
Dunamu is deploying its GIWA Chain on Optimism's OP Enterprise framework, aiming to secure control over its infrastructure for 13 million Upbit users.
Dunamu, the operator of South Korea’s dominant crypto exchange Upbit, has formalized a partnership with the Optimism Foundation to deploy its GIWA Chain using the OP Stack. This move marks the first implementation of the Self-Managed tier within the OP Enterprise framework, signaling a shift in how major centralized exchanges approach blockchain infrastructure. By moving from a rented network model to a self-managed layer 2, Dunamu aims to secure operational autonomy while maintaining the performance standards required for its 13 million registered users.
The core of this transition lies in the distinction between standard L2 usage and the Self-Managed tier of OP Enterprise. In the current market landscape, many exchanges rely on public networks where transaction sequencing and network governance are handled by third-party validators or decentralized protocols. Under the new arrangement, Upbit will retain control over the primary sequencer and all core network decisions. This architecture is designed to mitigate the risks associated with external network congestion or governance disputes that could otherwise disrupt trading operations.
Optimism’s role shifts from a network provider to an infrastructure partner. The foundation will provide institutional-grade support, including system monitoring and a backup sequencer for failover. This hybrid model allows Dunamu to leverage the open-source development of the OP Stack while keeping the critical path of transaction execution within its own internal environment. For an exchange that ranked second globally in cumulative spot trading volume between 2020 and 2024, this level of control is a prerequisite for scaling into high-frequency, real-world payment applications.
GIWA Chain is engineered for high-throughput environments, boasting one-second block times and full EVM compatibility. The latter is essential for ensuring that existing Ethereum-based applications can migrate to the chain without significant code refactoring. As of May 3, the testnet has processed nearly 100 million transactions, providing a stress-test baseline for the upcoming mainnet launch.
This performance profile is intended to support the integration of Korean won-backed stablecoins, a key component of Dunamu’s broader fintech strategy. By building on a proven stack, the company avoids the technical debt associated with developing proprietary consensus mechanisms from scratch, instead focusing its engineering resources on compliance and user-facing features.
Dunamu’s push into self-managed infrastructure is occurring alongside its pursuit of a merger with Naver Financial. This deal, which aims to create a fintech entity valued at approximately 20 trillion Korean won (roughly $13.5 billion), has faced regulatory delays that have pushed the expected closing to late 2026. The GIWA Chain is intended to serve as the technological backbone for this combined ecosystem.
For market observers, the shift toward self-managed L2s represents a broader trend of institutional operators seeking to internalize their blockchain stacks. As Jing Wang, director at the Optimism Foundation, noted, the largest exchanges are moving away from renting infrastructure in favor of owning the chains their users transact on. This trend is likely to influence how other major platforms approach crypto market analysis and infrastructure procurement in the coming years.
Investors tracking this development should monitor the mainnet launch timeline and the subsequent integration of won-backed stablecoins. The success of this deployment will be measured by the chain’s ability to maintain uptime and transaction throughput during periods of high market volatility. If the GIWA Chain can successfully handle the volume of Upbit’s retail and institutional base, it could set a new standard for exchange-operated L2s, potentially reducing reliance on public, multi-tenant networks for core trading activities. Conversely, any technical failures or delays in the mainnet transition would likely force a re-evaluation of the self-managed model for other large-scale operators looking to replicate this strategy.
AI-drafted from named sources and checked against AlphaScala publishing rules before release. Direct quotes must match source text, low-information tables are removed, and thinner or higher-risk stories can be held for manual review.