Delphi Digital is betting that Solana’s next major development cycle will reposition the network as an “exchange-grade” environment capable of supporting onchain order books that can address real-world issues around latency, liquidity depth, and market structure. In the post of Jan. 20 in X titled “2026 is Solana’s year”, the research firm argued that Solana’s 2026 road is “its most intense development cycle” yet, “changing.”[s] everything from compliance to infrastructure to be Nasdaq distributed.”
Why Delphi Digital Is Calling 2026 “The Year of Solana”
Delphi has laid out the roadmap less as a fund to hold performance enhancements and more as a lack of capital markets: “Solana’s roadmap is about transforming it into an exchange-level environment where native onchain CLOB can compete fairly with CEX latency, liquidity depth, and fairness. Here’s everything that makes this development possible.” From that point of view, shaving off milliseconds is only important as it produces predictable, compelling results for applications such as high-frequency trading and medium limit order books.
In essence, Delphi wrote, Alpenglow, a compromise redesign that it called “the most important protocol change in Solana’s history.” The company said Alpenglow is introducing a new architecture built around Votor and Rotor, and Votor is changing the way insurers reach a deal. Rather than “merge multiple rounds of voting together,” validators will merge votes out of series and “commit to the end in one or two rounds,” producing “a theoretical finish in the 100-150 millisecond range, down from the initial 12.8 seconds.”
Related Reading
Delphi emphasized Votor’s similar finishing methods as a feature of durability, not just speed play. If a block gets “big support (80% + stake)” it ends immediately; if the support is between 60% and 80%, the second round starts, and the conclusion follows if that also clears 60%. The goal, Delphi argued, is to maintain endpoints even if there are unresponsive parts of the network.
Alpenglow also presents what Delphi calls the “20+20” resilience model: security holds as long as no more than 20% of the share is malicious, while life continues even if the other 20% is offline, “endure[ing] up to 40% of the network is malicious or inoperable while maintaining an endpoint.” Under this design, Evidence of History has been “effectively phased out,” replacing it with fixed-slot programming and local timers. Delphi said development is expected to begin in early to mid-2026.
Delphi also identified Firedancer, Jump’s C++ authentication client, as a structural improvement aimed at mitigating long-standing operational vulnerabilities. Solana has historically relied on a single client, now known as Agave, and Delphi has described “single farming” as a central weakness because errors at the client level can reach network-wide disruptions.
The goal of Firedancer, says Delphi, is a deterministic, high-performance engine that can process “millions of TPS with minimal latency variation.” Before full readiness, Delphi highlighted “Frankendancer,” an interim architecture that combines Firedancer’s production modules with Agave’s runtime and compatibility components, as a “big” bridge that increases customer diversity.
In infrastructure, Delphi highlighted DoubleZero as a private fiber overlay for certifiers, matching the profile of transmission and interconnection with traditional exchanges: “similar exchanges of the same infrastructure as Nasdaq and CME rely on microsecond-rate transmission.” The argument is that as the validation sets grow, the variance of the distribution becomes the enemy of robust storage windows. By routing messages in “right paths” and supporting multicast delivery, Delphi said DoubleZero can reduce latency gaps across authenticators—an aid to both Votor’s quorum design and Rotor’s broadcast design.
Related Reading
Delphi also outlines Solana’s roadmap for building blocks as a market structure project. It described JIto’s BAM (Block Assembly Marketplace) as separating ordering from execution using a marketplace and privacy layer, with functions embedded in TEEs so that “verifiers or builders can see the raw content of transactions before the ordering is executed,” reducing pre-execution behaviors such as startups.
Harmonic, on the other hand, aims to compete with manufacturers by introducing an open integration layer so that verifiers can accept proposals from “multiple competing manufacturers in real time,” with Delphi summarizing: “Think of Harmonic as a meta market and BAM as a micro market.”
Raiku concludes the thesis by adding fixed delay and programmable execution guarantees close to the Solana guarantee set, using Ahead-of-Time (AOT) sales for pre-committed workflows and Just-in-Time (JIT) execution for real-time requirements—without changing L1 compatibility.
Delphi finally tied the road of technology and market demand: the gravity of Solana, the integration of onchain perps in several locations, and the need to reach operational parity with centralized platforms. It cited expectations for “Solana native innovations such as Bulk Trade coming early next year,” and pointed to products like xStocks bringing “onchain stocks directly to Solana,” arguing that money and attention converge in a chain with faster money, better UX, and tighter capital.
At press time, SOL traded at $127.
The featured image was created with DALL.E, a chart from TradingView.com
