Bitcoin's Fee Market Collapsed After the Halving. Here's Why That's the Most Important Long-Term Risk Nobody Is Talking About. | Bitcoin Index

The scariest chart in Bitcoin right now isn’t the price chart. It’s the fee market chart.

Transaction fees as a share of miner revenue collapsed from roughly 7% to 1% after the 2024 halving (April 20, 2024), according to The Block’s mining outlook. Median daily fees dropped over 80% since the halving, Galaxy reports. About 15% of blocks are now “free blocks” where the average fee is 1 satoshi per byte or less. Total fees fell 88.63% year-to-date as of August 2025, averaging just $0.87 per transaction.

This is not a short-term market quirk. It’s a structural problem for Bitcoin’s long-term security model.

How Bitcoin’s security model works

Bitcoin miners earn revenue from two sources:

The block subsidy: New bitcoins issued per block (currently 3.125 BTC, post-April 2024 halving)
Transaction fees: What users pay to get their transactions included in a block

The subsidy is Bitcoin’s inflation mechanism. It pays miners to secure the network. But it halves every 210,000 blocks (roughly every four years) and will eventually reach zero around the year 2140.

The halving schedule looks like this:

2012: 50 → 25 BTC
2016: 25 → 12.5 BTC
2020: 12.5 → 6.25 BTC
2024: 6.25 → 3.125 BTC (current)
~2028: 3.125 → 1.5625 BTC
~2032: 1.5625 → 0.78125 BTC
2140: subsidy → 0

When the subsidy hits zero, miners will earn only transaction fees. If those fees aren’t high enough to justify the cost of running mining hardware and consuming electricity, miners shut down. Hashrate drops. The network becomes vulnerable to 51% attacks.

Bitcoin’s entire security model rests on this assumption: transaction fees will grow to replace the subsidy.

The post-2024 halving data is alarming

Here’s what happened after the April 2024 halving:

Fees collapsed from 7% to 1% of miner revenue

The Block’s 2026 Bitcoin Mining Outlook (published December 30, 2025) states:

“Bitcoin miner revenue rose to ~$17.2B on higher bitcoin prices post-halving, but transaction fees as share of revenue collapsed from ~7% to ~1% as the 2024 onchain activity boom faded, increasing miners’ reliance on BTC price appreciation.”

Miners are now 99% dependent on the block subsidy, which will halve again around 2028.

TheMinerMag reported that fees dropped from 6.7% of block rewards in May 2024 (the first full month post-halving) to below 1% by June 2025. One year, and fees evaporated.

Median daily fees dropped over 80%

Galaxy’s “Where Did All the Fees Go?” report (August 20, 2025) found:

Median daily fee dropped more than 80% since April 2024
About 15% of daily blocks are now “free blocks” (average fee ≤1 sat/vB)
Nearly 50% of blocks don’t reach full capacity at times, meaning the mempool is often empty

Free blocks were virtually nonexistent during the Ordinals and Runes boom in 2024. Their rise to 15% of blocks signals a structural shift away from high-fee competition for blockspace.

Total fees down 88.63% year-to-date

AInvest reported in August 2025 that Bitcoin transaction fees plummeted 88.63% year-to-date, averaging $0.8732 per transaction. The peak fee in July 2025 was $2.85, compared to $7.68 a year prior.

For context, during the Runes launch (block #840,000, the halving block itself), fees temporarily spiked to 72% of miner revenue. That lasted a few weeks. Then the market adjusted, Runes hype faded, and fees collapsed back to sub-1%.

Hashprice fell ~35% year-over-year

Hashprice measures miner profitability: revenue per unit of mining power, expressed as $/PH/s/day (or $/TH/s). It combines block subsidy and fees.

As of February 2026, hashprice sits around $34/PH/s/day, down roughly 35% from a year earlier. Miners with electricity costs above $0.07/kWh are operating at a loss. More than 20% of hashrate dropped offline temporarily after the halving as smaller operations shut down.

Why did fees collapse?

The ETF paradox: off-chain demand doesn’t generate fees

Bitcoin’s price can rise without generating proportional fee revenue if demand is primarily through off-chain wrappers.

Spot Bitcoin ETFs now hold roughly 1.3 million BTC. ETF trades settle on traditional stock exchanges (NYSE, NASDAQ). They don’t create Bitcoin transactions. Institutional accumulation via MicroStrategy, corporate treasuries, and ETFs involves large OTC purchases that settle infrequently on-chain.

Compare this to the 2020-2021 cycle. After that halving, the bull run was driven by NFTs, DeFi, and retail speculation on-chain. Users were actively transacting, creating fee demand.

After the 2024 halving, the bull run was driven by ETF inflows. Institutional off-chain demand. Bitcoin’s price rose. Fee revenue collapsed.

Memecoin activity migrated to Solana

Galaxy observed:

“Memecoin activity has users who go for that sort of thing increasingly migrating to faster, cheaper L1s like Solana. It’s simply not worth dealing with the poor UX of trading Runes when the Solana memecoin trading experience has become so clean.”

Speculative trading generates high fee spikes. That activity has moved off Bitcoin. Runes and Ordinals provided a brief surge in Q2-Q3 2024, but the activity wasn’t durable. OP_RETURN transactions (used for Ordinals/Runes) peaked at 40-60% of daily transactions during Q2-Q3 2024, then declined to around 20% by August 2025.

Lightning and Layer 2s haven’t (yet) driven settlement demand

The optimistic case: Lightning Network, Ark, and other Layer 2 solutions will generate fee demand through batched on-chain settlement transactions (channel opens/closes, covenant transactions).

The current reality: Lightning has been live since 2018. Eight years later, fee revenue is still below 1%. A single Lightning channel open can support thousands of off-chain payments before needing to settle on-chain. Fee revenue per transaction is vastly lower than direct on-chain payments.

Lightning adoption is growing, but settlement transactions are infrequent and small.

What would a healthy fee market look like?

Industry analysts (including Bitdeer) suggest fees need to represent 20%+ of miner revenue consistently to signal long-term security sustainability.

Current baseline: ~1%
Required for long-term security: ~20%+
Gap: 20x increase needed

Let’s do the math:

Today (2024 halving):

Block subsidy: 3.125 BTC
Fees: ~0.03 BTC (~1% of subsidy)
Total miner revenue per block: ~3.155 BTC

Healthy fee market target:

Block subsidy: 3.125 BTC
Fees: ~0.78 BTC (20% of total revenue)
Total miner revenue per block: ~3.9 BTC
Needed fee increase: from 0.03 BTC/block to 0.78 BTC/block (26x)

By 2028 (next halving):

Subsidy will drop to 1.5625 BTC
If fees remain at 1%, total revenue drops to ~1.58 BTC/block (a 50% revenue cut)
To maintain today’s security budget (3.155 BTC/block), fees would need to rise to ~1.6 BTC/block
That’s a 53x increase from current levels

What could drive a 20x-50x fee increase?

Layer 2 settlement surge: If Lightning usage explodes, batched channel settlement could create sustained fee demand. Ark, statechains, and other L2s may add additional settlement transactions.
Inscription/token activity normalizes: If Ordinals, Runes, or future Bitcoin-native tokens find durable use cases (not just speculation), they could create structural blockspace demand.
Bitcoin as global settlement layer: If Bitcoin becomes the primary settlement layer for international payments, remittances, or corporate treasury movements, transaction volume could spike. This requires overcoming UX barriers and Lightning maturation.
Congestion becomes the norm: Bitcoin’s fee market is responsive to congestion. When blockspace is scarce, fees rise. The question is whether demand will be sufficient to create sustained congestion.

None of these are guaranteed. And Bitcoin can only process around 7 transactions per second on-chain due to the blocksize cap. For fees to rise 20x, either every transaction needs to pay 20x more, or demand must exceed supply so consistently that bidding wars become the norm.

The miner response: pivot to AI and consolidate

Miners aren’t sitting still. AMINA Bank’s post-halving report (November 19, 2025) notes:

“Profitability pressure led to operational innovation. Miners proactively invested in next-generation ASICs, diversified into AI/HPC (High Power Compute) workloads, and adopted hedging tools to protect cash flows. This transition positioned many miners as infrastructure providers, not just Bitcoin producers.”

Miners pivoting to AI/HPC is a workaround, not a solution to the fee market problem. It proves that Bitcoin mining alone is insufficient to sustain operations. So they’re diversifying revenue streams.

MARA, Riot, and CleanSpark are all exploring AI/HPC partnerships. Mining hardware (ASICs) can be repurposed, or data centers can lease excess capacity to AI workloads.

But here’s the question this raises: If miners derive most revenue from AI/HPC rather than Bitcoin mining, will they prioritize Bitcoin’s security when it’s no longer their primary revenue source?

AMINA also notes industry consolidation is accelerating:

“Smaller miners exited the market due to tighter margins, while larger firms capitalized on M&A to scale operations and secure power access. The top pools—led by Foundry USA and MARA Pool—now account for over 38% of global Bitcoin hashpower.”

Centralization risk: If a handful of large miners control more than 51% of hashrate, Bitcoin’s security model weakens. Geographic concentration (many large miners are U.S.-based) creates regulatory and geopolitical risks.

Long-term security implications

In 2016, researchers at Princeton published “On the Instability of Bitcoin Without the Block Reward,” the foundational academic paper on this risk. Key finding: when the subsidy becomes negligible, Bitcoin’s security relies entirely on fees. If fees are low, miners may engage in selfish mining or fee-sniping attacks to maximize revenue.

The current trajectory:

2028: Security budget cut in half (subsidy drops to 1.5625 BTC, fees still ~1%)
2032: Security budget cut in half again (subsidy drops to 0.78125 BTC)
2036: Security budget approaching dangerous territory (subsidy < 0.4 BTC)

At what point does the total security budget (subsidy + fees) fall below the cost of a 51% attack?

Three scenarios

Scenario 1: Hashrate collapse
Miners shut down unprofitable operations. Network hashrate drops (this already happened post-2024 halving: more than 20% of hashrate went offline temporarily). Lower hashrate makes 51% attacks cheaper and more feasible.

Scenario 2: Fee market maturation (optimistic)
On-chain demand surges (L2 settlement, institutional payments, token activity). Fees rise to 20%+ of revenue, filling the subsidy gap. Security budget remains robust.

Scenario 3: “Paper Bitcoin” dominance (dystopian)
Most Bitcoin is held in ETFs, custodial products, and wrapped tokens. On-chain transactions become rare (only used for large settlements). Fee market never develops. Miners exit. Network security degrades. Bitcoin becomes a “digital gold” asset that’s rarely moved on-chain.

Which scenario is most likely? Too early to tell. But the post-2024 data leans away from Scenario 2 and toward Scenario 3.

Counterarguments

“The security budget is fine for now.”
True. At current hashrate (~~1 zetahash) and BTC price (~~$68k as of February 2026), miners are still profitable overall. The security budget problem is 2028 or 2032, not today.

But waiting until the next halving to act gives zero time to adapt. Fee markets take years to develop. If we wait until 2028 when the subsidy halves again, there’s no runway left.

“Layer 2 activity will drive fees.”
This is the optimistic case. But Lightning has been live since 2018. Eight years later, fee revenue is still below 1%. A single Lightning channel can support thousands of off-chain payments before settling on-chain. Fee revenue per transaction is vastly lower than direct on-chain payments.

“Ordinals and Runes will normalize.”
Possible. But the data shows activity faded rapidly after the initial hype. OP_RETURN transactions dropped from 40-60% of daily transactions (Q2-Q3 2024) to around 20% by August 2025. Speculation-driven demand is episodic, not structural.

“Fee market mechanisms work; they just need demand.”
Correct, but tautological. The question is: Will demand be sufficient? The post-2024 data suggests:

ETFs and custodial products suppress on-chain demand
Memecoin speculation migrated to Solana
Lightning/L2s reduce on-chain settlement frequency

Without structural drivers of on-chain demand, the fee market can’t develop.

The 2028 deadline

Bitcoin’s security model is being stress-tested right now. The next halving (around 2028) is the critical deadline. If fees don’t develop meaningfully in the next two years, Bitcoin faces a structural security crisis.

Galaxy put it plainly:

“The long-term economics of Bitcoin security rest on a future with a robust fee market, and right now the fee market is anything but robust.”

This is not FUD. It’s math. The block subsidy halves every four years. Transaction fees need to replace it. The data from 2024-2026 shows fees moving in the wrong direction.

I’m genuinely uncertain how this resolves. Maybe Layer 2s mature and drive settlement demand. Maybe Bitcoin-native tokens find durable use cases. Maybe institutional on-chain activity picks up as custody solutions improve. Or maybe Bitcoin becomes a primarily off-chain asset, held in ETFs and rarely moved, with a slowly degrading security budget.

The next two years will tell us which path we’re on.