ZKsync Goes Enterprise: Why 5 Banks Chose ZK Tech for a $600B Tokenized Deposit Network

Ethereum's Layer 2 ecosystem just crossed a threshold that most observers did not see coming. ZKsync creator Matter Labs laid off a portion of its workforce in mid-June, announced a full strategic pivot to enterprise infrastructure, and is now the settlement layer behind a tokenized deposit network spanning six U.S. regional banks with over $600 billion in combined deposits. The open rollup that once competed with Arbitrum and Optimism for retail DeFi users has gone institutional — and it may be the smartest bet any L2 team has made in 2026.

Here is what happened, why five of America's largest regional banks chose ZK technology, and what the pivot means for developers building on Ethereum's scaling ecosystem.

The Pivot: From Public Rollup to Permissioned Privacy Chain

On June 17, Matter Labs CEO Alex Gluchowski confirmed that the company was restructuring around Prividium — a permissioned, privacy-preserving Layer 2 built on zero-knowledge proofs and designed exclusively for regulated financial institutions. The restructuring included staff cuts, which Gluchowski framed as a skills realignment: building infrastructure for banks requires a fundamentally different engineering profile than operating a public rollup serving retail users and DeFi protocols.

Prividium is not a side project. Matter Labs completed a SOC 2 Type I security examination for the platform this month — a certification that is table stakes for any software vendor hoping to serve U.S. financial institutions. No other Ethereum L2 has achieved this level of compliance certification. The company is also actively pursuing additional regulatory credentials required by institutional clients before onboarding enterprise software.

The pivot has drawn scrutiny from parts of the crypto community. Matter Labs raised substantial venture capital to build ZKsync as a public good — a permissionless scaling solution for Ethereum. Critics argue that the transition to a gated, bank-only chain represents a departure from those original commitments. The company has not yet clarified what role, if any, the ZK token will play in the Prividium architecture, leaving token holders in an information vacuum.

The Cari Network: Six Banks, $600 Billion, One Shared Ledger

The most concrete evidence that the pivot is working arrived in S&P Global's Q1 2026 stablecoin report. Six large U.S. regional banks — First Horizon, Huntington Bancshares, KeyCorp, M&T Bank, Old National Bancorp, and SouthState — have formed the bank-governed Cari Network, a shared tokenized deposit platform built on a private deployment of ZKsync's technology.

Collectively, these institutions hold over $600 billion in deposits. The Cari Network enables instant, 24/7 interbank settlement using tokenized deposits — effectively digital dollars that move between participating banks at the speed of a blockchain transaction rather than the two-to-three business days of the traditional ACH system. Each bank governs the network jointly through a consortium model, meaning no single institution controls the infrastructure.

The S&P Global report frames the Cari Network alongside proprietary efforts from BNY Mellon and JPMorgan, but the consortium approach is notably different. Rather than each bank building its own walled garden, the Cari Network is shared infrastructure — a common settlement rail that any qualified institution can join. This is the model that the GENIUS Act, which pulls stablecoin issuance onshore to regulated U.S. entities, seems designed to encourage.

Why Zero-Knowledge Won the Deal

Banks did not choose ZKsync because it was the fastest or cheapest L2. They chose it because zero-knowledge proofs solve a specific problem that no other scaling technology addresses: how to verify that a transaction is valid without revealing the underlying data to anyone who should not see it.

In a traditional blockchain, every node sees every transaction. That is a non-starter for banks, which are legally obligated to protect customer information and counterparty details. A ZK proof allows Bank A to cryptographically prove to Bank B that a transfer settled correctly — without exposing the identities, amounts, or account details of any other participant on the network. The privacy is mathematical, not policy-based. It cannot be circumvented by a rogue node operator or a database breach.

This property — called selective disclosure — is why ZK technology has been described as the 'holy grail' for institutional blockchain adoption since at least 2022. What changed in 2026 is that the technology finally crossed the compliance threshold. Prividium's SOC 2 certification, combined with ZKsync's architectural maturity after three years of mainnet operation, made the technology defensible in a bank boardroom in a way it was not two years ago.

Private Atomic DvP: Settlement That Cannot Fail Halfway

On June 22, ZKsync launched Private Atomic Delivery versus Payment (DvP) — a mechanism that allows two independently governed zones on the Elastic Network to settle a trade as a single trustless event, recorded on Ethereum. In plain terms: delivery of an asset and the corresponding payment either both happen, or neither does. There is no intermediate state where one party has the asset and the other has the money.

For banks, this is transformative. In the legacy financial system, settlement risk — the risk that one side of a trade fails after the other side has already delivered — is managed through clearinghouses, collateral requirements, and multi-day settlement cycles. Atomic DvP eliminates that risk at the protocol level. A $100 million bond trade settles with the same finality guarantees whether it occurs at 2:00 PM on a Tuesday or 3:00 AM on a Sunday.

The private variant is significant because it allows each bank to operate its own zone with its own governance rules, compliance policies, and access controls, while still settling atomically with counterparties in other zones. Banks get the operational isolation they require from a compliance perspective without sacrificing the interoperability that makes a shared settlement network valuable in the first place.

The Great L2 Bifurcation Is Here

ZKsync's pivot is not happening in isolation. It is the sharpest expression of a trend that has been building across the Ethereum L2 landscape throughout 2026: the separation of Layer 2s into two distinct product categories.

On one side, general-purpose public rollups — Arbitrum, Optimism, Base — continue to compete for retail users, DeFi protocols, and on-chain applications. Arbitrum had an extraordinary June in its own right, partnering with Mastercard for stablecoin settlement, landing LG Electronics as a custom L2 client for programmatic advertising across 216 million smart TVs, and reaching $341.9 million in tokenized real-world assets. Optimism activated its Isthmus hardfork in mid-June, bringing Ethereum's Pectra upgrade features to the Superchain, and launched an enterprise-grade managed chain product called OP Enterprise Fully Managed.

On the other side, permissioned institutional chains — ZKsync's Prividium, and to a lesser extent the enterprise variants of Arbitrum and Optimism — are targeting a fundamentally different customer: regulated financial institutions that need privacy, compliance certifications, and operational control, not just high throughput and low fees.

Vitalik Buterin acknowledged this bifurcation in a June essay, noting that many general-purpose L2s 'no longer have a reason to exist' as standalone chains. The ones that survive, he argued, will either capture a distinct user base or provide specialized infrastructure that general-purpose rollups cannot. ZKsync appears to have taken that advice literally.

What Developers Should Watch

For smart contract developers and web3 builders, the ZKsync pivot carries several implications that are worth tracking over the next six to twelve months.

First, the ZK token's utility model is now an open question. ZKsync was originally positioned as a public L2 where the ZK token would govern the protocol and capture value from retail and DeFi activity. If Matter Labs is now fully focused on permissioned infrastructure where banks — not retail users — are the customers, the demand dynamics for ZK change materially. Developers building applications that integrate with ZK or rely on its tokenomics should monitor official communications closely.

Second, the institutional ZK stack is becoming a distinct developer surface. Prividium's private DvP, its zone-based architecture, and its compliance tooling represent a new category of blockchain infrastructure that requires a different skill set — one that blends zero-knowledge cryptography with an understanding of regulatory requirements. Developers who can bridge that gap will be in high demand.

Third, the Cari Network is a real-world benchmark for tokenized deposits at scale. If six regional banks can successfully operate a shared blockchain settlement network, it validates the entire thesis behind on-chain finance and will accelerate adoption across the banking sector. If it stumbles — on security, governance, or throughput — it will set the institutional narrative back by years.

The Pragmatic Bet

Matter Labs' decision to step back from the public L2 race and bet the company on enterprise infrastructure is the clearest strategic signal the Ethereum scaling ecosystem has produced in years. The company looked at the landscape — dozens of general-purpose rollups competing for a finite pool of retail users and DeFi liquidity — and concluded that the bigger opportunity was building the plumbing that traditional finance will actually use.

Whether that bet pays off depends on execution at a level of reliability that crypto companies have historically struggled to deliver. Banks do not tolerate downtime. They do not accept 'the bridge had a bug.' They require SOC 2 certifications, disaster recovery plans, and 24/7 support SLAs. Matter Labs has achieved the first of those milestones. The Cari Network will test the rest.

If you are building on Ethereum's L2 ecosystem — whether for DeFi, gaming, or enterprise applications — understanding how privacy-preserving infrastructure like ZKsync's fits into the broader roadmap is increasingly essential. The lines between public rollups and permissioned chains are blurring, and the developers who understand both worlds will be the ones best positioned for what comes next. If you're ready to build, thirdweb offers developer plans that scale with your project, with SDKs that support every major EVM chain including ZKsync.