Technology & Design Choices

1. Global Landscape — Architecture Models

Central bank digital currency technology choices fall into four broad architecture models. Each reflects a jurisdiction's priorities regarding scalability, sovereignty, vendor risk, and innovation ambition. No single model dominates; the choice depends on the central bank's existing infrastructure, legal framework, population scale, and policy objectives.

This guidance covers all 20 jurisdictions tracked in this domain. Of these, only six have CBDCs that are launched/live (

Bahamas Sand Dollar,

Nigeria eNaira,

Jamaica JAM-DEX, Eastern Caribbean DCash [pilot concluded / successor suspended],

China e-CNY [pilot-scale], and

India e₹ [pilot]); the rest are at pilot, research, or design stage, with the

United States uniquely anti-CBDC.

Model A: Centralized Ledger

Countries:

China (e-CNY, hybrid centralized core),

Russia (digital ruble),

Jamaica (JAM-DEX),

Nigeria (eNaira — account-based on permissioned DLT),

EU (digital euro, centralized N-XT)

Centralized architectures place the core ledger under direct central bank control, without distributed ledger technology as the authoritative production record. The central bank operates the single authoritative record of all issuance, redemption, and — in some designs — individual transactions.

China (e-CNY)

• Platform: In-house PBOC core infrastructure (Digital Currency Research Institute, established 2017), with 22 authorized operators as of April 2026 (12 banks added 2 April 2026); integrated with Alipay and WeChat Pay as front-ends
• Ledger model: Hybrid centralized/distributed architecture — a centralized PBOC ledger, not a public blockchain. From 1 January 2026 the model combines accounts, token strings, and smart contracts
• Throughput: Reported 300,000+ TPS (NOT VERIFIED against primary PBOC technical documentation — figure appears in secondary sources only)
• Scalability evidence (self-reported, PBOC): 3.48 billion cumulative transactions worth CNY 16.7 trillion (~USD 2.37 trillion) by November 2025; 2.25 billion wallets across 26-29+ cities in 17 provinces

Russia (digital ruble)

• Platform: Centralized in-house CBR platform with secure electronic digital signatures, deliberately built in-house to mitigate sanctions risk on technology imports
• Ledger model: Centralized hybrid platform; the specific DLT (if any) is not publicly disclosed by the CBR. One wallet per person on the CBR platform
• Scalability evidence (self-reported, CBR, June 2025 pilot): ~2,500 wallets, 15 banks, 150+ localities; 63,000+ P2P transfers, 13,000+ B2C payments. Narrow live government/budget use since 1 January 2026; legally-mandated large-scale launch planned 1 September 2026

Jamaica (JAM-DEX)

• Platform: eCurrency Mint (Ireland) — DSC3 (Digital Symmetric Core Currency Cryptography); ~0.3s settlement, quantum-resilient, 24/7, standards-based APIs; integrated with JamClear RTGS
• Ledger model: NOT blockchain / NOT DLT — a centralized ledger controlled by the Bank of Jamaica; "digital bearer instrument" tokens (DSC3 cryptography secured independently of the ledger/wallet)
• Scalability evidence (self-reported, BOJ): ~282,274 registered wallets (end-2024); ~0.1% of currency in circulation; transaction value +550% YoY, volume +267% YoY in 2025 (driven by existing users, not new sign-ups)

Nigeria (eNaira)

• Platform: Hyperledger Fabric (permissioned DLT), built by Bitt Inc. (Barbados); CBN sole validator; stress-tested >2,000 TPS (up to 3,000); ISO 20022 via Bitt Numa; NIBSS-integrated September 2022
• Ledger model: Account-based — a centralized CBN ledger of all wallets/balances (not token-based). Although Hyperledger Fabric is technically DLT, the CBN is the sole validator and the architecture functions as a centrally-controlled ledger
• Scalability evidence (self-reported, CBN/IMF — flag): 98.5% of wallets inactive after one year (IMF WP/23/104); ~2.2 million transactions by July 2024; 0.37% of currency in circulation (Feb 2025). Cautionary tale of failed adoption; Nov 2025 pivot to wholesale CBDC focus

EU (digital euro)

• Platform: Centralized Eurosystem platform built on TIPS; six NCBs build the core settlement components; five competitively procured external components from EU-controlled providers (Sapient, Feedzai, Almaviva, Giesecke+Devrient, Senacor; framework agreements est. EUR 432.1 million, max EUR 1.164 billion)
• Ledger model: Centralised settlement (NOT DLT) — the in-house N-XT (New Euro Exchange Technology) engine using a UTXO data model (the same data model as Bitcoin, but on a centralized platform). Listed here under Model A on the strength of its centralized settlement; it draws DLT concepts (UTXO) without DLT consensus (see Model C discussion)
• Throughput: NOT PUBLISHED by the ECB as of June 2026

Key characteristics of Model A:

• Maximum central bank control over the ledger and all monetary operations
• No dependency on DLT consensus mechanisms — simpler operational model
• Scalability limited only by database engineering, not by consensus protocol overhead
• Risk concentration: single point of failure if core infrastructure is compromised
• Privacy architecture is a policy choice, not a technology constraint

Model B: DLT-Based Architecture

Countries:

Bahamas (Sand Dollar / Cortex DLT), Eastern Caribbean (DCash / Hyperledger Fabric [suspended]),

Sweden (e-Krona pilot / R3 Corda),

Singapore (wholesale SGD Testnet),

Thailand (Inthanon / retail pilot),

Australia (Project Acacia multi-DLT)

These projects use distributed ledger technology as the core settlement infrastructure, distributing transaction validation and record-keeping across multiple nodes.

Bahamas (Sand Dollar)

• Platform: NZIA Limited (now Movmint) — proprietary Cortex DLT; partners IBM, Zynesis, Norton Rose Fulbright. HCI infrastructure, two-layer encrypted storage
• Ledger model: Private, permissioned distributed ledger validated through a network (not public mining)
• Offline capability: Yes — supports offline/pre-authorized transactions that sync when connectivity returns (critical for hurricane/archipelagic resilience); hybrid wireless network + hardware nodes. Under development, not yet fully deployed
• Scalability evidence (self-reported, CBOB): Circulation B$2,482,045 (~0.39% of physical cash, early 2025); ~144,000 consumer wallets, 2,100 business wallets (Nov 2024)

Eastern Caribbean (DCash) — PILOT CONCLUDED, SUCCESSOR SUSPENDED

• Platform: Bitt Inc. on Hyperledger Fabric, hosted on Google Cloud; nodes across ECCU states
• Ledger model: Distributed ledger (Hyperledger Fabric), private permissioned with certificate-based identity
• Offline capability: DCash 1.0 had none — its absence was flagged, and the 2022 outage exposed reliance on connectivity. DCash 2.0's RFVI required offline payment capability, but DCash 2.0 development was suspended by the ECCB Monetary Council on 13 February 2026 in favor of a Fast Payment System + CAPS
• Critical failure: On 14 January 2022 the system went offline for ~56 days due to expired Hyperledger Fabric identity certificates — a certificate-management failure that became a globally cited cautionary case study
• Outcome: DCash pilot concluded January 2024; DCash 2.0 suspended February 2026 (Communiqué of the 112th Meeting, 4 May 2026)

Sweden (e-Krona)

• Platform: R3 Corda DLT pilot platform (concluded 2024), developed with Accenture; EMV-compatible payment card for offline
• Ledger model: DLT-based pilot on R3 Corda (digital tokens). No issuance decision has been taken; no enabling legislation exists
• Offline capability: Explored in Phase 4 (2023-2024) via a payment card ("shadow wallet"); offline wallet cap SEK 3,000, max 5 consecutive offline payments before reconnection; technically feasible but needs substantial further development
• Status: Longest-running CBDC research programme globally (since 2017); four pilot phases concluded 2024; trajectory now contingent on the ECB digital euro

Singapore (wholesale SGD)

• Platform: SGD Testnet (DBS, OCBC, Standard Chartered, UOB); GL1 public permissioned blockchain (core group BNY Mellon, Citi, JPM, MUFG, SocGen-FORGE); Purpose-Bound Money protocol (form-factor agnostic)
• Ledger model: DLT — SGD Testnet shared ledger for wholesale CBDC; Orchid Blueprint building blocks (Settlement Ledger, Tokenisation Bridge, Programmability Protocol/PBM, Name Service)
• Offline capability: NOT VERIFIED / not a stated feature (wholesale-first; retail via PBM on bank deposits)
• Status: First live SGD wholesale CBDC issuance on SGD Testnet (2024-2025); first live wholesale transaction (overnight interbank lending) November 2025. Retail (Project Orchid) remains research/infrastructure only

Thailand (Inthanon / retail pilot)

• Platform: Dedicated mobile app for the retail pilot (Krungsri, SCB, 2C2P); PromptPay (QR) integration tested; Programmable Payment Sandbox (case-by-case testing, segregated 1:1 reserves)
• Ledger model: DLT-based — Project Inthanon (wholesale, DLT); the retail pilot used a DLT-based system; the Programmable Payment Sandbox uses DLT and smart contracts with Thai Baht-backed digital units at a fixed 1:1 peg
• Offline capability: Not publicly specified for the retail CBDC pilot; PromptPay considered the most suitable channel
• Status: Retail CBDC pilot (Bang Khun Phrom) concluded 2024 with no plans for full-scale retail issuance; pivot to a Programmable Payment Sandbox for stablecoin experimentation; world-class wholesale work (mBridge founding member, Inthanon-LionRock)

Australia (Project Acacia)

• Platform: Pilot wCBDC issued onto AP+ HashSphere (private Hedera), Canvas Connect, and Redbelly Network
• Ledger model: Multi-DLT (Canvas Connect private, Ethereum public, Hedera public+private, Redbelly public, XRP Ledger public); ERC-20 standard on EVM platforms with RBA mint/burn/pause/allow controls
• Offline capability: NOT VERIFIED for wholesale; the RBA identified offline payment as a potential retail resilience use case
• Status: No retail CBDC (assessed, no public interest case). Project Acacia wholesale experimentation concluded February 2026 (20 use cases: 12 pilot + 8 PoC); final report 18 May 2026 — A$4.4m pilot wCBDC issued across 8 use cases (research issuance, not a live launch)

Key characteristics of Model B:

• Distributed validation adds resilience through redundancy across nodes
• DLT consensus mechanisms can impose throughput constraints at scale
• Vendor dependency is significant — the DCash outage illustrates the risk of relying on a single vendor's blockchain implementation
• Privacy-on-chain is technically challenging (see Brazil's experience below)
• Certificate and key management become critical operational risks

Model C: Hybrid / DLT Pivot to Centralized

Countries:

Brazil (Drex / Hyperledger Besu pivot),

India (wholesale DLT + retail API),

South Korea (Project Hangang two-layer)

These projects began with or draw from DLT concepts but combine DLT and non-DLT layers, or pivoted toward centralized designs after encountering scalability, privacy, or regulatory constraints.

Brazil (Drex)

• Platform: Hyperledger Besu (Phase 1, decommissioned 10 November 2025); the project is rebuilding "from scratch" with technology yet to be selected. Privacy solutions tested — Anonymous Zether (JPMorgan), Starlight (EY), Rayls (Parfin), ZKP Nova (Microsoft), Hamsa — none deemed production-mature
• Ledger model: Was permissioned DLT (Hyperledger Besu, EVM-compatible) in Phase 1; post-Nov-2025 pivot to centralized infrastructure, with blockchain no longer guaranteed; technology TBD
• The "privacy trilemma": The BCB concluded that privacy, decentralization, and programmability cannot all be maximized simultaneously with current technology. ZKPs that protected user data also prevented the BCB from performing monetary controls and KYC/AML — privacy was the single most consequential obstacle and drove the pivot
• Status: Active pilot but restarted/de-scoped; MP 1,303/2025 lapsed 8 October 2025; Phase 3 (collateral/lien reconciliation) to begin 2026; no firm launch date (Galipolo: "blockchain tested did not prove viable")

India (e₹)

• Platform: Token-based retail (e₹-R) wallets; account-based wholesale (e₹-W) on DLT; NPCI manages the backend; UPI QR interoperability (HDFC, Axis, Kotak, SBI)
• Ledger model: Hybrid — wholesale (Tier 1) on DLT (Hyperledger Fabric); the retail layer is intentionally NOT on DLT — API-based, inspired by UPI architecture for scalability
• Offline capability: Yes — officially launched at Global Fintech Fest October 2025; NFC tap-and-pay, telecom-assisted/feature-phone payments, stored-value offline wallets
• Status: Active retail (e₹-R) and wholesale (e₹-W) pilots; e₹-R circulation ₹771.7 crore (~US$90M) at 31 March 2026, down ~24% YoY (first reported contraction; self-reported, RBI AR 2025-26 — flag)

South Korea (Project Hangang)

• Platform: Bank mobile apps with QR; LG CNS core infrastructure partner; Samsung (earlier phases); biometric (fingerprint) auth, auto top-up, P2P transfers
• Ledger model: Two-layer hybrid — a wholesale CBDC layer (BOK issues to banks against 100% collateral, real-time interbank settlement) plus a deposit-token layer (banks issue won-pegged tokens convertible 1:1). DLT in cloud testing; node management centralized at BOK (efficiency dropped with more nodes/users)
• Offline capability: Offline payments tested in a Phase 2 simulation (2022); not emphasized as a core deployed feature
• Status: BOK describes it as "an intermediate stage between a CBDC and stablecoins" — not a full retail CBDC. Phase 1 paused June 2025; Phase 2 restarted March 2026 with a government-subsidy focus

Key characteristics of Model C:

• Draws on DLT design concepts (UTXO, tokenization, deposit tokens) while either retaining or pivoting toward centralized control
• Avoids DLT consensus overhead on the highest-throughput layer (India's API-based retail; Brazil's centralized rebuild)
• Privacy can be engineered architecturally rather than relying solely on zero-knowledge proofs
• Pragmatic approach: adopt what works from DLT innovation without the operational complexity

Model D: Platform Model with Open API / Technology-Agnostic Core

Countries:

United Kingdom (digital pound),

Japan (digital yen — shared account ledger)

These designs emphasize a central bank-operated platform with a standardized API layer and an explicitly undecided underlying technology, enabling competitive private-sector innovation.

United Kingdom (digital pound)

• Platform: BoE core ledger + API layer + analytics hub + alias service; tested via Project Rosalind (33 API functions) and the Digital Pound Lab (sandbox, August 2025)
• Ledger model: Undecided — the "core ledger" concept; explicitly no decision on the underlying technology (DLT vs centralized database); funds held on the BoE core ledger, not commercial bank balance sheets
• Offline capability: Deferred offline payments prioritised for launch (funds on the ledger, settle on reconnection); device offline payments deferred (benefits uncertain, risks unproven at scale)
• Status: Design phase; no launch decision taken; decision on whether to proceed expected later in 2026

Japan (digital yen)

• Platform: Experimental pilot system; API sandbox (18 companies as of Oct 2024); 50,000 TPS achieved vs a 100,000 TPS target
• Ledger model: Account-type CBDC ledger shared between central and intermediary systems; tested centralized account-based, centralized token-based, and shared account-based designs; NoSQL databases; researching UTXO models; public blockchain deemed not currently viable; exploring a controlled Layer 2
• Offline capability: Under development — for disaster resilience (earthquakes, typhoons) and poor connectivity; multiple patterns evaluated (online/offline payer/receiver terminals)
• Status: No CBDC issued; the BOJ maintains "no plan to issue a CBDC"; a 2026 retail-CBDC decision is pending. February 2026 consolidation expanded research to wholesale CBDC and tokenization

Key characteristics of Model D:

• Technology-agnostic core — the platform model works regardless of whether the ledger is DLT or centralized
• Maximizes private-sector innovation through competitive API access
• Separates concerns: the central bank handles money issuance and settlement; the private sector handles user experience and value-added services
• Requires robust API governance and standardization
• Untested at production scale

Research-Only / Anti-CBDC

Four jurisdictions have no deployed system and no design-stage platform:

• United States (anti-CBDC): No deployed system. Project Hamilton (Boston Fed/MIT, OpenCBDC, MIT license) found that a distributed ledger was NOT needed — a central-actor trust model suited the design — and reached 1.7M TPS in research; RLN proof-of-concept and Project Cedar (NY Fed) for wholesale. Any potential US CBDC would be intermediated (per the 2022 "Money and Payments" paper), but EO 14178 bars all federal CBDC activity (statutory moratorium passed both chambers in differing forms, not enacted).
• Canada (research, scaled down): No deployed system. Research evaluated OpenCBDC 2PC (two-tier) for retail; Project Samara used Hyperledger Fabric (separate bond/cash ledgers) for wholesale (CAD $100M tokenized bond settled 5 March 2026). Retail work scaled down September 2024; the Stablecoin Act (Bill C-15) received Royal Assent 26 March 2026 but is not yet in force and excludes CBDCs.
• Ghana (pilot completed, design phase): Token-based offline component (stored-value, cash-like) using a Giesecke+Devrient prototype; two-tier model; offline-first design (offline NFC POS, consecutive offline payments, dual-mode tested in the 2022 pilot). The BoG declared the pilot "completed" on 22 May 2026 and is now designing wholesale/cross-border use; the retail eCedi remains unlaunched and enabling legislation is unpassed.
• South Africa (research/PoC): PoC/sandbox only — no production ledger. Khokha 1 used Quorum (Ethereum-based permissioned, IBFT consensus, Pedersen commitments + range proofs); Project Khokha 2 used a wCBDC Zone on R3 Corda + a Khokha Hub on a custom Cosmos-based ledger with a bidirectional bridge; Project Dunbar on R3 Corda + Partior/Quorum. SARB's November 2025 Position Paper found a retail CBDC technically feasible but with "no compelling immediate need."

2. Key Design Decisions Table — All 20 Jurisdictions

The table below mirrors the four technology-design facts (ledger model, intermediated/two-tier, offline capability, platform) for every tracked jurisdiction.

Notes:

• TPS figures are rarely published by central banks. China's 300,000+ claim is widely cited but NOT VERIFIED against primary documentation; the EU and UK have not published throughput targets; Japan achieved 50,000 TPS against a 100,000 target; Nigeria stress-tested >2,000 TPS; the US Project Hamilton reached 1.7M TPS in research only.
• "Intermediated (two-tier)" means commercial banks or licensed PSPs act as intermediaries between the central bank and end users. Every one of the 20 jurisdictions uses (or, for research/design-stage projects, anticipates) an intermediated two-tier model — no jurisdiction issues, or plans to issue, directly to retail users without intermediaries.
• "Offline?" reflects deployed or tested offline capability for the relevant CBDC; "NOT VERIFIED" means the matrix records no offline design fact for that jurisdiction.
• All adoption and transaction figures are self-reported by the issuing central banks unless otherwise verified; see the per-jurisdiction profiles.

3. Cross-Country Aggregates (N = 20)

Recomputed from the fact matrix for all 20 jurisdictions:

3.1 Ledger model

• Centralized ledger (no production DLT): 4 — Jamaica (explicitly not DLT), EU (centralized N-XT/UTXO), Russia (centralized hybrid), Japan (account-type shared ledger, public blockchain deemed not viable).
• Account-based on permissioned DLT, central bank sole validator (effectively centrally controlled): 1 — Nigeria (Hyperledger Fabric).
• DLT-based (core settlement on distributed ledger): 7 — Bahamas (Cortex), Eastern Caribbean (Hyperledger Fabric, suspended), Sweden (R3 Corda), Singapore (SGD Testnet/GL1), Thailand (Inthanon), Australia (multi-DLT), South Africa (Khokha PoCs, Quorum/Corda/Cosmos).
• Hybrid (DLT + non-DLT layers, or DLT pivoting to centralized): 4 — China (hybrid centralized/distributed), India (wholesale DLT + retail API), South Korea (wholesale CBDC + deposit tokens), Brazil (was Besu DLT, pivoting to centralized).
• Undecided / research with no production ledger: 4 — US (Project Hamilton found DLT not needed), Canada (research), UK (explicitly undecided), Ghana (token-based offline component, pre-launch).

(Counts overlap by design where a jurisdiction's classification is genuinely a judgment call — e.g. the EU borrows the UTXO data model from DLT but settles centrally; Nigeria runs on Hyperledger Fabric but is centrally validated and account-based. The matrix language, not the surface technology label, governs the placement.)

3.2 Intermediated (two-tier) distribution

• All 20 of 20 use or anticipate intermediated two-tier distribution. No jurisdiction in this domain issues — or plans to issue — directly to retail users. This is the single most universal CBDC design choice.

3.3 Offline capability

• Deployed or operationally targeted offline (Yes): 5 — China (deployed, dual offline), India (launched Oct 2025), Bahamas (under development), EU (core design feature), Ghana (offline-first design).
• Tested / explored but not a core deployed feature: 3 — Sweden (Phase 4 card), South Korea (2022 simulation), UK (deferred offline prioritised; device offline deferred).
• Under development: 1 — Japan.
• Research only: 1 — Canada.
• Explicitly No / required connectivity: 2 — Nigeria (required internet), Eastern Caribbean (DCash 1.0 had none; 2.0 required it but is suspended).
• NOT VERIFIED / not applicable (no offline fact in matrix): 8 — US, Brazil, Jamaica, Australia (wholesale), Singapore, Thailand, Russia, South Africa.

3.4 Launch / deployment status (technology relevance)

• Launched / live: 6 — Bahamas (Sand Dollar), Nigeria (eNaira), Jamaica (JAM-DEX), Eastern Caribbean (DCash, pilot concluded / successor suspended), China (e-CNY, pilot-scale), India (e₹, pilot). Only these six have production technology in the field.
• Active pilot / live wholesale pilot: Singapore (live wholesale), South Korea (Project Hangang), Brazil (restarted), Russia (narrow live govt use, large-scale launch 1 Sep 2026), Australia (completed wholesale experimentation), Thailand (completed retail pilot).
• Design / research / anti-CBDC (no production technology): EU (advanced preparation), UK (design), Japan (research), Sweden (research, pilot concluded), Canada (scaled-down research), Ghana (pilot completed, design phase), South Africa (PoC), US (anti-CBDC).

4. What Works and What Doesn't (Technology Lens)

What Works

1. Two-tier distribution is the proven and universal model. All 20 jurisdictions use or anticipate intermediated two-tier distribution. China's 2.25 billion wallets across 22 authorized operators (self-reported) and India's distribution through ~19 banks demonstrate the model's scalability. The BIS foundational principles (2020) codified this as a core design requirement, and no jurisdiction has departed from it.

2. Centralized or hybrid ledgers dominate the highest-throughput layers. The jurisdictions handling the most volume — China (hybrid centralized), India (retail API, intentionally not DLT) — deliberately avoid DLT consensus on the retail layer. The EU's N-XT borrows the UTXO data model from DLT but settles centrally. Even the US Project Hamilton concluded a distributed ledger was not needed.

3. Offline capability is achievable and is the clearest retail differentiator. China's dual-offline (NFC/Bluetooth/SIM/hardware), India's October 2025 offline launch, Ghana's offline-first design, and the EU's secure-element device-to-device approach show offline is technically deliverable. In jurisdictions with strong existing fast payments, offline resilience is often the only compelling reason to build a retail CBDC.

4. Multi-vendor and EU-controlled procurement reduces dependency risk. The EU's competitive procurement of five external components from EU-controlled providers (with core settlement built in-house by six NCBs) is the most robust vendor strategy observed — a direct response to the single-vendor failure mode below.

What Doesn't Work

1. Single-vendor DLT dependency: the Eastern Caribbean DCash outage. The ~56-day DCash outage (January-March 2022), caused by expired Hyperledger Fabric identity certificates on Bitt Inc.'s platform, is the most consequential CBDC infrastructure failure to date. It destroyed user confidence and contributed to the pilot's closure and DCash 2.0's suspension (February 2026). The lesson: multi-vendor strategies and robust certificate lifecycle management are non-negotiable.

2. Over-ambitious DLT without a privacy resolution: Brazil's trilemma and pivot. Years of testing on Hyperledger Besu could not produce a privacy solution satisfying both LGPD/banking secrecy and the BCB's regulatory-visibility needs. Brazil decommissioned the Besu platform on 10 November 2025 and is rebuilding "from scratch" — the most significant public repudiation of DLT for CBDC by a major central bank. Resolve privacy architecture before committing to a platform.

3. Account-based centralized ledger without offline or adoption strategy: Nigeria's eNaira. eNaira required internet connectivity (no offline), and 98.5% of wallets were inactive after one year (IMF, self-reported). Technical readiness on Hyperledger Fabric did not translate into adoption; the CBN has pivoted to a wholesale focus. Lesson: a centrally-controlled ledger and high TPS headroom mean nothing without offline access and an adoption strategy.

4. Fast payment systems may obviate retail CBDC entirely. The ECCB's suspension of DCash 2.0 in favor of a Fast Payment System + CAPS, Thailand's pivot from retail CBDC to a Programmable Payment Sandbox (PromptPay deemed sufficient), and South Africa's "no compelling immediate need" Position Paper all show that jurisdictions with functional fast payments may not need a retail CBDC. The technology must offer something existing rails do not.

5. Recommended Framework (Technology & Design)

5.1 Architecture decision tree

Step 1 — Primary objective:

• Payment resilience / cash decline: prioritize offline capability and hardware wallets (Model A or D; Ghana/EU/China offline patterns).
• Financial inclusion: tiered KYC, low-barrier wallets, offline NFC (any model with offline).
• Programmable finance / tokenized assets: evaluate DLT-based or hybrid (Model B/C), but resolve privacy first (Brazil's lesson).
• Cross-border / wholesale settlement: DLT shared ledger for institutional participants (Singapore GL1, Thailand mBridge, South Africa Khokha).

Step 2 — Existing digital payment landscape:

• Strong fast payment system (India UPI, Brazil Pix, Thailand PromptPay, EU SEPA Instant): retail CBDC must offer distinct value — programmability, offline resilience, or privacy.
• Weak infrastructure: retail CBDC may serve as foundational payment infrastructure.

Step 3 — Population scale:

• Large (>100M): DLT consensus may not scale on the retail layer; centralized or hybrid recommended (China, India, EU).
• Small / island state: DLT-based with offline is feasible but single-vendor dependency risk is higher (DCash lesson).
• Multi-jurisdictional: harmonized governance required (DCash showed the complexity).

Step 4 — Privacy legal framework:

• Strong data-protection regime (GDPR/LGPD/POPIA): resolve privacy architecture before deployment; the EU's UTXO-based privacy-by-design is the reference model.

5.2 Minimum technical requirements

5.3 Vendor strategy

Recommendation: The EU's approach — competitive procurement of five external components from EU-controlled entities, with core settlement developed internally by six national central banks — is the most robust vendor strategy: maximum sovereignty over critical infrastructure, private-sector expertise for non-core components, and no single point of failure of the DCash kind.

Wikilinks

Country profiles (all 20):

United States,

Bahamas,

Brazil,

Canada, Eastern Caribbean,

Jamaica,

Australia,

China,

India,

Japan,

Singapore,

South Korea,

Thailand,

EU,

Russia,

Sweden,

UK,

Ghana,

Nigeria,

South Africa

International frameworks: BIS

Cross-cutting dimensions: Privacy & Data Protection, Cross-Border Interoperability, Financial Inclusion

Verification Notes

1. China's 300,000+ TPS claim: NOT VERIFIED against primary PBOC technical documentation; widely cited in secondary sources only. From 1 January 2026 the model combines accounts, token strings, and smart contracts; 22 authorized operators as of 2 April 2026 (full operator list NOT fully verified).
2. EU digital euro throughput: NOT PUBLISHED by the ECB as of June 2026. Ledger is centralised (N-XT/UTXO), not DLT. Preparation phase concluded October 2025; ECON Committee vote scheduled 23 June 2026.
3. UK digital pound technology choice: explicitly undecided — the BoE has confirmed no decision on DLT vs centralized database. Design phase; decision expected later in 2026.
4. Brazil Drex ledger: Hyperledger Besu platform decommissioned 10 November 2025; rebuilding "from scratch" with technology TBD. Offline capability NOT VERIFIED.
5. Russia digital ruble DLT platform: the specific DLT (if any) is NOT disclosed by the CBR. Offline capability NOT VERIFIED. Narrow live govt/budget use since 1 January 2026; large-scale launch planned 1 September 2026.
6. Jamaica JAM-DEX: NOT blockchain/DLT — a centralized BOJ ledger (eCurrency Mint, DSC3). Offline capability NOT VERIFIED.
7. India retail e₹ TPS capacity: NOT PUBLISHED by the RBI. Retail layer is intentionally NOT DLT (API-based). e₹-R circulation ₹771.7 crore at 31 March 2026, ~24% YoY decline (self-reported, RBI AR 2025-26).
8. Nigeria eNaira: account-based on Hyperledger Fabric with the CBN as sole validator; no offline (required connectivity); stress-tested >2,000 TPS. Adoption figures self-reported (IMF/CBN).
9. Singapore, Australia, Thailand, South Africa offline: NOT VERIFIED / not a stated feature in the matrix for these (wholesale-focused or pilot-stage) programmes.
10. US Project Hamilton (1.7M TPS) and Canada OpenCBDC: research figures only; neither jurisdiction has a production CBDC. The US is anti-CBDC (EO 14178 in force; statutory moratorium passed both chambers in differing forms, not enacted).
11. All adoption / transaction figures: self-reported by issuing central banks unless independently verified; flagged accordingly in the per-jurisdiction profiles.

Last updated: 2026-06-04