By the middle of 2026 a single validator client was running under more than 95% of the stake securing Solana, and the tips it collected accounted for more than 60% of every priority fee paid on the network. That is not a competitive market. It is a utility with one meter.

Flowra launched its Open Orderflow Auction, a block-building framework for Solana that lets registered searchers bid openly for transaction inclusion rather than routing flow through closed channels. The Seoul-based company is also shipping Programmable Block Policy, which lets a validator define its own transaction inclusion rules at the block-building layer, and it has already announced a collaboration with compliance provider Honeypot to bring sanctions and risk screening into that layer.

The claim underneath it sounds backwards on first hearing. Expose orderflow to more competitors and validators earn more, not less. It is a claim with a precedent.

The experiment Ethereum already ran

Flowra's architecture is explicitly modelled on proposer-builder separation, the design that reshaped Ethereum's MEV market after the Merge, and the honest way to assess it is to look at both halves of what that experiment produced.

The first half worked. Separating the job of proposing a block from the job of building it created a competitive market of builders who bid away most of the value to win the right to have their block proposed, and a peer-reviewed empirical study measured a 261% increase in proposer revenue after the switch. Validators stopped extracting MEV personally and started auctioning the right to extract it, which turned out to pay considerably better.

The second half is the cautionary part, and Flowra should be judged on whether it has learned from it. Roughly 90% of Ethereum blocks now come through the MEV-Boost auction, and a handful of firms build almost all of them. Between October 2023 and March 2024 three builders produced about 80% of MEV-Boost blocks. By March 2025 two firms accounted for around 86%. By mid-2026 the top two were at roughly 73% and the top four were above 95%. An auction that is open in principle concentrated in practice, because access to private orderflow compounds: builders with exclusive flow win more blocks, winning more blocks attracts more exclusive flow, and the subsidy a newcomer needs to buy a 1% share rose from close to nothing before May 2023 to more than 1.4 ETH by March 2024.

The lesson is not that open auctions fail. It is that openness is a property that has to be maintained rather than declared, and the mechanism that erodes it is exclusive orderflow. That is precisely the thing Flowra's design targets.

What Solana looks like now

Solana never ran the Ethereum experiment. It arrived at concentration by a different route and ended up further along.

Jito tips accounted for roughly 10% of Solana priority-fee volume in April 2023. By 2025 they were above 60% and they have stayed there. Over the same period the Jito-Solana client went from running under 92% of active stake in late 2024 to more than 95% by mid-2026. One firm supplies the client, the block engine and the relay, which means it holds the whole pipeline from transaction submission to block production.

There is a structural reason Solana ended up here rather than somewhere more plural. Gulf Stream, Solana's transaction forwarding design, routes transactions privately to the upcoming leader rather than into a public mempool. There is no shared pool for anyone to observe, so visibility into MEV opportunity is a function of who you have a relationship with. Flowra's answer is to build the missing layer directly, having multiple validators share pending transactions to form an accessible stream that searchers reach through a standard API.

How concentrated is concentrated

Crypto commentary tends to describe market structure with adjectives. There is a standard unit for it, and it is worth applying.

The Herfindahl-Hirschman index squares each participant's market share and sums the results, so a single firm holding an entire market scores 10,000. Competition authorities in the United States treat anything above 1,500 as moderately concentrated and anything above 2,500 as highly concentrated. Ethereum's builder market has been measured at about 3,892, already well past that line. A 95% share on Solana implies a floor of 9,025 before anyone else is counted.

Put plainly, if Solana block building were an industry rather than a protocol, the number would not be a talking point. It would be a filing.

The pilot number, and what it does and does not show

Flowra is launching with one piece of performance evidence, and the piece is worth reading carefully because it is narrower and more interesting than a headline percentage.

On a single validator, a Flowra-enabled setup raised compute units per block by 20.6%, moving that node from 84% of the network average to 101%. Alongside it the company reports higher block fees than comparable validator software, 100% block production and 99.999% block engine uptime.

Compute units per block is the right metric to lead with, and not the obvious one. A validator's revenue is a function of how much billable work it fits into the slots it is given, so a node sitting at 84% of network average is leaving roughly a sixth of its earning capacity unused every time it leads. Moving it past 100% is a claim about block packing efficiency rather than about MEV extraction, which makes it independently checkable and harder to dress up. It is also self-reported and covers one node, which the company says plainly, and the number that will matter is what a cohort of institutional validators reports after a quarter rather than what one node did in testing.

Why the policy layer may matter more than the auction

The auction is the headline. Programmable Block Policy is the part that is likely to decide which validators Flowra actually wins. An institutional validator running stake for regulated counterparties has an unresolved problem. It is expected to apply sanctions and risk controls to what it processes, and the tooling available to it operates after transactions have already reached a block. Flowra's approach moves that check into block construction, where each validator writes its own policy rather than inheriting someone else's, and the Honeypot collaboration supplies the screening intelligence behind it.

The design decision that makes this defensible is that the policy sits at the block-building layer rather than in the protocol. Solana itself stays neutral. A validator that wants a compliance filter runs one, a validator that does not runs none, and no consensus rule changes either way. That is the difference between a network with configurable operators and a network with a filtered base layer, and it is a distinction the piece is deliberately careful about because it is the one a critic will reach for first.

What has to go right

Flowra is attempting to open a market that has every structural reason to stay closed, and four questions will decide whether it works.

Liquidity is the first, and it is the hardest. An auction with few bidders is a worse price discovery mechanism than a closed channel with one motivated buyer, so the searcher side has to arrive at roughly the same time as the validator side. Flowra has approached this from the stake end, onboarding institutional validators first, which is the correct sequencing because searchers follow blockspace rather than the other way round.

Exclusivity is the second, and it is the failure mode Ethereum demonstrated. Nothing prevents a large searcher from striking a private deal that routes its best flow away from the auction, and every incentive pushes toward it. The counterweight in Flowra's design is verifiability, since an auction whose bids and outcomes are auditable makes exclusive arrangements visible in a way that closed channels never do.

Latency is the third. Solana's slot times leave very little room for an auction round trip, and a framework that adds meaningful delay costs a validator more in missed blocks than it gains in bids. The 99.999% block engine uptime and 100% block production the pilot reports are the early evidence on this, and the figure to watch as the cohort grows is skipped slots rather than uptime.

Incumbency is the fourth. A validator switching away from the dominant client is not making a software decision, it is repricing its delegation relationships, its MEV commission and its stakers' expected yield at the same time. Flowra's advantage here is that the switching decision is being made by institutional operators with the analytical capacity to model it, which is a slower sale and a considerably stickier one.

What to watch

The honest read on this launch is that Flowra is running a well-understood playbook in a market that has not seen it, and the reason the timing works has less to do with technology than with what Solana's fee reform already changed. Since SIMD-0096 passed, validators keep 100% of priority fees, and tips already account for more than half of Solana's real economic value. The two largest components of what the network earns now sit with whoever builds the block.

That is the pool Flowra is opening to bidding, and it explains why a validator would take the switching risk. If the Ethereum result transfers even partially, the difference between auctioning blockspace and accepting a single counterparty's terms is not a rounding error in validator economics. It is the majority of the revenue line.

Three things will show whether it lands, and none of them is a press release. The first is the number of distinct searchers clearing bids, because that single figure separates an open auction from a rebranded private channel, and it is the number Flowra should publish first and most often. The second is whether the compute-unit result holds across a cohort. One node moving from 84% to 101% of network average is a promising engineering result and nothing more until twenty nodes do it, and the company has been careful to describe it as early testing on a single validator rather than a network claim. The third is what institutional validators do with Programmable Block Policy, because if compliance-driven operators turn out to be the ones who switch first, then the policy layer rather than the auction is the wedge, and Flowra will have found its market somewhere adjacent to where it aimed.

Underneath all of it is a question about what a blockchain's ordering rights are worth and who should be allowed to bid for them. Solana's engineering answer to transaction routing was privacy and speed, and Gulf Stream delivered both by handing transactions straight to the leader with no public pool for anyone to watch. That was a performance decision, and it produced a market structure nobody designed, in which value flows through channels most participants cannot see and one firm sits at the junction of nearly all of them. Flowra's wager is that the fix is not less privacy but more competition, and that a transparent auction beats a private relationship on price for the same reason it does in every other market where someone finally opened the book. Whether that is true on Solana is now an empirical question, which is a considerable improvement on it being a theoretical one.

Don’t forget to like and share the story!

Vested Interest Disclosure: HackerNoon has reviewed the report for quality, but the claims herein belong to the author. #DYOR.