Four hundred milliseconds is a perfectly respectable interval if you are waiting for a lift. It becomes rather less respectable when a market price has moved and your old offer is still available to anyone quick enough to take it. This is the sort of problem that occupies Ellipsis Labs: the small delay with a large financial consequence.
- Phoenix began as a fully on-chain spot order book on Solana.
- Atlas takes the lessons downward, into blockchain infrastructure.
- Phoenix perpetuals carries the trading work into derivatives.
- Verifiable builds let developers check what actually runs on-chain.
Founders Eugene Chen and Jarry Xiao first met in middle school. They later brought backgrounds in mathematics, programming and high-frequency trading to a company founded in 2022. Their subject was decentralized finance. Their reference point was a functioning market: prices, spreads, cancellations and the people prepared to put capital behind a quote.
The market had to earn its visitors
The original Phoenix, introduced in 2023 and now called Phoenix Legacy, moved a familiar financial instrument onto Solana: the limit order book. Buyers and sellers state the prices they will accept. Market makers compete to supply those prices. The book is on-chain, and trades settle atomically, within the same transaction.
There is a pleasingly technical adjective attached to it: crankless. Some on-chain designs require a separate operation to finish processing trades. Phoenix Legacy dispenses with that step. For a trader, the useful result is fewer moving parts between an execution and its settlement. For a developer, the public code makes the mechanism available for inspection.
The company’s August 2023 seed announcement reported more than $250 million in spot volume without external incentives. That qualification deserves attention. A trading venue can buy activity by handing out rewards; Phoenix’s early proposition was that the trading opportunity itself should bring liquidity. Volume alone cannot prove a healthy market, but removing the subsidy makes the experiment more revealing.
An exchange discovers the floor beneath it
Phoenix also supplied a practical education in constraints. Haun Ventures’ 2024 investment account described an order book operating on roughly 400-millisecond Solana blocks, compared with a cited 15-millisecond latency at Binance. Those are different measures, not a controlled benchmark. They nevertheless explain the founders’ preoccupation: a market maker needs to update or cancel a quote before yesterday’s price becomes somebody else’s bargain.
Atlas, announced in September 2024, followed that experience. Ellipsis Labs proposed a blockchain specifically for financial applications, using a custom implementation of the Solana Virtual Machine and settling to Ethereum. Existing Solana programs were meant to carry over, while transaction scheduling, storage and verification received more specialized treatment.
The advertised priorities are unusually concrete: reliable transaction delivery, fast pre-confirmations, frequent oracle updates and honest sequencing. An oracle supplies outside price information. A sequencer determines transaction order. Both sit close to the question a trader actually cares about: what price did I get, and why?
The distinction from a general-purpose chain is deliberate specialization. Atlas targets developers whose financial applications are constrained by existing infrastructure. Its proposed benefits depend on reliable implementation and usable liquidity. Compatibility helps a program move; it cannot make a new market liquid by decree. The published launch roadmap is a statement of intent, rather than evidence that every promised milestone occurred.
The order book was a beginning
The product history has another turn. In December 2025, Ellipsis Labs introduced Phoenix perpetuals, initially in private beta. Perpetual futures let traders hold derivative exposure without a fixed expiry. The company described cheaper liquidity updates for market makers and gasless trading for retail users, applying proprietary AMM-style ideas to derivatives.
That announcement also identified SolFi, introduced in 2024, as a proprietary automated market-making system. The progression matters: Ellipsis Labs’ expertise extends beyond choosing an order book over an AMM. It is concerned with how liquidity gets priced and updated. Traders are the visible users; market makers and application developers are part of the machinery that makes their experience possible.
Capital, without confusing the columns
Three public funding announcements describe the resources behind the work: a $3.3 million seed led by Electric Capital in August 2023; a $20 million Series A led by Paradigm in April 2024; and a $21 million round led by Haun Ventures that October, intended to accelerate Atlas.
These are amounts raised, not the cost of building each product. The current careers page describes the company as profitable. That is a company statement, not a published income statement. Trading fees belong to the economics of these markets, but protocol volume, protocol fees and Ellipsis Labs revenue should occupy separate columns.
A fair launch needs a different starting gun
Gavel, introduced in May 2025, applies the same market-design interest to token distribution. A launch that rewards the fastest bot can leave ordinary buyers paying for somebody else’s speed. Gavel combines public sales with sandwich-resistant liquidity bootstrapping. Its demonstration token, IBRL, was explicitly described as having no present or future utility.
The example uses a fixed sale window and proportional allocations. That removes one race while introducing a tradeoff: depositors do not know their final price when they commit. The broader lesson is practical. Choose the distribution mechanism for the market you want, and align the sale’s closing price with the opening liquidity price.
Show the work. Check the deployment.
Ellipsis Labs also contributes research and engineering tools. Phoenix Legacy’s repository publishes an OtterSec audit and instructions for verifying that deployed code matches the source. The distinction is useful: reading a recipe does not establish what was served. Build verification connects the two, although it cannot establish that a program is free of bugs.
“Process exists only to support outcomes.”Ellipsis Labs’ published hiring principles
The team works in person in New York and emphasizes autonomy and technical ability. Its most transferable idea is the sequence of work: operate an application, identify the constraint, then justify the infrastructure. Ellipsis Labs’ products make that sequence visible. The clock remains a demanding customer.