Drex: proposal, architecture and project status

Since 2023, the name Drex has appeared in headlines as “the digital real”, sometimes with promises that it would replace cash or Pix. Neither happened. The project of the Banco Central do Brasil (BC, Brazil’s central bank) tested a platform for about two years, ran into a serious privacy obstacle and was redesigned.

This article explains the original proposal, the architecture tested in the pilot, what the official reports concluded and what is known about the current stage, with the date of each piece of information. When the information comes from news reports rather than a BC document, that will be stated.

What Drex set out to be

According to the official project page, Drex would allow transactions with digital assets settled within a platform run by the BC itself, an “integrated and standardized environment”. Access would always be through an authorized intermediary, such as a bank, which would move money from your account into a Drex wallet.

The page uses a simple example: buying a car. Today, buyer and seller distrust each other, because someone has to act first. In the Drex proposal, a smart contract (a program that enforces rules automatically when conditions are met) would make the money and the ownership change hands at the same time. If one party failed, everything would return to its previous state.

This mechanism has a technical name: delivery versus payment (DvP). When the exchange is indivisible, happening in full or not at all, it is called atomic settlement. This is the central idea of asset tokenization, a topic Token & Tech explains in how real estate tokenization works.

The BC also explains the name. The letters “d” and “r” refer to real digital, the “e” comes from eletrônico (electronic) and the “x” suggests connection, as in Pix (Brazil’s instant payment system).

An important point: Drex was classified as a central bank digital currency (CBDC), but not as “digital cash” you would hold directly at the BC. To understand the variations of this concept around the world, see what CBDCs are.

The tested architecture: wholesale and retail

The pilot’s phase 1 report, published by the BC in February 2025, describes the design in two layers.

Wholesale Drex. This is the part that would actually be central bank money. The BC would issue it to institutions that already hold accounts with it. In practice, it would be the tokenized version of bank reserves, settlement accounts and the National Treasury Single Account.

Retail Drex. This is the part that would reach customers. It would not be issued by the BC, but by banks and payment institutions: tokenized versions of demand deposits and electronic money, one-to-one with the real. The report explains the choice: preserving financial intermediation and banks’ ability to extend credit.

This model resembles what the Bank for International Settlements (BIS) called a unified ledger: a programmable platform where central bank money, tokenized deposits (private money issued by banks, which remains a claim against the issuing bank) and other tokenized assets coexist. In the BIS 2023 annual report, central bank money in this arrangement plays a role similar to that of reserves in the current system: it is the anchor that gives transactions final settlement.

How would a transfer between customers of different banks work? According to the phase 1 report, in three simultaneous steps:

1. the payer’s bank “burns” (removes from circulation) the customer’s retail Drex;
2. the payer’s bank transfers the equivalent amount in wholesale Drex to the recipient’s bank;
3. the recipient’s bank issues retail Drex into the receiving customer’s wallet.

Technically, the pilot ran on a distributed ledger network (DLT, the family of technologies to which blockchain belongs) based on Hyperledger Besu, compatible with Ethereum. The network had 26 nodes (computers that keep copies of the ledger) operated by the BC and the participants. If you want to understand the technical basis, see what Ethereum is.

Phase 1: testing privacy

The pilot’s guidelines were approved by Vote 31/2023-BCB, of February 14, 2023, and the regulations came with BCB Resolution 315, of April 27, 2023, according to the Drex Pilot page. The BC began onboarding participants in July 2023. There were 16 consortia or companies, including banks of various sizes, credit cooperatives, payment institutions and technology companies, about 70 participants in total, according to the opening of the 5th plenary of the Drex Forum, in November 2024. End users did not take part; their transactions were simulated.

Phase 1 had a single use case: the purchase and sale of tokenized federal government bonds between customers of different institutions, with delivery versus payment. The choice was not random. This type of transaction forces several participants to exchange information, which puts privacy at the center of the test.

The report names the problem: the trilemma between privacy, programmability and decentralization. On a shared network, participants can see the recorded transactions. But the Brazilian financial system must respect banking secrecy (Complementary Law 105/2001) and the LGPD (Brazil’s General Data Protection Law), as the report itself notes when citing the project’s guidelines.

The BC tested four privacy solutions based on zero-knowledge proofs (cryptographic techniques that make it possible to prove a piece of information is valid without revealing it) and on network segregation: Anonymous Zether, Starlight, Rayls and Microsoft Nova ZKP.

The report’s conclusion was direct. The solutions made progress on anonymization, but “have limitations that compromise, for now, their adoption”. The text points to “the lack of control by the authorities over the tokens and the limitations in programmability and composability” and states that the solutions “lack maturity” for the requirements of a robust financial system.

In May 2025, BC director Renato Dias de Brito Gomes put a number on this cost in a speech published by the BIS: with one of the privacy solutions tested, Drex processed fewer than 10 transactions per second. He called the problem the “privacy trilemma”: it is hard to achieve privacy, scale and programmability at the same time. In the newsroom’s assessment, a mass payment system such as Pix could not work with that limit.

Phase 2: market use cases

The second phase began in September 2024, with 13 themes chosen from 42 proposals, according to a BC news release. The idea was to test financial services created by third parties, through smart contracts, and to include assets regulated by the CVM (Comissão de Valores Mobiliários, Brazil’s securities regulator).

The themes included credit backed by CDBs (bank certificates of deposit) and government bonds, assignment of receivables, foreign exchange, debentures, agribusiness assets, decarbonization credits, real estate and vehicles.

In February 2025, when publishing the phase 1 report, the BC announced it would not add new cases to phase 2, despite having received 101 proposals. The official note says the pilot “proved technologically challenging” and that the BC “will only move forward with solutions that ensure privacy, data protection and transaction security”.

In the same May 2025 speech, Gomes said the first results of phase 2 confirmed the potential of delivery versus payment with atomic settlement. But he also said that, in the short term, the BC would remain technology “agnostic”, improving existing infrastructure, and would keep the flexibility to migrate to a unified ledger on DLT “when the time comes”.

The 2025 change of course

This part relies on news reports, because the newsroom did not find an official BC note detailing the decision. The outlets are cited one by one.

In August 2025, according to Exame, the BC’s Drex coordinator, Fabio Araujo, told the newspaper Valor Econômico that the privacy solutions tested were good, but apparently insufficient. The project would be split into a long-term part, with DLT, and a short-term part, without DLT, focused on lien reconciliation: checking which assets have already been pledged as collateral for credit, and where. The timeline cited was the second half of 2026.

On November 4, 2025, Finsiders Brasil reported that the BC had told participants, at a meeting, that it would shut down the pilot platform on November 10 and begin building a new infrastructure, defining the use cases first and the technology afterward.

In December 2025, Exame reported statements by BC governor Gabriel Galípolo. According to him, the technology tested “did not prove viable”, but the project was not abandoned: “Today there are other ways to achieve what we want”. The report describes the new stage as an interoperable environment for tokenized assets, settled in money issued by the monetary authority.

Where it stands in October 2026

What is official, checked on October 8, 2026:

• the Drex Pilot page lists phase 1 and phase 2 as “Closed”;
• the phase 2 report appears as “under construction”;
• the project page still describes Drex in the future tense (“will allow”), with the same integrated platform design.

What the newsroom did not find: an official announcement with a delivery date, final scope or technology for the new stage, nor confirmation that lien reconciliation is already in operation. The second-half-of-2026 timeline comes from 2025 news reports and should be read as a reported plan, not as fact.

So be wary of any text that treats the 2023 design (a Drex wallet for citizens, tokenized retail money on DLT) as the current plan. It was the plan tested in the pilot, and the BC itself concluded the technology did not meet the requirements.

What this means for you

First, there is no Drex to buy, invest in or “activate”. The project never reached the public, and the stage under discussion is infrastructure between institutions. Messages asking you to register, deposit money or provide data to “receive your Drex” are scams. The guide to digital financial scams shows how to respond.

Second, asset tokenization keeps advancing outside Drex, in private and regulated initiatives. That does not eliminate issuer, liquidity or legal risks, as explained in tokenization does not eliminate risk.

Third, comparing other forms of digital money helps separate concepts. Stablecoins are issued by private companies; CBDCs, by central banks. The differences are in stablecoins versus CBDCs. And financial services embedded in non-financial apps follow another logic, explained in embedded finance.

Key takeaways

• Drex was designed as a BC platform with two layers: wholesale, issued by the BC to institutions, and retail, tokenized deposits and e-money issued by banks and payment institutions.
• The pilot, on a Hyperledger Besu network, showed that the privacy solutions tested were not mature enough for the banking secrecy and data protection required in Brazil.
• As of October 2026, both phases of the pilot are listed as closed. The new stage, without blockchain and focused on credit collateral, has been reported by the press, but no official date has been announced.