Timechain Wiki

Network capacity and fee-market critiques

14 min read

Bitcoin's base layer sustains roughly 7 transactions per second, with the post-SegWit 4 MB weight limit serving as a consensus parameter the Block Size Wars (2015-2017) demonstrated cannot be raised by hard fork. The critique: this throughput is inadequate for global money, and congestion-driven fee spikes (2017 bull market; 2023-2024 Ordinals era) contradict the whitepaper's peer-to-peer electronic cash framing. The principled response treats the constraint as a deliberate trade-off preserving full-node operability and propagation security, with the layered-scaling thesis — Lightning for routine payments, base layer for settlement — as the engineered answer. Whether layered scaling can serve mainstream-money use cases remains genuinely uncertain and depends substantially on Lightning's continued development trajectory.

Why this note matters

Paired controversy note (cross-section): the event-level BIP-110 filtering proposal and the broader Ordinals/Inscriptions/Runes dispute live in The Ordinals, Inscriptions, and BIP-110 controversy. This note engages the analytical throughput-and-fee-market question.

The “Bitcoin doesn’t scale” critique is among the oldest and most-engaged. It produced the 2015-2017 Block Size Wars and continues to shape network-evolution debates. It establishes the throughput numbers and structural constraints that produce them, engages the congestion-event empirical record at high resolution, and articulates layered scaling as a principled response rather than retreat. It also distinguishes the technical throughput question from the adoption-friction question treated in User experience and accessibility friction and Lightning Network operational critiques.

The defensible position: the base-layer constraint is real, principled, and unlikely to change. Whether layered scaling can serve global-money use cases is uncertain; evidence will accumulate across 2025-2030.

The critique

Bitcoin’s transaction-throughput limit:

  • Block size: 1 MB (pre-SegWit) → effectively 4 MB weight units (post-SegWit, 2017) → practically 1.5-2.5 MB of useful transaction data depending on transaction-type mix
  • Block interval: ~10 minutes on average
  • Throughput: approximately 7 TPS sustained, 10-12 TPS at peak with favourable transaction mix
  • Comparison: Visa processes ~24,000 TPS; SWIFT averages much higher; PayPal averages thousands

The structural constraint is a consensus parameter, not a technical limit:

  • Raising the block size requires a hard fork that all full nodes must adopt
  • The Block Size Wars (2015-2017) demonstrated that hard-fork attempts at block-size increase are politically infeasible — the larger blocks splits (Bitcoin Cash, BSV) lost most of their support and market value
  • SegWit (soft fork, August 2017) effectively raised capacity to ~1.7x base; further base-layer increases are not on the development roadmap

The critique:

  • At ~7 TPS, Bitcoin cannot serve as global money for ordinary payments. Even modest adoption growth produces persistent congestion.
  • Congestion events produce high fees: during the 2017 bull market, average fees reached 50-200+ per transaction.
  • High fees make ordinary use cases economically impossible: a 50 fee makes no economic sense.
  • The fee-market is structurally insufficient for the original “peer-to-peer electronic cash” framing in the Bitcoin whitepaper.
  • Lightning Network’s promised solution has been slow to deliver in mainstream usage (see Lightning Network operational critiques).

Critics argue Bitcoin has effectively failed at the original “currency” use case and has been redefined as “store of value” to accommodate the throughput constraint — a retroactive narrative shift.

Key proponents

The critique is advanced by diverse voices:

  • Roger Ver and Bitcoin Cash supporters — argued for larger blocks as the principal solution; lost the Block Size Wars but have continued to argue the position
  • Craig Wright and BSV — advocated for “Bitcoin Original” larger-block scaling; mostly marginalized
  • Various academic economists — Eswar Prasad, others — cite throughput limits as part of “Bitcoin isn’t money” critiques
  • Mainstream economic critics — Krugman, Roubini reference scaling limitations
  • Within-Bitcoin technical thinkers — Mike Hearn (left in 2016 over scaling debates); various developers who supported XT/Classic/Unlimited proposals
  • Critics including Frances Coppola, David Gerard, Molly White — cite scaling as evidence of structural issues

The critique evolved meaningfully:

  • 2015-2017: primarily a block-size debate; resolved (controversially) in favour of small-block + Lightning
  • 2017-2023: primarily a Lightning-adoption-pace debate
  • 2023-2026: primarily an Ordinals-congestion and fee-market-equilibrium debate

What’s right about the critique

Several factual points are correct:

Bitcoin’s base-layer throughput is structurally low. The ~7 TPS limit is empirically correct and is not realistically expandable at the base layer.

Congestion events produce real friction. During the 2017 bull market, fees averaged 50+; during 2023-2024 Ordinals activity, fees spiked into similar ranges. These are not theoretical; users experienced them.

Layered scaling has been slower to deliver than originally hoped. Lightning Network has grown but mainstream-payment adoption remains modest as of 2026. The “any day now” framing for Lightning mainstream adoption has been repeated since 2018-2019.

On-chain payments for ordinary purchases are no longer economically viable at typical fee levels. A median transaction fee of 100+ purchases but not for 50 grocery purchases.

The narrative shift is real. The Bitcoin whitepaper’s title is “A Peer-to-Peer Electronic Cash System”; the contemporary Bitcoin discourse emphasizes “store of value” much more than “medium of exchange.” This shift is partly a response to scaling limitations.

The Bitcoin-side response

The throughput constraint is a deliberate trade-off

Bitcoin’s block-size limit preserves several properties that critics underweight:

  • Full-node operability on commodity hardware: smaller blocks mean ordinary users can run full nodes on consumer hardware (~$200-500 setup), which is essential for credible decentralization. Larger blocks would push validation cost higher, eventually requiring data-center-grade hardware that only institutional actors could afford.
  • Block-propagation speed: small blocks propagate to the global network quickly, minimizing orphan rates and supporting the proof-of-work security model.
  • Initial block download (IBD) cost: small blocks keep the historical-blockchain size manageable. As of 2026, the Bitcoin blockchain is ~750 GB; larger blocks would multiply this.

The defensible Bitcoin-side position: throughput could be expanded but only at the cost of decentralization (full-node operability) and security (block-propagation speed; long-term IBD cost). The trade-off has been deliberate.

The layered-scaling thesis

Bitcoin’s scaling design intentionally moves payment activity to upper layers:

  • Layer 1 (base chain): settlement layer for high-value transactions; channel-management for Layer 2; institutional settlement
  • Layer 2 (Lightning Network): routine payments; micropayments; cross-border remittances
  • Layer 3 and higher: more specialized payment systems built on Lightning (Fedimint, Cashu, custodial channels)
  • Custodial services: hot wallets and payment apps that bundle multiple users’ activities into fewer on-chain transactions

The argument: the throughput-per-user can be much higher than 7 TPS because each user doesn’t need their own on-chain transaction per payment.

The honest counter: layered scaling shifts the throughput problem upward but doesn’t solve it. If Lightning doesn’t deliver on its promise (per Lightning Network operational critiques), the layered architecture doesn’t help.

The fee market is the rationing mechanism

Bitcoin’s fee market prices block space efficiently:

  • High-value transactions can pay high fees; they get included quickly
  • Low-value transactions can wait for low-fee periods; they get included at low cost when demand drops
  • The market prices block space at the marginal value of inclusion

This is economic-textbook market behaviour, not a failure mode. The criticism that “fees are too high during congestion” is essentially “the price-signaling-mechanism produces price signals” — which is what it’s supposed to do.

The genuine concern is whether the equilibrium fee level is high enough to exclude ordinary users from the on-chain layer. As of 2026, base-layer fees during non-congested periods are typically 5 coffee.

Layered scaling distinguishes use-case appropriateness

A key reframing: not every payment should be on-chain. The right architecture pairs use cases with appropriate layers:

  • High-value, settlement-grade transactions: on-chain ($1M+ transfers; institutional settlement; channel-management transactions)
  • Routine payments: Lightning ($1-1000 transfers; coffee, retail, micropayments)
  • Sub-cent and high-frequency payments: Lightning-or-higher with bundling (streaming payments; metered services)
  • Identity-and-credential transactions: emerging Layer 2 protocols (BitVM, rollups proposals)

The critique sometimes treats Bitcoin as if all payments should happen on-chain. This is not the right framing for any modern payment system; SWIFT and ACH have similar low-throughput characteristics because they serve high-value settlement, not retail commerce.

Ordinals demonstrate the fee market works

The 2023-2024 Ordinals/Inscriptions era is sometimes cited as a critique (high fees forced ordinary users out). It is also a demonstration that the fee market works:

  • New use cases (Ordinals; BRC-20; Inscriptions) emerged organically
  • Block space became valuable; fees rose
  • Users with higher-value uses outbid users with lower-value uses
  • Lightning and adjacent infrastructure handled the displaced routine-payment activity
  • The fee market reached an equilibrium that rewarded miners (good for Long-term security budget) and priced block space at its marginal value

This is the system working as designed. The criticism that “high fees during Ordinals were bad” is implicitly arguing that the fee market shouldn’t price block space — which would be a problem in its own right.

Counter-arguments and tensions

”Lightning hasn’t delivered; the layered-scaling thesis is theoretical”

The tension: Lightning’s mainstream adoption has been slower than promised. As of 2026, Lightning capacity is ~5,000 BTC; usage is concentrated in specific applications (Strike, Cash App, El Salvador’s Chivo); it has not become the universal-payments layer for Bitcoin. The layered-scaling argument depends on Lightning succeeding; if it doesn’t, the throughput-constraint problem remains.

Response: Real concern; see Lightning Network operational critiques for the detailed treatment. Mitigations: (1) Lightning is growing meaningfully in absolute terms even if mainstream-adoption metrics lag; (2) adjacent Layer 2 technologies (Fedimint, Cashu, BitVM) are emerging and may fill gaps; (3) custodial-payment-app integrations (Strike, Cash App) handle a meaningful fraction of routine Bitcoin payments. But the success of layered scaling is genuinely contingent on Lightning’s continued development.

”The block-size debate isn’t settled; it just hasn’t been re-litigated”

The tension: The Block Size Wars ended with a soft-fork solution (SegWit) and the split-off of Bitcoin Cash. But the underlying disagreement about base-layer scaling vs Layer-2 scaling is not resolved; it has been deferred. A future protocol-evolution debate could re-open it.

Response: Real concern. The empirical observation: Bitcoin Cash’s market value has collapsed (~99% lower than Bitcoin); BSV has similarly failed; the larger-block alternatives have been market-rejected. This is the strongest evidence that the current resolution is community-accepted. But the debate could re-open if Lightning fails to deliver or if specific block-size proposals address current technical concerns more carefully than the 2015-2017 proposals did.

”Most users don’t need on-chain transactions; the ‘global money’ framing was wrong”

The tension: The Bitcoin community has, in some quarters, retreated from the “global money” framing in favour of “store of value.” This is partly a response to scaling limitations. The critique: Bitcoin originally promised broad-adoption payment functionality; the redefinition is moving goalposts.

Response: Partially valid. The framing has shifted; some Bitcoin advocates emphasize store-of-value more than medium-of-exchange. But: (1) the layered-scaling architecture does enable mass-adoption payments via Lightning; (2) the store-of-value role is itself a substantial economic function that justifies Bitcoin’s existence; (3) the original “global money” framing implicitly assumed mass adoption with no scaling constraints — that framing has been refined, not abandoned. The honest position: Bitcoin is currently better-suited to store-of-value than medium-of-exchange, with Layer 2 enabling growing medium-of-exchange capability.

”Fee-market congestion experiences are bad for mainstream adoption”

The tension: From a mainstream-user perspective, occasionally paying $50 for a transaction is unacceptable. The fact that the fee-market produces this outcome (when block space is scarce) makes Bitcoin’s mainstream adoption harder.

Response: True for on-chain mainstream payments. Mitigations: (1) Lightning-based payment apps handle most routine mainstream payments without on-chain congestion exposure; (2) custodial wallets bundle multiple users into fewer on-chain transactions; (3) high-value users (institutions, large transfers) can pay congestion fees comfortably. The honest answer: Bitcoin’s on-chain layer is not for mainstream payments; the Layer 2 layer is. Whether Layer 2 is sufficient is the real question.

”Covenant proposals could improve throughput without hard forks”

The tension: Proposals like CTV (BIP-119), OP_CAT re-enablement, and various other covenant proposals could enable new scaling mechanisms (vault constructions; rollups; channel improvements) via soft forks. Bitcoin’s resistance to even soft-fork upgrades (see Protocol-evolution constraints) may be blocking real improvements.

Response: Real opportunity; the protocol-evolution-debate is genuine. Various covenant proposals could enable meaningful scaling improvements via soft fork rather than the politically-impossible hard-fork block-size increase. The debate over which covenants to adopt is active in 2026 and is the most likely path to meaningful base-layer evolution.

Verdict: Throughput constraint is real and principled; layered-scaling response is the right architecture; success is contingent on Lightning trajectory

The base-layer throughput limit is real and is structurally unlikely to change. The layered-scaling response is the principled architecture. Whether it will be sufficient for mainstream-money use cases is the genuinely open question.

A serious assessment:

  • Base-layer throughput: ~7 TPS; structurally constrained by decentralization-preservation trade-offs; unlikely to change materially
  • Congestion-event UX: real friction during demand spikes; mitigated by Lightning and custodial services for ordinary users
  • Layered-scaling: technically working; mainstream-adoption pace has been slower than promised; success is contingent on Lightning trajectory (per Lightning Network operational critiques)
  • Covenant proposals: could enable meaningful improvements via soft fork; political feasibility is the limiting factor (per Protocol-evolution constraints)
  • Ordinals-style novel uses: demonstrate fee-market function; provide Long-term security budget support; trade-off against ordinary-user payment-cost during congestion

This critique is worth taking seriously as a structural limitation that shapes Bitcoin’s use cases. It is not a critique that justifies dismissing Bitcoin; it is a critique that justifies recognizing Bitcoin’s specific architectural choices.

Open questions for further development

  • What is the realistic timeline for Lightning to handle mass-adoption payments? 2-5 years? 5-10 years? Never?
  • Which covenant proposals (BIP-119 CTV, OP_CAT, others) are most likely to be activated, and what scaling improvements would they enable?
  • The interaction between Long-term security budget and scaling is concerning — high fees support miner revenue but exclude ordinary users; low fees support ordinary use but underfund security. What’s the equilibrium?
  • Mainstream-payment alternatives (stablecoins on Tron/Solana; CBDCs; Lightning-via-Cash-App) are growing; how do they affect Bitcoin’s medium-of-exchange role?
  • The 2017 fee-market and 2023-2024 Ordinals fee-market produced different patterns. What does the equilibrium fee-market look like at sustained mass-adoption?

Canonical sources for this note

Foundational technical:

  • Nakamoto, Satoshi — Bitcoin Whitepaper (2008) — the original “peer-to-peer electronic cash” framing; see The Bitcoin whitepaper - Explainer
  • BIP-141 (SegWit) — the 2017 soft-fork that effectively increased capacity
  • Various BIPs and proposals on scaling

The Block Size Wars:

Engineering and academic:

Within-Bitcoin engagements:

  • Carter, Nic — various essays on Bitcoin’s scaling architecture
  • Adam Back, Pieter Wuille, Greg Maxwell — technical talks on Layer-2 scaling philosophy
  • Lopp, Jameson — practitioner perspective; see Jameson Lopp

Critic engagement:

As of 2026-05-15: Bitcoin base-layer throughput remains ~7 TPS; congestion fees vary widely; Lightning adoption growing but partial; covenant proposals active in development.

Within the Criticisms section:

Scaling and Layer 2 section (cross-listed):

Technical foundations section:

History-section adjacency:

Adjacent thinker pages:

The sub-MOC home:

Graph View

Backlinks