Skip to main content

Lightspark and Visa Bring Self-Custodial Bitcoin and Stablecoin Debit Cards to 100+ Countries

· 11 min read
Dora Noda
Dora Noda
Software Engineer

For most of the last decade, "spending crypto in the real world" meant handing your coins to an exchange, waiting for them to issue you a Visa or Mastercard, and accepting that the spending balance was no longer yours in any meaningful sense. The Coinbase Card, the Crypto.com card, the BVNK-powered programs — all of them solved the merchant-acceptance problem by re-introducing a custodian.

That model just cracked.

On April 29, 2026, Lightspark and Visa announced a partnership to issue stablecoin- and Bitcoin-backed Visa debit cards across 100+ countries, plugged directly into Lightspark's Grid platform. The same week, Lightspark's Grid Global Accounts launched at Bitcoin 2026 Las Vegas, and a new wave of issuers — including a self-custodial multi-asset wallet called Avvio — began onboarding to the rails. The pitch is blunt: a Visa swipe at any of 175 million merchants, funded by a balance the user actually holds the keys to.

If the architecture sticks, this is the first global Visa product where "your card, your coins" stops being a slogan and starts being a default.

What Lightspark and Visa Actually Shipped​

The headline number is 100+ countries, but the more important detail is what Grid is. Lightspark Grid is an API platform that lets any fintech, neobank, or app behave like a global financial institution without becoming one. Through a single integration, a partner can offer:

  • Branded dollar accounts backed by stablecoins
  • Visa debit cards, virtual and physical, that swipe at 175 million merchants in 33 countries at launch
  • Real-time payouts to bank accounts and mobile money providers in 65+ countries, across 14,000 banks
  • Instant Bitcoin/fiat conversion routed over Lightning or the new Spark protocol
  • Stablecoin support including USDC on Solana, Base, and Spark

According to Lightspark, the network as configured already reaches roughly 5.6 billion people across an aggregated $93 trillion in GDP. The first phase rolls out in the United States and Europe, with planned expansion into Asia Pacific, Africa, and the Middle East later in 2026.

For Visa, this is a continuation of a clear 2025–2026 strategy. The card network now captures more than 90% of on-chain card volume through partnerships with crypto-native infrastructure providers, and its on-chain stablecoin settlement for issuers reached an estimated $3.5 billion annual run rate by late 2025. Lightspark gives Visa something it didn't have before: a partner whose entire stack is built around Bitcoin and Lightning settlement, not just stablecoins.

The Avvio Wedge: Self-Custody as a Product, Not a Compromise​

The Lightspark–Visa announcement on its own would already be a big payments story. What pushes it into "architectural shift" territory is the type of issuer now showing up on Grid.

Avvio is one of the first card-issuing wallets to launch on the Lightspark+Visa stack as an explicitly self-custodial, multi-asset product. The pitch is unusually direct for a consumer payments app: real USD and EUR accounts, payouts into 120 countries, and a spending balance collateralized by self-custodial Bitcoin, gold, and tokenized-stock exposure. The wallet keys never leave the user's device, and the Visa rail sits on top.

This matters because every prior attempt at a "real" crypto debit card has eventually hit one of two walls:

  1. Custodial issuers (Coinbase Card, Crypto.com Card, the original BVNK pilots) had to take ownership of user funds to authorize merchant pulls in real time. Convenient — but the user is back to trusting an intermediary, with all the failure modes that implies.
  2. Pseudo-self-custodial wrappers typically required moving funds into a centralized intermediate balance the moment you swiped. Self-custodial in marketing copy, custodial at the moment of truth.

A Lightspark+Visa+Avvio-style stack threads the needle by separating roles. The user holds the keys. The wallet authorizes a draw against a verified balance. Lightspark Grid handles the conversion and settlement to Visa in real time over Lightning or Spark. The merchant gets dollars. Visa gets a clearing event. Nobody in the chain ever needed sole custody of the asset.

That is a meaningfully different security model from anything that has shipped at this scale before.

How This Stacks Up Against BVNK, MoonPay, and Coinbase​

To understand how big a shift this is, it helps to look at where the other three contenders sit in May 2026:

  • BVNK + Visa Direct (2025–2026): BVNK's stablecoin payments infrastructure powered Visa Direct payouts to issuers in select markets, handling roughly $30 billion in annual stablecoin volume. The model was issuer-locked and operated through custodied balances. In a notable plot twist, Mastercard acquired BVNK for around $1.8 billion in March 2026, effectively migrating that infrastructure off Visa's roadmap.
  • MoonPay MoonAgents Card (May 1, 2026): MoonPay launched a stablecoin debit card aimed at AI agents and consumers, on the Mastercard network via Monavate. It links a self-custodial wallet to a virtual Mastercard, with revocable approvals and no transfer of custody at issuance. It is genuinely closer to self-custodial than older custodial-card products, but it lives on Mastercard rails and on a single chain.
  • Coinbase Cards and Base App: Coinbase still operates one of the most widely held crypto cards in the U.S., funded from the centralized exchange wallet. The Base App, launched as a self-custodial consumer wallet, points in the same direction as Avvio — but Coinbase has not yet plugged Base directly into a Visa-issuing path that bypasses the exchange custody layer.

Stack those four side by side and a clear pattern emerges. Mastercard's bet is on acquiring custodial stablecoin infrastructure (BVNK) and licensing it to AI-agent and fintech use cases (MoonAgents). Visa's bet, via Lightspark, is on building a programmable global rail where the issuer can be self-custodial by default. They are not the same architecture, and within 12–18 months one of them is going to start to look obviously correct.

The Numbers Behind the Inflection​

The market context makes the timing less surprising. The total stablecoin market capitalization crossed $317 billion in early 2026, with USDT at roughly $187 billion and USDC at around $75.7 billion — and USDC growing 73% year over year, faster than USDT for the second consecutive year. Crypto card spending hit an $18 billion annualized run rate by January 2026 as everyday payments shifted on-chain. Some analysts now project stablecoins to settle more than $50 trillion in transactions during 2026, a figure that would put on-chain dollar transfers comfortably ahead of the legacy card networks on raw transfer volume.

What was missing from those numbers was a credible self-custodial spending experience at global scale. Card programs were either niche, custodial, or both. The Lightspark+Visa launch is the first piece of infrastructure that lets that $317 billion of dollar-pegged tokens, plus Bitcoin, plus tokenized assets like gold and equities, become spendable in 100+ countries without forcing the user to hand over the keys.

It also reframes the agent economy story. MoonPay positioned MoonAgents around AI agents that need to spend. Lightspark and Avvio are quietly building the same capability for humans first, with agent-callable controls bolted on top via Grid's "agent permissions" layer. Both groups are converging on the same insight: the spending experience and the custody decision should be decoupled.

What This Means for Web3 Infrastructure​

For builders sitting one layer below the card network, the Lightspark+Visa launch reshapes demand in three concrete ways:

1. Continuous balance attestation becomes the new hot path. A self-custodial card has to verify "the user has X dollars of spendable balance" in milliseconds, every swipe, often across multiple chains and assets. That is not a one-shot RPC pattern. It looks much more like a high-QPS read workload — eth_call, getBalance, oracle lookups, and Lightning channel state — sustained 24/7 against millions of wallets. RPC providers are about to feel this.

2. Multi-asset price feeds move from analytics to settlement-critical. When the spending balance is collateralized by BTC, gold, USDC, and tokenized stocks at once, the price feed that values that basket is no longer a UX detail. It is part of the authorization flow. Latency, freshness guarantees, and feed redundancy become payments-grade requirements rather than dashboard features.

3. Lightning/Spark settlement attestation becomes a queryable surface. For Bitcoin-backed swipes, the issuer needs to prove that a Lightning payment cleared, that a Spark transfer is finalized, and that a USDC swap settled — all in time to authorize the Visa transaction. Every one of those is a new RPC pattern that today's Ethereum-shaped infrastructure was not designed for.

The shape of all this is different from how centralized exchange wallets generated load. Exchange wallets concentrated traffic at a few endpoints. Self-custodial spending wallets fan out load across millions of independently keyed addresses, each polling for balances, each requiring its own authorization checks, each potentially live across multiple chains.

What to Watch Next​

Three open questions will decide whether this becomes the new template or a well-funded experiment:

  • Does MiCA and the GENIUS Act compliance overhead force self-custodial issuers like Avvio back behind a custodian for licensing reasons in Europe and the U.S.? The technical architecture is ready. The regulatory architecture for self-custodial card programs is genuinely unclear.
  • Does Mastercard counter with a self-custodial Visa-style stack of its own, or double down on the BVNK-MoonPay custodial agent thesis? The two networks are now visibly diverging on architecture for the first time in years.
  • Do other issuers — BVNK successors, Bridge, regulated neobanks — follow Avvio onto Grid, or do they wait for the regulatory dust to settle? The first 90 days of issuer onboarding will be telling.

Either way, the era where "spending Bitcoin" required surrendering Bitcoin is ending. The infrastructure to keep the keys and swipe the card now exists, in 100+ countries, on the world's largest card network.

BlockEden.xyz provides enterprise-grade RPC and indexing infrastructure for the chains powering this new self-custodial payments stack — including Solana, Base, and the Bitcoin-adjacent Lightning ecosystem. If you're building wallets, card programs, or agent-callable financial services on top of this architecture, explore our API marketplace to ship on rails designed for the workload.

Sources​

Share on Twitter