Who counts as a crypto broker mattered less than what brokers could not know

The week

Most infrastructure bills are about roads, bridges and broadband. In August 2021, one of the biggest fights over the largest infrastructure bill in years was about cryptocurrency, and specifically about a single definition.

The bill was the bipartisan infrastructure package negotiated by a group of senators over the summer, with roughly $550 billion in new spending. Its authors needed ways to pay for it without raising tax rates. One of the offsets was tighter tax reporting for digital assets, which the Joint Committee on Taxation estimated would raise about $28 billion over ten years. The estimate assumed that better reporting would bring in tax that was already owed but going unpaid.

The IRS had been moving in this direction for years. Since 2014 it had treated virtual currency as property, so each sale or exchange could produce a taxable gain or loss. In 2017 a federal court ordered Coinbase to hand over records on thousands of its customers in response to an IRS summons. For the 2020 tax year, the IRS put a question about virtual currency transactions on the front page of Form 1040. Routine reporting by brokers, of the kind long applied to stocks, was the next step.

On August 10, 2021, the Senate passed the Infrastructure Investment and Jobs Act by 69 votes to 30. Information reporting is the system of forms, like the 1099s many people receive each January, through which banks, brokers and employers tell the IRS what they paid or sold on someone’s behalf. It is how the IRS knows what to expect on a tax return.

The bill defined a digital asset as any digital representation of value recorded on a cryptographically secured distributed ledger or similar technology. It then widened the definition of a broker to include any person who, for consideration, is responsible for regularly providing any service effectuating transfers of digital assets on behalf of another person. That covers, roughly, anyone who is paid to regularly move crypto for other people.

It did several more things. It treated digital assets as specified securities for cost basis reporting. Cost basis is simply what you paid for something, which is needed to work out your gain or loss when you sell. It extended broker-to-broker transfer statements to digital assets, the paperwork that follows an investment when it moves from one firm to another. It added a new return for transfers from a broker account to an account not kept by a broker. And it brought digital assets within the rule that requires businesses to report cash receipts over $10,000. The changes were to apply to returns filed after 2023.

The broker definition set off alarms. Industry groups and several senators argued that it could reach miners and validators, the computers that process and confirm transactions on a blockchain, as well as software developers who never touch a customer’s money. Competing amendments to narrow it circulated that week. One, from Senators Wyden, Lummis and Toomey, would have excluded miners, validators and developers. Another, from Senators Warner, Portman and Sinema, would have exempted some miners but not validators. A merged compromise was announced, but it failed to get unanimous consent for a vote amid an unrelated dispute over a defense spending amendment.

The rest of the market was busy too. Earlier in the month, the chair of the Securities and Exchange Commission, Gary Gensler, told a security conference that crypto markets needed stronger investor protection and asked Congress for more authority. On the day of the Senate vote, the cross-chain protocol Poly Network lost roughly $600 million to an exploit, most of which the attacker later returned. Regulators and customers alike were learning how little of the activity on a blockchain maps neatly onto the records traditional finance relies on.

The text stood. The House passed the bill on November 5 by 228 votes to 206, and it was signed into law on November 15, 2021. The details of how brokers would report were left to Treasury and the IRS to write in regulations. For crypto’s builders, the week was about whether they were in or out. For banks and brokerages, a different and harder question was waiting.

Most of the August debate was about the edges of the definition: whether a miner, a validator or a wallet developer could be a broker. That mattered to the crypto industry. It barely touched banks, broker-dealers and wealth managers considering digital asset services. A firm that buys and sells digital assets for customers and holds them in custody was a broker under every version of the text.

For those firms, the provision did something familiar. It applied the securities information reporting regime, gross proceeds, cost basis and transfer statements, to a new asset class. Gross proceeds is the total amount a customer received from a sale. Brokerages have reported these figures for stocks and bonds for years.

Familiar did not mean easy. Securities reporting works because the data it needs exists inside the brokerage system and travels between brokers in standard formats. Digital assets were designed to move without intermediaries, so much of that data lies nowhere a broker can reach. The provision therefore exposed three data gaps, each harder than the scope question: cost, what the customer paid; movement, where the asset has been; and counterparties, who was on the other side of a payment.

The sections below begin with a short primer on wallets and custody and then take the three gaps in order. A hypothetical wealth manager’s first year shows the work inside one firm. The closing sections set out what the gaps cost, what a reporting-ready service looks like, and the questions to answer before the first client trade.

Who was clearly in, and how they got there

Regulated firms reached digital assets by different routes in 2021, and the route shaped who would carry the reporting duty. National banks had interpretive letters from the Office of the Comptroller of the Currency, issued in 2020 and early 2021, confirming they could provide crypto custody. Broker-dealers faced unresolved custody questions under securities rules, and the SEC had only recently set out a narrow path for special purpose broker-dealers. Credit unions generally reached the market through partnerships with third-party providers.

Registered investment advisers usually do not hold client assets themselves. Where an adviser’s clients hold crypto at a custodial platform or bank, the reporting duty generally falls on that custodian, not the adviser. The adviser still sees the consequences, because its clients and their accountants will ask why the figures do not reconcile. Which entity is the broker for a given arrangement is a question for counsel.

Wallets, custody and the blockchain

A digital asset such as bitcoin is, at bottom, an entry in a shared public ledger called a blockchain. It works like an enormous accounting book that thousands of computers keep identical copies of. Every transfer is a new line in the book, visible to anyone, recording that a certain amount moved from one address to another.

An address is a long string of letters and numbers, a little like an account number with no name attached. Control of the assets at an address comes from holding a private key, a secret code that authorizes transfers. Whoever holds the key controls the funds.

A wallet is the software, or sometimes a small device, that stores those keys. There are two broad kinds. A custodial wallet is held by a firm on the customer’s behalf, much as a brokerage holds shares in street name. A self-hosted wallet is held by the customer directly, with no institution involved. Custody, in financial terms, simply means holding assets for someone else.

Why the ledger is not the firm’s record

Custodial firms rarely keep each customer’s assets at a separate address. More commonly they pool many customers’ holdings in shared addresses, called omnibus wallets, and track who owns what in their own internal books. A customer’s sale inside the platform may never touch the blockchain at all; it is an entry in the firm’s ledger.

That has two consequences. The firm’s internal books, not the blockchain, are the record that tax reporting must draw on. And the firm must reconcile its internal books against what it actually holds on chain, every day, because a mismatch means either a missing asset or a misstated customer balance.

This design is the root of the reporting challenge. In traditional finance, almost every transaction passes through an intermediary that knows who the customer is and keeps records. On a blockchain, assets can move straight from one address to another with no intermediary at all. The ledger shows that something moved. It does not show who owned it, what they paid for it or why it moved.

A few other terms appear below. Staking is a way of earning new tokens by helping to secure certain blockchains, somewhat like earning interest. An airdrop is a distribution of free tokens to many addresses, often as a promotion. A fork happens when a blockchain splits into two versions, which can leave holders with units of a new asset. Each creates assets with no obvious purchase price.

Cost: basis the broker never saw

For securities, brokers report cost basis for covered securities, those bought in the account after the basis rules took effect, and track adjustments such as splits. A broker knows the basis because it executed the purchase.

Digital asset customers commonly buy on one platform, move assets to a self-hosted wallet, and later deposit them with another provider to sell. The selling broker sees the sale. It has no reliable record of the purchase. The customer may have bought years earlier, on an exchange that no longer exists, in many small lots. Some assets arrived through staking rewards, airdrops or forks whose tax treatment was itself unsettled in 2021.

Consider a customer of a regional bank’s new digital asset service. Over several years she bought small amounts of bitcoin on two different exchanges, moved them to a hardware wallet in her desk drawer, and now deposits the whole amount with the bank and sells. The bank can report precisely what she received from the sale. What she paid is scattered across two exchanges’ records, one of which she can no longer access. The bank has nothing to report for basis, and anything she tells it, it cannot check.

Two ways of counting that do not agree

A subtler problem is how basis is counted. For securities, basis is tracked account by account, and a broker applies a method such as first in, first out to the shares in that account. Many crypto holders in 2021 calculated basis across all their holdings everywhere, treating units in different wallets and exchanges as one pool.

The two approaches give different answers for the same sale. A broker can only count the lots it holds. A customer counting across every wallet may pick a different, older or cheaper lot. Until the rules said which approach governed, a broker’s form and a customer’s return could disagree with neither being clearly wrong.

Assets with no purchase price

Assets received rather than bought raised their own questions. The IRS had addressed hard forks and airdrops in a 2019 revenue ruling, treating new units as income when the holder could control them. The treatment of staking rewards was still contested in 2021. A broker holding such assets had to decide what basis to carry, if any, and whether to report anything at the point of receipt.

Corporate actions without a corporation

Securities operations teams handle corporate actions, such as splits, mergers and spin-offs, using announcements from issuers and data vendors. Digital assets have equivalents with no issuer to announce them: chain splits, token migrations to new contracts and airdrops to holders.

Each one forces a custodial firm to decide whether to support the event. Supporting it means crediting customers with a new asset, assigning it a basis and adding it to reporting. Declining means customers who expected the new units may complain, and the firm must explain why. Firms needed a written policy for these events, with tax, operations and legal sign-off, before the first one arrived.

Building a two-tier book

The practical consequence was a two-tier book. Assets bought and sold inside the firm could carry clean basis. Assets transferred in would carry none, or a customer-supplied figure the firm could not verify. Firms needed to tag every lot by origin from day one, decide how to treat customer-supplied basis, and design statements that made the gaps clear. A lot is a single purchase, with its own date and price; tagging it by origin means recording whether it was bought inside the firm or arrived from elsewhere.

A firm that mixed the two tiers would issue forms that were wrong in ways customers would find at tax time. Few things damage a client relationship faster than a tax form that tells the IRS a customer made a gain they did not make.

Customer-supplied basis also carries a trade-off. Accepting it makes statements more useful and reduces surprises at tax time. It also puts the firm’s name next to a number it cannot verify, and an inflated figure would understate gains. Firms in 2021 had to choose whether to display such figures, how to label them and whether to carry them into any future form. Whatever they chose, they needed to keep the customer’s supporting evidence.

What a statement should say

Statements are where the two tiers meet the client. A useful statement shows each lot with its acquisition date and cost, and marks the source of every cost figure. Lots with no known cost should say so plainly rather than show zero, since a zero basis would overstate gains if anyone copied it into a return. A short note can explain why the firm cannot know the cost of assets bought elsewhere and what records the client should keep. Clients and their accountants can then reconcile the statement with their own records instead of calling to ask what the numbers mean.

Movement: transfers without a common language

The second gap is movement. In securities, when a customer moves shares between brokers, the transferring broker sends a statement with basis and acquisition dates. The law extended that duty to digital assets moving between brokers. It could not extend it to self-hosted wallets, where no broker exists to send anything, and in 2021 the brokers themselves had no shared format or network for exchanging such statements.

In the securities world, those transfers travel over well-established industry plumbing, standard messages that every brokerage understands, so the paperwork arrives automatically with the shares. Digital assets had nothing comparable. Even where two brokers were both willing to share data, there was no agreed way to do it.

The new return on transfers out

The law tried to narrow the self-hosted wallet gap from the other side. Its new return for transfers out of a broker account to an account not kept by a broker meant the IRS would learn that assets had left the reporting system. It would not learn where they went or what happened to them next. The return records a door being used, not the room on the other side.

Self-hosted wallets were already a sensitive subject. In December 2020, FinCEN, the Treasury’s financial crimes unit, had proposed recordkeeping and reporting for certain transactions with self-hosted wallets. The proposal drew heavy criticism and was not finalized. Tax reporting met the same wall: the law could reach the firm on one side of a transfer, but never the person on the other.

Classifying every deposit and withdrawal

That made transfer data an industry plumbing problem layered on a firm problem. Inside the firm, every deposit and withdrawal needed to be classified: an incoming transfer from another broker, from the customer’s own wallet, or a payment from someone else. Each answer implies different reporting. The blockchain shows that an asset moved. It does not show whose wallet it came from or why. Firms with mature securities operations already had transfer processing teams. They needed those teams to learn a new asset class with none of the standard messaging they relied on.

A deposit looks the same on the ledger whether it is the customer’s own assets returned from elsewhere, assets bought from another dealer, or a payment of a debt. Only the surrounding records, and sometimes a phone call to the customer, reveal which it is. The firm must build the habit, and the systems, to ask and record the answer every time, because the asset will never explain itself.

Fees, dust and partial transfers

Small details multiply the work. Network fees are often paid in the asset being moved, so a customer who sends one unit receives slightly less. The firm must decide how to record the difference and whether any part of it is a disposal. Tiny unsolicited deposits from unknown addresses, sometimes sent to track wallets, arrive in customer accounts and still need classifying. A transfer system that cannot handle these cases produces a steady stream of exceptions for staff to clear by hand.

One classification, three users

There is a control benefit here too. The same classification that drives tax reporting feeds anti-money laundering monitoring and sanctions screening of wallet addresses. Anti-money laundering monitoring looks for patterns that suggest criminal funds; sanctions screening checks that the firm is not dealing with prohibited people or entities. A firm that builds one transfer classification service, used by tax, compliance and customer service, avoids three teams reaching three different answers about the same deposit.

The anti-money laundering side was ahead of tax in one respect. Under existing rules, FinCEN had made clear that its travel rule, which requires certain information to accompany transfers of $3,000 or more, applied to convertible virtual currency. International standard-setters had adopted a similar expectation in 2019. By 2021, industry groups were building messaging protocols so that exchanges could pass sender and recipient details to each other.

Those protocols were the closest thing to shared plumbing the industry had. A firm choosing one for anti-money laundering purposes was, in effect, choosing the channel through which tax transfer data might one day travel. Tax and compliance teams had reason to make that choice together.

Counterparties: the $10,000 rule and the unknown sender

The third gap came from the cash reporting rule in section 6050I of the tax code. It requires any trade or business that receives more than $10,000 in cash in one transaction, or related transactions, to file a report with the payer’s name, address and taxpayer identification number. Treating digital assets as cash for this purpose meant that a business accepting a large crypto payment would need to identify the sender.

Why the cash rule fits badly

The cash rule was built for a world where the payer is physically present. A business files the report, Form 8300, within 15 days of the payment and must give the payer a written statement by the end of January of the following year. The related-transactions test stops payers from splitting a large payment into smaller ones to dodge the threshold.

This duty fell on recipients, not brokers, which is what made it so broad. A car dealer, a law firm or a commercial bank customer accepting crypto would carry it. In a peer-to-peer payment, one person paying another directly with no institution in between, the recipient may know only a wallet address. Critics called it unworkable and a privacy problem.

Suppose a boat dealer agrees to accept a large crypto payment for a yacht. With cash, the buyer is standing in the showroom, and the dealer can ask for identification. With crypto, the payment may simply arrive at an address from another address. Unless the dealer has collected the buyer’s details separately, the report the law contemplates cannot be filled in.

Valuation adds a further wrinkle. Whether a payment crosses $10,000 depends on the asset’s value at the moment of receipt, which can move sharply within a day. A business needs a consistent pricing source and a record of the time of receipt to know whether it even has a duty.

Where banks come in

For banks, the relevance was indirect but real: business clients accepting digital assets would ask how to comply, and some would expect their bank or payment processor to collect sender information on their behalf. Whether the reporting duty reaches a particular business, and what its bank may collect on its behalf, is a question for lawyers.

Payment processors that convert crypto to dollars before the merchant receives anything change the picture. If the merchant receives dollars, its own cash reporting duty may look different, and the processor may face broker obligations instead. Banks offering merchant services, and their partners, needed to understand which side of that line each product was on.

Which business clients are exposed

The businesses most likely to meet the rule were those already used to large cash payments: car and boat dealers, jewelers, real estate professionals and some law firms. Many of them already filed cash reports and had procedures for identifying payers. Their bankers could help most by pointing out that the same procedures now needed to cover crypto, and that identification had to happen before payment, not after.

A wealth manager’s first year

To see what closing these gaps involves, follow a hypothetical mid-sized wealth manager as it prepares to let clients buy, hold and sell digital assets.

The first decision is about records, not technology. The operations team agrees that every holding will be recorded lot by lot, and that each lot will carry a field saying where it came from. The options are bought here, transferred in from another firm, deposited from a client’s own wallet, or received through something like staking. A second field records where the cost figure came from, if there is one: the firm’s own trade record, another firm, or the client. That small design choice, made before the first trade, is what later lets the firm separate clean figures from unverifiable ones.

Next, the team builds a single place to classify movements. Whenever an asset arrives or leaves, the service asks the same questions. Did it come from a known firm, from an address the client has registered as their own, or from an address nobody recognizes? Is it leaving for a client’s wallet or going to pay a third party? Where the answer is not obvious, a client service representative contacts the client and records the answer. The tax team, the financial crime team and the client service team all read from this one service rather than keeping their own spreadsheets.

The firm then writes its basis policy. When a client supplies a purchase price for assets transferred in, the firm accepts it, labels it on statements as client-provided, and keeps the supporting note the client sent. Statements spell out which figures the firm can stand behind and which it cannot.

The firm also chooses its custodian carefully. It will not hold keys itself, so it selects a regulated custodian and writes into the agreement what data the custodian will provide each day: lot-level positions, every deposit and withdrawal with its on-chain reference, and the custodian’s own classification. A daily reconciliation compares the custodian’s records with the firm’s books. Breaks are investigated the same day, before they turn into wrong statements.

At onboarding, clients are asked to register any self-hosted wallets they intend to use, and to confirm they control them. The step adds a few minutes to account opening. It saves far more time later, because deposits from registered addresses can be classified automatically.

Finally, the firm prepares for questions from business clients. A family-owned business that banks with the firm mentions it is thinking of accepting crypto for large orders. The relationship manager has a short note, reviewed by counsel, explaining that identifying the payer may be the business’s responsibility and suggesting it collect payer details before accepting payment.

Before its first year end, the firm runs a dry run of tax statements on test accounts that include every awkward case: lots transferred in, client-supplied basis, a staking reward and a withdrawal to a client wallet. The test exposes one gap: the statement template has nowhere to show a client-provided figure. It is fixed before real statements go out.

The firm names an owner for each piece. Operations owns the lot and transfer data model. The tax reporting team owns the basis policy and statements. Financial crime owns screening rules that use the transfer classification. Technology owns the integration with the custodian. A single executive sponsor resolves disputes between them, because the first argument over whose definition of a transfer wins will come within weeks.

Nothing here requires exotic technology. It requires agreeing on what to record, recording it every time, and making sure different teams share one version of the facts. The firms that found this hard were usually those that launched with a trading interface first and tried to reconstruct the records later.

One cause, three costs

The three gaps share a cause. Information reporting assumes that intermediaries see both ends of a transaction and talk to each other. Digital assets often have neither property.

For a bank, brokerage or wealth manager, that shows up in several places on the balance sheet and in the risk register:

• Operating cost. Lot tagging, transfer classification and basis handling require systems and staff that a pilot budget rarely includes.
• Regulatory risk. Incorrect information returns carry penalties and invite scrutiny, and the IRS will compare what firms report with what customers file.
• Customer trust. Wealthy clients and their accountants notice wrong tax forms, and they remember which firm sent them.
• Compliance integration. The same data supports anti-money laundering and sanctions controls, so weak data creates financial crime risk as well as tax risk.

How the burden differs by firm

The weight of these costs depends on what a firm already has. A large broker-dealer has tax reporting engines, transfer teams and lot-level systems for securities, and its work is to extend them to a new asset class. A regional bank entering through a partner may have none of that in-house and depends on the partner’s data. A credit union offering access through a fintech provider is further removed again, and its exposure is mostly to member trust and to the quality of its provider’s records.

The common thread is that a firm can outsource the work but not the consequences. If a partner’s records are wrong, the customer’s complaint arrives at the institution whose name is on the account.

Second-order effects on products

Reporting duties shape the products firms offer. A firm worried about unverifiable basis may discourage transfers in, or accept them only from registered wallets. That keeps its books cleaner but frustrates customers who value moving assets freely, which was much of the appeal of crypto in the first place.

There is also a displacement effect. Reporting applies to intermediaries that hold customer assets, so activity that moves to self-hosted wallets and non-custodial services leaves the reporting net. A regulated firm that makes its service harder to use may simply push customers toward venues with fewer controls. Firms had to balance clean records against a customer experience that kept assets, and data, inside the regulated perimeter.

The case for waiting

The strongest argument against acting in 2021 was that the rules did not yet exist. Treasury would have to write regulations, and their shape was unknown. A firm that built detailed reporting logic early risked rebuilding it when the rules arrived, and a firm unsure whether it would stay in the business risked spending on a service it might close.

That argument was right about forms and wrong about data. Form layouts, thresholds and deadlines did change. The need to know each lot’s origin, each transfer’s purpose and each payment’s sender did not, because every plausible version of the rules depended on those facts. Firms that waited on the forms lost little. Firms that waited on the data found it could not be recreated later for assets already moved.

Firms that started in 2021 by mapping, for each customer asset, what they knew about its cost, path and counterparties found the eventual forms manageable. Firms that waited for the forms found that the hard work was the data model, not the form. A data model is simply the agreed structure for what the firm records about each asset and transaction: the fields, their meanings and how they connect.

Records built for reporting from day one

The target is a digital asset service whose records are designed for reporting from the first day. Every lot carries its origin, acquisition date and, where known, its cost, with a clear flag where the cost came from the customer and was not verified. Every deposit and withdrawal is classified by source and purpose by one shared service that tax, compliance and client service all use. A written policy sets out how customer-supplied basis is accepted, labeled and reported. Client statements explain plainly where the firm’s information is complete and where it is not. And relationship managers have a short, counsel-approved note for business clients asking about large crypto payments.

Telling clients before the forms do

Clients rarely read regulations, but they do read tax forms. A firm that explains its approach before the first form arrives avoids most of the anger that follows a surprise. The explanation can be short: what the firm will report, what it cannot know, which records clients should keep and how to give the firm a purchase price for assets transferred in.

Advisers and relationship managers need the same explanation in more depth, because they will take the first calls. A one-page internal guide, with answers to the questions clients ask most, keeps those answers consistent across the firm and stops staff from improvising tax advice.

Evidence an examiner or auditor would ask for

A firm that has built this can show it quickly. The evidence is ordinary:

• The data dictionary defining each lot and transfer field, with an owner for each.
• Daily reconciliations between the firm’s books and on-chain or custodian holdings, with breaks and their resolution.
• The basis policy and a sample of client-supplied figures with their supporting evidence.
• Transfer classification records showing how unclear deposits were resolved and by whom.
• Statement testing results from dry runs before each year end.

Controls that keep the records clean

Good data degrades without controls. Changes to a lot’s basis or origin should require a second person to approve them and leave an audit trail. Reclassifying a transfer after the fact should record who changed it and why. Changes to the logic that calculates gains should go through the same change management as any other financial system, with testing against known cases. Access to edit customer tax data should be limited to a small, named group.

These controls are standard in securities operations. The risk in a new digital asset service is that they are skipped in the rush to launch, on the theory that volumes are small. Volumes grow faster than control frameworks catch up, and errors made in the first year carry forward in every later basis calculation.

The usual errors

• Treating it as a forms project. The form is the last step. The data has to exist first.
• Mixing clean and unknown basis. Blending lots bought in-house with lots transferred in produces numbers that look precise and are wrong.
• Building separate answers for each team. Tax, compliance and customer service classifying the same transfer three different ways.
• Assuming the blockchain will tell you. It shows movement, not ownership or purpose.
• Leaving business clients to guess. Clients accepting digital assets will turn to their bank first.

Before the first client trade

• For every digital asset we hold for customers, can we say whether it was bought here or transferred in?
• What do we do when a customer tells us what they paid, and how do we show that figure on their statement?
• Who classifies each incoming and outgoing transfer, and do tax, compliance and client service all rely on the same classification?
• How would we exchange transfer information with another firm today, and what happens when the other side is a self-hosted wallet?
• If the reporting forms were finalized tomorrow, which data would we be missing?

With hindsight

The rules took far longer than the law implied. In late 2022 the IRS told brokers they need not apply the new reporting provisions until final regulations were issued. Treasury and the IRS proposed rules in August 2023, introducing Form 1099-DA, the new information return for digital asset sales. They finalized rules for custodial digital asset brokers in mid-2024. Five years passed between the law and its first full reporting cycle.

The market changed in the meantime in ways that sharpened the cost gap. The collapse of FTX in November 2022, and the bankruptcies of lenders such as Celsius earlier that year, left many customers without access to the exchange records that held their purchase history. The scenario of a customer who bought on an exchange that no longer exists stopped being hypothetical for a large number of people.

Gross proceeds reporting began with 2025 transactions, and basis reporting applies to assets acquired in the customer’s account from 2026. Basis for assets transferred in from elsewhere remains the customer’s problem. That is the cost gap, written into the rules: brokers report what they can see, and the rest stays with the taxpayer.

The counting problem was settled too. Alongside the final rules, the IRS required taxpayers to track basis wallet by wallet and account by account from 2025, and gave them a one-time way to allocate their remaining basis across wallets. The universal pool many holders had used gave way to the account-level logic of securities. Separately, a 2023 IRS ruling said staking rewards are income when the holder gains control of them, which settled one of the open questions about assets with no purchase price. The IRS also gave penalty relief for good-faith reporting of 2025 transactions and deferred reporting of some activity, including staking, lending and wrapping, until further guidance.

The final rules did not implement broker-to-broker transfer statements for digital assets, or the return on transfers to outside wallets. The movement gap is still open, because the shared plumbing the law assumed has yet to be required.

A separate rule extending broker reporting to certain decentralized finance participants, finalized in December 2024, was overturned by Congress under the Congressional Review Act in April 2025. The argument about the edges of the definition, which dominated August 2021, ended up being settled years later and largely in favor of a narrower scope.

For the $10,000 rule, the IRS said in January 2024 that businesses need not report digital asset receipts until regulations are issued. The counterparty gap remains unresolved in practice.

The context for banks shifted as well. In January 2025 the SEC withdrew a 2022 staff accounting bulletin that had made crypto custody expensive for banks to offer. In March 2025 the OCC confirmed that national banks could provide crypto custody without first obtaining supervisory non-objection. In July 2025 the GENIUS Act created a federal framework for payment stablecoins. More regulated institutions now have a path into the business, and each one inherits the data problems described above.

Abroad, the same gaps drove similar rules. The OECD’s Crypto-Asset Reporting Framework set an international standard for exchanging crypto tax data, and the EU’s implementing directive applies from January 2026.

All three gaps identified above are still visible in what the rules require, and firms should confirm current obligations with counsel. The hard part of a new reporting regime is rarely the form; it is whether the organization records the right facts, in a consistent way, from the first transaction. Firms that build that foundation early tend to find each new rule an adjustment rather than a scramble.