Bitcoin may not be a Ponzi scheme, but does that settle the harder question? A network can be open and honest about its rules and still impose costs that outweigh its benefits. It can help criminals move money. It can consume large amounts of electricity. It can create hardware waste. The strongest objection is not that Bitcoin hides these costs; it is that society should not have to bear them for a monetary system that may be unnecessary.
The previous chapter separated Bitcoin's protocol from fraudulent businesses built around it. That distinction matters here too. Bitcoin is not an exchange, a mixer, a mining company, or a power plant. Yet the protocol's properties make certain uses possible, and its proof-of-work design deliberately requires real-world resources. The question is not whether those consequences exist. It is how large they are, who bears them, what service they support, and whether the benefits justify the costs.
The crime objection is real, but the comparison is often wrong
Bitcoin has been used to receive ransom payments, sell drugs, launder stolen assets, evade sanctions, and solicit funds for violent groups. The U.S. Department of Justice has documented cryptocurrency fundraising campaigns involving al-Qaeda and ISIS. (DOJ: disruption of terror-finance campaigns) Calling every such example a misunderstanding would be dishonest. A pseudonymous address can receive value from anyone, and the network does not ask why a transaction is being made.
But public Bitcoin transactions are not anonymous in the ordinary sense. They are recorded on a public ledger. Addresses do not automatically reveal a person's legal identity, but investigators can follow flows between addresses and combine that record with exchange records, seized devices, communications, and other evidence. In a case involving the 2016 Bitfinex theft, the U.S. Department of Justice described how investigators traced stolen bitcoin through transactions and laundering attempts. The FBI also recovered part of the Colonial Pipeline ransom after tracing it through Bitcoin's public ledger. Those cases show both sides: criminals can use Bitcoin, and the permanent transaction record can help investigators follow it. (DOJ: Bitfinex case; DOJ: Colonial Pipeline seizure)
The ledger does not make every transaction attributable. Users can transact through intermediaries, split funds, use privacy tools, trade peer to peer, or move into assets and services where public attribution becomes harder. Blockchain analysis is probabilistic and depends on data outside the chain. A cluster of addresses is not a name; an address associated with a service does not establish that every user of the service committed a crime. Transparency helps tracing, but it does not eliminate uncertainty or guarantee law-enforcement success.
The right comparison is also not “Bitcoin crime versus no crime.” Cash, bank wires, shell companies, trade-based laundering, and other payment systems are used in crime too. Their records, visibility, and enforcement mechanisms differ. Raw dollar totals across these channels are not a fair scorecard: they cover different populations, definitions, reporting systems, and periods. Nor is it enough to divide a crypto-crime estimate by every on-chain transfer and announce a definitive “crime rate.”
For example, Chainalysis's 2025 report estimated that 0.14% of attributed on-chain transaction volume in 2024 involved illicit activity. That is a dated, method-specific estimate, not a current or complete census of crime and not a Bitcoin-only figure. Chainalysis revises prior estimates as it attributes additional addresses: its estimate for 2023 rose from $24.2 billion in the earlier report to $46.1 billion in the 2025 report. In its 2026 report, Chainalysis revised the 2024 estimate from $40.9 billion to $57.2 billion. The firm defines which addresses and economic transfers it counts and excludes internal transfers and certain other activity. Its illicit category spans multiple cryptoassets, and stablecoins have become prominent in illicit activity. These changing vintages illustrate why estimates should be dated and treated as lower-bound, methodology-dependent measures rather than definitive crime rates. (Chainalysis 2025 Crypto Crime Trends; Chainalysis 2026 Crypto Crime Report)
That distinction prevents two opposite mistakes. The first is to say “crypto crime is a tiny share of transactions, so it does not matter.” A small share can still represent severe harm, and measured activity can omit unknown addresses and off-chain crime. The second is to treat the presence of criminal transactions as proof that Bitcoin is uniquely criminal. The Financial Action Task Force continues to report gaps in countries' implementation of anti-money-laundering rules for virtual-asset service providers, while also identifying risks involving stablecoins, offshore platforms, and other channels. The policy challenge includes intermediaries and cross-border enforcement, not only the base protocol. (FATF 2024 targeted update; FATF 2026 offshore VASP report)
Bitcoin's permissionless design makes censorship resistant access possible for people who cannot or will not rely on a conventional intermediary. The same openness also means the protocol does not screen participants. That is not an incidental defect that can be removed without changing the system's character. It is a tradeoff. Whether the tradeoff is acceptable depends partly on the value placed on open access and partly on the real effectiveness of law enforcement against harmful uses.
Energy is part of the security budget
Bitcoin's energy use is not mainly caused by users checking balances or nodes independently verifying transactions. It comes from proof-of-work mining. Miners compete to find a valid block by repeatedly hashing candidate data. The work is probabilistic: no miner can know in advance which attempt will succeed. The network adjusts difficulty so blocks arrive, on average, roughly every ten minutes. Nodes then independently check that the proposed block follows Bitcoin's rules.
Proof of work makes rewriting recent transaction history costly. An attacker seeking to replace confirmed blocks would need to produce competing work and outpace the honest chain, while bearing the cost of machines, electricity, facilities, and the risk of losing revenue if the attempt fails. Energy is therefore not a useful computation in the same sense as weather forecasting or rendering a film. It is a costly signal that helps make consensus expensive to counterfeit.
The skeptical response is powerful: if the hashes do no independent task, then energy spent finding them is waste. Bitcoin's defenders reply that the security service is the task. A bank's ledger, a payment network, or a government bond market also uses real resources, but those services are different and are measured in different ways. Comparing only electricity per on-chain transaction leaves out what each system secures, how many users it serves, settlement finality, custody, and the rest of the financial infrastructure. But pointing to those differences does not make Bitcoin's energy use free. It means the comparison requires a clear boundary and a clear account of the service being purchased.
We should distinguish electricity consumption from carbon emissions. The same number of kilowatt-hours can have different emissions depending on when and where it is consumed and which generation responds to that demand. Average grid mix can differ from the marginal source that turns on when demand rises. A miner connected to a fossil-heavy grid, a hydro-rich system, a renewable project with surplus output, or a dedicated off-grid generator will not have the same footprint.
The Cambridge Bitcoin Electricity Consumption Index is useful partly because it is explicit about uncertainty. It cannot meter every decentralized mining site, so it estimates network electricity demand from mining economics and a basket of real hardware, and gives lower, best-guess, and upper-bound estimates. Its greenhouse-gas work then combines estimated electricity use with estimates of mining locations and electricity-generation mixes. Cambridge notes that the emissions model is not a full life-cycle assessment and assumes, for simplification, that marginal emissions equal average emissions. It also says estimates change as data and methodology improve. The correct lesson is not to select whichever point estimate supports a preferred conclusion; it is to preserve the ranges and disclose what the model does not capture. (Cambridge CBECI methodology; Cambridge GHG methodology)
There are other environmental costs besides operational electricity. Mining hardware requires materials and manufacturing, and specialized machines may become economically obsolete before they stop functioning. A cradle-to-gate assessment of mining equipment can help quantify that part of the footprint, but it does not on its own tell us the network's full environmental impact. A complete accounting would need reliable global equipment lifetimes, reuse and recycling rates, manufacturing data, cooling, facilities, and local effects. Those data are less complete than the energy estimates. (Equipment life-cycle study)
“Renewable-powered” does not finish the argument
Mining advocates often point to hydropower, wind, solar, methane capture, or curtailed electricity. These may matter, but labels are not enough. A mine buying renewable certificates is not necessarily consuming new renewable power at the same time and place. A project using otherwise-curtailed energy may help finance generation that would not otherwise be built, but that claim requires evidence about actual curtailment, grid connections, contracts, and what would have happened without the mine.
The relevant question is often counterfactual: what generation changes because the mine operates? If a mine uses electricity that would otherwise have been spilled because transmission was constrained, its marginal emissions could be low, though construction and equipment still have impacts. If the mine's demand causes a gas or coal plant to run more, the marginal emissions may be higher than the system average. If it provides a bankable buyer that enables new renewable construction, it may support investment; if it consumes clean power that households or factories would otherwise use, the opportunity cost matters.
Methane capture claims need the same discipline. Capturing gas that would otherwise be vented or flared and using it to generate electricity can reduce methane release, but combustion still emits carbon dioxide and the result depends on the counterfactual, methane leakage, generator efficiency, measurement, and time horizon. A miner's report that it operates on “waste gas” is not by itself a lifecycle emissions study. The chapter's conclusion should change as audited, site-level data becomes available, not on the basis of slogans from either side.
Flexible load can help a grid, under specific conditions
Bitcoin miners have one genuine operational characteristic that many loads lack: they can often curtail quickly. An ASIC can be switched off without losing the work already completed by the rest of the network, and the operator can resume when electricity prices fall. In a system with variable renewable output or short periods of scarcity, a large interruptible load may absorb electricity at some times and reduce demand at others.
This is not simply a theoretical claim. ERCOT created a voluntary program for large flexible customers, explicitly including Bitcoin mining facilities, to reduce power use during periods of high demand. Researchers have also modeled how flexible mining loads can affect prices, reliability, and emissions in the Texas grid. Such research suggests flexibility can reduce some market stress under the modeled conditions, while also showing that location and response rules matter. (ERCOT program announcement; grid impact study)
But a mine is not automatically a grid battery. It does not store electricity for later delivery; it only stops consuming. Its incentive to curtail depends on mining revenue, power prices, contracts, demand-response payments, and facility design. An operator may reduce load when it is economically beneficial, but grid operators need reliable response during the hours that matter, not just flexibility in principle. ERCOT's own planning documents treat large flexible loads as a distinct forecasting and reliability issue, reflecting uncertainty about how much demand will actually be present or responsive during stressed conditions. (ERCOT large-load analysis)
There is also a public-policy question when a grid operator or utility pays a miner to curtail. The payment may be cheaper than building peaking generation or shedding household load. It may also shift costs among ratepayers, miners, and generators. To judge it, one needs the contract, the counterfactual cost, measured performance, and who ultimately pays. “The miner shut down during a heat wave” is evidence of a response; it is not enough by itself to prove the mine lowered total emissions or electricity costs.
What does society get for the cost?
The hardest question is not whether proof of work consumes energy. It does. The harder question is whether the security and monetary service justify that expenditure, and how that judgment should be made when people disagree about Bitcoin's social value.
One view is that an open monetary network with no central operator is valuable precisely because no company or state can unilaterally rewrite its ownership rules. If people use it as a reserve asset, settlement layer, or collateral, then mining expenditure is an operating cost of providing that security. The amount spent may respond to Bitcoin's price and transaction-fee revenue, so energy use is linked to the market's valuation of the service.
The opposing view is that the service is narrow relative to the electricity and hardware consumed. The base layer processes a limited number of transactions, and users could use more energy-efficient systems for many ordinary payment needs. If Bitcoin's monetary value is mostly speculative, society may be subsidizing an expensive contest whose private benefits accrue to holders while emissions, local pollution, grid costs, and electronic waste are shared more broadly.
Neither argument is settled by comparing Bitcoin's consumption to a country. The analogy makes scale legible but not value comparable. Nor does a large number of transactions on a second layer automatically settle it: layers may change throughput and user experience, but they do not remove proof-of-work security costs at the base layer. The analysis has to connect energy to the security service, then ask who benefits and who pays.
That is why both “Bitcoin uses energy, therefore it is waste” and “Bitcoin secures value, therefore any amount of energy is justified” are incomplete. The first assumes the service has no social value. The second assumes its value exceeds its full external cost. Both are conclusions that require evidence.
What evidence would change the conclusion?
The crime assessment would improve with independent, comparable statistics that state clearly what counts as illicit activity, distinguish Bitcoin from other cryptoassets and stablecoins, account for non-crypto payment systems on comparable terms, and disclose how much remains unobserved. Evidence that illicit activity is increasingly concentrated in intermediaries or assets with weaker transparency would shift attention toward those services. Evidence that Bitcoin's base layer is systematically essential to severe criminal activity at a scale that cannot be mitigated through lawful enforcement would strengthen the case for targeted restrictions. Neither conclusion follows from a handful of anecdotes.
The energy assessment would improve with independently audited mining-location and time-of-use data; marginal emissions measured at the relevant grids; equipment production, lifetime, reuse, and disposal records; and verifiable site-level evidence for curtailed-energy, methane, and demand-response claims. It would also benefit from transparent ways to compare the service Bitcoin provides with the costs of alternative settlement and monetary systems.
Until those data are stronger, the careful conclusion is mixed. Bitcoin is used in crime, but its transparent ledger has also aided investigations, and available estimates have substantial attribution limits. Proof-of-work consumes significant electricity by design, and that consumption underwrites the cost of attempting to rewrite the ledger. Some miners can provide flexible demand, but the benefit depends on the site, grid, contracts, and measured behavior. Renewable and methane claims may be valid in particular cases, but they need counterfactual accounting rather than promotional labels. Environmental costs remain real even when the energy mix is cleaner.
The previous chapter asked whether Bitcoin's market is a fraudulent payment chain. This one asks whether the costs of an open proof-of-work monetary network are socially justified. A fair answer cannot be “yes” just because Bitcoin is useful to some people, or “no” just because the hashes do not produce a conventional product. We need to understand the mechanism, measure the costs as honestly as the data allow, and compare them with the service people actually receive.
Next chapter
The next chapter should examine whether governments can stop Bitcoin or whether an open network can be copied and replaced. That moves from the costs of the network to the political and competitive risks that could undermine the service those costs are meant to secure.