Crypto Market Makers Could Become Compute Allocators

Ritual asked one short question on X on 18 November 2025: what if market makers could evolve beyond trading. The post, by the project’s @ritualnet account, framed two options for the role: stay the price-quoter the asset class already knows, or step into the architect’s job for any scarce resource in crypto. A second post by Naveen Durvasula, quoted in the question Ritual posted on X, laid out a mechanism that tries to answer yes.

That question lands on a market that has spent the year widening its own definition of what a market maker does. Acheron Trading secured the first crypto MiCA license in 2025; DWF Labs already trades on 60+ venues; Keyrock opened a U.S. office. The firms are moving outward, not yet toward the resource-allocation shift Ritual is asking about. That gap is the story the rest of this piece follows, and it is the gap the next year of crypto market structure will be measured against.

The Question That Reframes a Role

Ritual’s question on 18 November 2025 was two sentences long. The first asked whether market makers could evolve beyond trading. The second asked whether those entities could become the architects of allocating any scarce resource in crypto. Both questions came tagged to a new post by Naveen Durvasula breaking the framing down.

Market makers today are paid to keep an order book alive: narrow the bid-ask gap and recapture the spread. The November post pointed at a different compensation surface, one tied to the value a market maker extracts by allocating rather than quoting. The shift relocates where a market maker’s profit comes from, not what its bid and ask screens look like. The November post reframes a job the rest of crypto still treats as plumbing, the kind of work that does not get a keynote at any conference because no one is meant to notice it.

Durvasula’s contribution is the part the question foreshadows. The piece, hosted on Ritual’s blog under the title “Markets for Decentralized Computation,” argues that market-making can be generalized past asset pricing and into any decentralized computing system. The mechanism adds allocation to a market maker’s brief, treating which nodes should execute which requests as the core pricing problem. Posted prices and posted node rewards both come out of the same calculation. That formulation turns the November question into a concrete research problem rather than a slogan, and it sits inside a category that has been quietly forming since automated market makers took over trading on Uniswap.

How Durvasula Designs a New Kind of Market Maker

Durvasula’s mechanism targets the messy part of decentralized compute: requests that require specialized hardware, nodes with hard capacity limits, and allocations that have to satisfy many constraints at once. The protocol picks the most economically valuable set of user requests the network can serve under those constraints, and it outputs posted prices for users and posted rewards for nodes. Both sides see a single price and never a bidding war. Details of the design are laid out in Ritual’s “Markets for Decentralized Computation” post, the same piece Durvasula linked from his own X account the morning of 18 November 2025.

The mechanism obtains the information it needs without forcing users or nodes to bid. On a central limit order book, market makers post prices for a traded asset and update them as their inventory moves. Durvasula’s design uses third-party market makers as intermediaries that learn user willingness to pay and node costs, then internalize that information to assemble the most valuable allocation. The third party here is paid for solving the allocation problem, not for quoting a price. That single swap is the conceptual leap the question on X was pointing at, and it is what makes the protocol worth a posted-price experience rather than another auction layer glued onto an existing chain.

To make the role change concrete, here is how the new job compares with the old one along three working attributes.

The mechanism also has to handle the constraints that come with general compute. In Ritual’s setting, a request might require specialized hardware on a particular node, or several nodes redundantly, or no node at all. The mechanism finds the highest-value allocation those constraints allow and pays out accordingly. Both sides agree to the price because the design guarantees neither party is extracted and neither side subsidizes the other.

Market Makers Are Already Pushing Past the Order Book

The role definition at the major crypto market-making shops in 2025 looks nothing like the textbook description. Across the year the firms that dominate token order books spent their time pushing into adjacent businesses: regulated venues, OTC desks, expanded footprints, and U.S. ground. The moves don’t yet look like the compute allocation shift Ritual is asking about, and that is the point of cataloguing them. They look like market makers preparing for one, even if the firms themselves wouldn’t name the preparation that way.

Four signals from the year show what that preparation looks like in practice.

• Acheron Trading became the first crypto market maker to secure a Markets in Crypto-Assets (MiCA) CASP license in the EU, in 2025.
• DWF Labs trades spot and derivatives on 60+ top crypto exchanges including Binance and Bybit, and manages a portfolio of more than 1,000 projects.
• Keyrock, founded in Belgium, opened a U.S. office in 2025 as part of a broader international push.
• Wintermute broadened beyond on-screen trading, building out OTC and institutional services through the year.

Each fits the pattern the post on X described without quite matching its ambition.

Where Ritual Sits in the Decentralized AI Stack

Ritual is positioned as a decentralized AI infrastructure network. Its main page describes the project as building the infrastructure “autonomous intelligence needs: native compute, privacy, verification, coordination, and markets for long-lived agents.”

Ritual’s economics are anchored in crypto’s venture cycle, not its trading floor. The protocol sits in the same category of decentralized-AI compute plays as Bittensor, Morpheus, and others chasing demand for AI compute on rails that don’t route through a hyperscaler. That category is also where any deployment of the new mechanism would land. For broader context on how the wider crypto market sits against AI flows, how Bitcoin’s recent distribution looks against ETF outflows tracks the macro backdrop this category sits inside.

The opportunity is one CoinDesk sized up in February 2026. An opinion column in the publication put centralized AI enterprise value at roughly $12 trillion and decentralized AI at about $12 billion. The same piece projected the blockchain AI market growing from $6 billion in 2024 to $50 billion by 2030, a 42.4% compound annual growth rate, and AI infrastructure spending hitting $300 billion in 2025 alone. Those figures belong to a single forecaster, and the column’s author acknowledged they might be conservative. For the underlying sizing, see the CoinDesk opinion sizing the decentralized AI buildout.

• 60+ – crypto exchanges DWF Labs trades spot and derivatives on, including Binance and Bybit.
• 1,000+ – projects in DWF Labs’ market-making portfolio.
• $300 billion – AI infrastructure spend projected for 2025, including the $500 billion Stargate initiative, per a CoinDesk opinion piece dated 22 February 2026.
• $6 billion to $50 billion – projected blockchain AI market size, 2024 to 2030, at a 42.4% CAGR, per the same CoinDesk opinion piece.

Market makers tend to sit where new volume and new asset classes form. The decentralized AI buildout creates both at once: tokens that need deep books and compute resources that need priced allocation. A role that sits at that intersection has more surface to clear than one that sits at either end of it.

The fit is concrete: tokens need deep order books, and compute needs priced allocation. A market maker that does both would meet both demands. Whether the firms already running the order books are the ones to take that job on is the open question the rest of the piece turns to.

The Gap Between the Mechanism and the Market

The mechanism Durvasula wrote up solves a real problem. What it doesn’t yet have is a live deployment at scale, a regulator willing to bless it, or a market maker willing to be paid for resource allocation when the existing spread already clears. The November post was, by Ritual’s own framing, a question. The Durvasula contribution was, by its own header, an early post meant to be read on its own. Neither ships the design into production on its own.

The market on the other side doesn’t read like the design’s input assumptions. Volume and liquidity on major venues in 2025 ran through firms with regulator-friendly footprints and broad product coverage: Wintermute’s drift into OTC and institutional flow, Acheron’s MiCA license, DWF Labs’ reach across 60+ venues, Keyrock’s U.S. expansion. The shape of those moves is outward into new products and new geographies, not into resource allocation. What those firms build next is the test the question on X was pointed at.

selects pricing in a way that is not extractive, nor requires subsidies.

The line comes from the summary of Naveen Durvasula’s post “Markets for Decentralized Computation” on Ritual’s blog, which lays out the design choices the mechanism makes and the property it claims for itself.

Whether the mechanism becomes the working template for crypto market structure, or stays one more elegant answer waiting for a market that hasn’t shown up yet, will be settled by what those firms do next. The answer belongs to them.

Frequently Asked Questions

What did Ritual ask about crypto market makers?

Ritual, the account @ritualnet, posted on X on 18 November 2025 asking whether a market maker’s job could expand past trading into the broader work of allocating scarce resources in crypto. The post was framed as a question and pointed to a new piece by Naveen Durvasula that lays out a mechanism to do it.

What is Durvasula’s mechanism for decentralized compute?

Durvasula’s mechanism generalizes market-making beyond the central limit order book and into any decentralized computing system. A third-party market maker learns user willingness to pay and the costs nodes bear, assembles the most valuable allocation of requests to nodes within the network’s constraints, and produces posted prices for users and posted rewards for nodes that both sides will accept.

Are crypto market makers already changing what they do?

Several large firms have already widened their definition of the role. Acheron Trading secured the first MiCA license for a crypto market maker in the EU during 2025. DWF Labs trades spot and derivatives on 60+ crypto venues. Keyrock, founded in Belgium, opened a U.S. office in 2025. Wintermute has spent recent years pushing further into OTC and institutional flow rather than staying on screen only.

How would market makers become compute allocators in practice?

Under the Durvasula design, a market maker is paid for solving an allocation problem, not for quoting a price. The market maker internalizes information about what users will pay and what it costs nodes to run a workload, then assembles a portfolio of request-to-node matches that maximizes network value, with posted prices and posted node rewards attached to the result.

When would a Durvasula-style mechanism actually deploy?

No source consulted places a live deployment at scale. Ritual’s post was, by its own framing, a question, and Durvasula’s piece self-describes as an early post meant to be read on its own. Whether the firms already pivoting into adjacent lines of business adopt it, ignore it, or build something adjacent is the open question that the rest of 2026 will start answering.