalibrated me. At no point did I decide to increase my ETH exposure by $600. The decision made itself, quietly, inside a cost-control scheme, the moment the premise underneath the scheme expired — and premises expire without sending notice. I was now long an asset as a side effect of budgeting, which is how, I suspect, most accidental positions in the world are actually formed. Nobody chooses to be overweight. Their utility closet fills up and one day it is an asset class.

the unwind, at the wrong hour

I spotted the drift in March 2023 during a portfolio review — the reserve sat 22% above its purchase value while still labeled "gas" — and then did the classic thing: nothing. It felt wasteful to convert a winning position, and the label kept whispering that it had a purpose. Two months later I needed the transaction budget for something real, the price had given back the gain, and I unwound the whole thing in two tranches within 3% of my original cost. Round trip: a $90 saving, a $600 asset held for ten months for no reason, and approximately forty minutes of taxes on conversions that had never needed to exist.

the rule that came out of it

The rule I run now has a boring name — the expiry check. Every prepaid or earmarked position gets a labeled premise and a quarterly question: is the reason this money exists still true? Not "is it up," not "do I feel lucky" — is the premise alive. The gas reserve's premise died in September 2022 and nobody called it, including me, cuz I had written the premise in a wallet label field where no review process would ever look. Premises now live in a spreadsheet with a date column, like an adult.

what I do about fees now

As for gas itself, the modern answer humbles the whole project: I moved nearly all the activity that used to justify the reserve onto layer 2s, where a year of transactions now costs less than the coffee I buy while waiting for one old-mainnet confirmation. The $600 prepaid toll card is a relic of a network I barely touch. My current transaction budget is so slight that prepaid anything would be a joke, which is, I imagine, the market solving my problem while charging me a tuition of time.

back to the label field

I still keep that wallet. The label still reads "gas reserve — six months of fees at Aug 2022 rates," and I have never changed it, because it functions as a museum exhibit now: the most confident prediction I ever wrote down, expired in place. The fee-hedging plan was not stupid; it was correct for about five months, which is five months longer than most plans survive. But the balance it left behind taught the actual lesson — every earmarked pile of money is a bet, whether you admit it or not, and the day the reason goes away is the day you either sell it or confess that you are an investor now. I confessed ten months late. The label field still hasn't.