The House Edge

Appearance

Chapter 8 — Research Note (Phase 1) ​

Status: RESEARCH (Phase 1 — written by agent, awaiting Nick's review) Date: 2026-05-14 Working chapter: 08 — Solana: AMMs, Auctions, and the Speed Game Part: III — The Architectures (Ch 8 is Part III's opener; Chs 9 Hyperliquid, 10 Ethereum + L2s follow)

How to read this note ​

Ch 8 is Part III's opener and the first of three chain-architecture chapters. Its job is synthesis, not introduction: Chs 2/3/5/6/7 already developed every Solana mechanism the chapter needs. Ch 8 assembles them into one integrated picture and lands its load-bearing observation: Solana is the chain where the validator IS the slot leader IS the block constructor; client choice plus infrastructure relationships dominate validator revenue; the access-vs-operational gap (Ch 7's +33–101% vs ~+3%) is an order-of-magnitude difference that is Solana-specific because of this vertical integration.

Four things to know up front:

  1. No concepts re-introduced. Jito, BAM, Harmonic, Beam, Frankendancer, Rakurai, Jupiter, the prop-AMMs, Helius, the June 2024 enforcement, the Vpe / DeezNode case, the +101% Harmonic gap, the Chorus One timing-games paper — all already in Chs 2/3/5/6/7. Ch 8 references with chapter-citations and develops the cross-layer integration.

  2. The "Pittsburgh" item in the prompt is wrong. Jito's North American mainnet block engines are New York and Salt Lake City (Dallas is testnet). Full mainnet footprint: Amsterdam, Dublin, Frankfurt, London, NY, Salt Lake City, Singapore, Tokyo. Recent BAM additions: Dublin, Dallas. Lithuania next; Brazil / Hong Kong / Washington DC / Miami / South Africa under evaluation. The chapter should name the actual sites.

  3. The chapter's strongest integrating fact is geographic. Top three cities by Solana stake: Frankfurt 19%, Amsterdam 16%, London 12% (together 47%). Top six (adding Vilnius 6%, Tokyo 4%, Ashburn 4%) ≈ 61% of stake. Jito block engines run in five of those six cities. The "top-quintile validator" is co-located in Equinix FR5/AM3 with the block engine and major searchers, on a Jito-derivative client, with a BAM-node relationship — every compounding variable is geographic, client-side, and relationship-side, not operational.

  4. The Q1 2026 prop-AMM picture has shifted from Ch 2. Ch 2 cited HumidiFi ~65% of the prop-AMM segment (Helius mid-2025; Blockworks Jan 2026). March 2026 is different: prop-AMM share of Solana DEX volume ~55% (down from 68% in February); BisonFi (Forward Industries, Q4 2025 launch) briefly overtook HumidiFi, both then declining >55% through March. Ch 8 should frame the segment as competitively contested rather than HumidiFi-dominant.

1. Key claims ​

Each numbered claim is something the chapter is allowed to state. Sources cited inline.

The chapter's central architectural argument

  1. Solana is the chain on which the validator is simultaneously the slot leader, the block constructor, the transaction-sequencer host, and (under BAM) the executor of TEE-attested ordering decisions made by separate node operators. No other major chain integrates these roles into one actor: Ethereum separates proposer from builder via PBS (Ch 5); Hyperliquid runs matching inside HyperBFT consensus with no separable construction role (Ch 5, Ch 9). The integration is what makes client choice, scheduler mode, and infrastructure relationships dominate Solana validator economics in a way they do not dominate the others.

  2. The access-to-operational revenue gap on Solana is an order of magnitude. Per Ch 7: Harmonic Performance validators earn ~+101% priority fees vs median (Balanced +39%, Agave Harmonic +36%); Chorus One's August 2025 research caps the timing-games-plus-scheduler-optimisation gain at ~+3.0% of total rewards (~27 bps annualised). Ch 8 lands the architectural version of the claim: the gap exists because Solana's vertical integration makes the validator's revenue ceiling a software-and-relationships function rather than a physical-operations function. On Ethereum the gap collapses into the builder layer; on Hyperliquid it collapses entirely. (Syndica — Deep Dive: Solana Onchain Activity, March 2026; Chorus One — Timing Games)

Aggregation: Jupiter's structural dominance of the retail interface

  1. Jupiter routes ~93.6% of Solana aggregator volume in Q1 2026; aggregator-routed volume is ~74% of total Solana DEX volume (up from ~40% six months prior). The interface-replacing-the-pool dynamic from Ch 2 is now the chain's structural shape: the pool is inventory, the aggregator is the front of the market. 2025 cumulative routed volume: ~$716B. (SolanaFloor — 93.6%; SolanaFloor — 70% routed through aggregators)

  2. Jupiter Beam — the private transaction-landing engine routing signed transactions to Jito as private bundles — is the default routing inside Jupiter Ultra in 2026. Beam opened as a public Transaction Submission API in April 2026 (any signed Solana tx, not just Jupiter-routed). Published Ultra V3 slippage: ~+0.006% on Beam-protected paths vs −0.14% on unprotected (Ch 3). Ultra handles ~95% of aggregator volume and ~50% of total DEX volume. (Jupiter Ultra docs)

Why no Solana CLOB has won spot

  1. The Solana CLOB experiment has ended for spot. Phoenix Q2 2024 peak $3.7M → Q1 2026 $68,604 quarterly revenue (~54× decline); renamed Phoenix Legacy, pivoted to perpetuals (Ch 2). OpenBook is small; Manifest at ~$3.9B 30-day volume is the only above-noise spot CLOB and is <2% of Solana's ~$284.5B Q1 2026 DEX spot volume. The architectural reason: mempool surface (Ch 3) plus prop-AMM displacement (Ch 2) — retail routes through Jupiter, Jupiter routes to prop-AMMs, the order-book design produced too little volume to compensate the market makers who would host it. Those market makers now operate prop-AMMs instead. (DefiLlama — Phoenix; DefiLlama — Manifest)

Prop-AMMs as the dominant spot venue

  1. Prop-AMMs ~55% of Solana DEX volume in March 2026 (down from 68% in February), ~$36B combined monthly volume of $65B total. Competitively contested: BisonFi (Forward Industries, Q4 2025 launch) overtook HumidiFi in late January 2026; both then declined >55% in March. Named operator map: HumidiFi → Temporal (reported, Ch 2); SolFi → Ellipsis Labs (publicly claimed, Ch 2); Tessera V → Wintermute (Wintermute-confirmed to DL News Aug 2025); BisonFi → Forward Industries (Dec 2025); ZeroFi, Goonfi → undisclosed. (Syndica — DeFi March 2026; SolanaFloor — BisonFi; DL News — dark AMMs)

The Jito auction: what's actually being auctioned

  1. The Jito Block Engine auctions three things in one product, not "MEV from a public mempool" (Solana's Gulf Stream produces none, Ch 3): (a) inclusion priority for searcher bundles into the next slot's block, against SOL tips; (b) atomic bundle execution with revert protection (whole-or-nothing); (c) routing access to validators on Jito-derivative clients. ~32% of stake runs Agave Jito; ~28% JitoBAM; Jito-Solana variants total ~94% of stake by Q2 2025 (Ch 5, Ch 6). Jito tips: late-2024 ~$2.5M/day average, single-day high $14.7M (17 Nov 2024). Post-JIP-24 the entire 6% Block Engine + BAM cut routes to the Jito DAO treasury; projected $15–50M annual run-rate. (JIP-24)

BAM and TEE-encrypted block construction

  1. BAM launched on Solana mainnet 21 July 2025 with four launch BAM Node operators: Helius, Triton One, SOL Strategies, Figment. BAM Nodes run inside AMD SEV-SNP TEE enclaves; sequencing delegated to BAM Nodes, execution stays with the validator (Ch 3, Ch 5). Late February 2026: ~25% of stake connected through BAM. New nodes in Dublin and Dallas; Lithuania next; Brazil / Hong Kong / Washington DC / Miami / South Africa under evaluation. Open-source target mid-Q2 2026 as prerequisite for onboarding third-party operators at scale. (Jito — Introducing BAM; Chainflow Feb 2026)

Validator client diversity in March 2026

  1. Solana client distribution (March 2026): Agave Jito 32%, JitoBAM 28%, Agave Harmonic 17%, Frankendancer 12%, Rakurai 6%, Firedancer 2%. ~86% on Agave-based clients. Maintainers: Anza (Agave); Jito Labs (Jito-Solana variants + JitoBAM); Jump Crypto (Frankendancer + Firedancer); Rakurai (closed-source Agave fork). Structural point: none of these clients is sandwich-default in 2026. The combination of the Foundation's June 2024 enforcement, Harmonic's SFDP-compliant strategies, the BAM TEE architecture, Beam's protected routing, and the 8 April 2026 protocol patch (all per Chs 3, 5) collectively closed the formal-sandwich surface. (Syndica March 2026)

Rakurai and the "no timing games" positioning

  1. Rakurai is a closed-source Agave fork from an ASIC/SoC-background team; ~2% of stake Jan 2026 → ~6% Mar 2026. $3M seed in 2025 led by Anagram Ventures (Ch 5). The architectural differentiator: explicit commitment that Rakurai "do[es] not resort to block timing game manipulation with intentional delays past the target block time of 400ms" (Figment migration report, Ch 5). The 2 March 2026 Figment migration is the chapter's cleanest published institutional-adoption case (Ch 5). Other named adopters: P2P.org, GlobalStake. (PRNewswire; Figment)

Geographic proximity to block engines — Ch 8's signature new material

  1. Solana stake concentrates in six cities: Frankfurt 19%, Amsterdam 16%, London 12%, Vilnius 6%, Tokyo 4%, Ashburn (N. Virginia) 4% — together ~61% of all staked SOL. West Europe + North America host ~70%. (Helius — Decentralization)

  2. Jito mainnet block engines run in eight cities: Amsterdam, Dublin, Frankfurt, London, NY, Salt Lake City, Singapore, Tokyo. Optimal round-trip latency block-engine-to-validator <50ms; <100ms operates within current bundle throughput. Measured slot latencies: 0.25 ± 0.9 Amsterdam→Europe; 0.47 ± 1 Utah→N. America; 0.29 ± 0.46 Tokyo→Asia. The block-engine footprint maps almost exactly onto the stake footprint. The geography is not accidental; it is the equilibrium of where validators colocate given where block engines run, which is itself a function of where major searchers colocate. (Jito — Low-Latency Send)

  3. Standard Solana validator colocation: bare-metal AMD EPYC 9354P + 512GB RAM + 3× NVMe (one-time $8–15K), plus $200–500/mo rack and power in Equinix-tier data centres. Premium colocation $400–800/mo. The "next to the TPU port" pattern from Ch 3 (non-voting validator alongside paid RPC in same rack, forwarding to current leader via QUIC) is the load-bearing latency play; RPC Fast reports 5–10× latency reduction from LAN-local vs remote cloud. Equinix FR5 (Frankfurt) and NY5 (NY) are the named anchor data centres. (RPC Fast; Good Shell)

The SFDP compliance regime

  1. SFDP (Solana Foundation Delegation Program) is the largest single-source stake-delegation pool and the chain's de facto soft-governance lever. Starting 1 May 2026: participants must operate on an ASN and hosting provider holding <25% of network stake; data-centre concentration must be <15% across all staked validators at any single data-centre provider. The "onboard 1, offboard 3" policy retires validators with <1,000 SOL external stake after 18 months. ~150 validators projected to lose foundation stake under the May 2026 rule. Two structural levers: (a) anti-concentration (the data-centre rule counteracts the Frankfurt/Amsterdam colocation gravity); (b) anti-extraction (censorship-resistance + "no sandwich-enabling private mempools" rules; June 2024's ~32-validator removal, Ch 3). (Solana Foundation — Delegation Criteria; Phemex News)

Alpenglow and the forward-looking architectural shift

  1. Alpenglow (SIMD-0326) activated on testnet 11 May 2026; mainnet target late Q3 / early Q4 2026. Full architectural treatment is Ch 12. Three Solana-specific shifts: (a) vote transactions removed from blockspace (~70% of all Solana transactions at peak), freeing block capacity; (b) VAT ~1.6 SOL/epoch entirely burned — first explicit per-epoch fixed validator cost; (c) profitable-validator stake threshold drops from ~4,850 SOL to ~450 SOL (~90% reduction). (SIMD-0326; CoinDesk 11 May 2026)

The chapter's verdict frame

  1. The chapter's verdict frame — three-actor stratification: (a) retail traders lose via slippage, residual long-tail sandwiches, and cumulative take at every stack layer — but lose less than in 2024 because Beam + BAM + the April 8 patch + Foundation enforcement closed the formal-sandwich surface; (b) market makers win spot trading by internalising adverse selection via prop-AMMs and solvers; (c) validators with infrastructure relationships win the most, with the largest dollar magnitude per actor and lowest visibility. The "top validator" — Frankendancer Harmonic Performance in Frankfurt FR5 with a Helius-backed RPC and a BAM-node relationship — earns ~+101% priority fees vs a stock Agave-default median validator with no relationships and no SFDP delegation. The difference is overwhelmingly software + infrastructure access, not operational excellence. (Ch 7 develops; Ch 8 lands the chain-specific version.)

  2. Solana's 2025 REV ~$1.4B; Jito tips 41.6%–66% of REV by month; January 2025 peak $551.7M / single-day high $56.8M (19 January). Q1 2026 REV ~$89.5M, down 68% YoY as memecoin volume cooled. Solana ranks #2 by REV behind Hyperliquid in Q1 2026 — itself a structural data point about value capture under different architectures. (Blockworks REV dashboard; PANews Q1)

2. Numbers to verify ​

#NumberSourceDateFlag
N1Jupiter aggregator share Q1 2026: ~93.6%; aggregator-routed share of total DEX volume ~74%SolanaFloorQ1 2026Already cited Ch 2
N2Solana DEX volume March 2026: ~$65B (33% contraction from October 2025 peak); prop-AMM share 55% / $36B; HumidiFi and BisonFi both declined >55% in MarchSyndica DeFi March 2026March 2026New, supersedes Ch 2's HumidiFi-65% framing
N3Prop-AMM operator map: HumidiFi → Temporal (reported); SolFi → Ellipsis Labs; Tessera V → Wintermute (confirmed); BisonFi → Forward Industries; ZeroFi/Goonfi → undisclosedDL News; SolanaFloor BisonFi2025–2026New: Wintermute/Tessera and Forward/BisonFi confirmations
N4Jupiter 2025 cumulative routed spot volume: ~$716BYellow / blockchainreporter.net2025–2026Strong
N5Phoenix peak Q2 2024 $3.7M → Q1 2026 $68,604 quarterly revenue (~54× decline); pivoted to Phoenix PerpetualsDefiLlama Phoenix; Ellipsis Labs — Phoenix PerpetualsQ1 2026Already cited Ch 2
N6Manifest 30-day spot volume ~$3.9B; Solana Q1 2026 DEX spot volume ~$284.5BDefiLlama Manifest; Pine AnalyticsQ1 2026Already cited Ch 2
N7Solana stake-weighted client distribution March 2026: Agave Jito 32%, JitoBAM 28%, Agave Harmonic 17%, Frankendancer 12%, Rakurai 6%, Firedancer 2%SyndicaMarch 2026Already cited Chs 3, 5, 6
N8Harmonic priority-fee gap per block vs network median: FD Harmonic Performance +101%; FD Harmonic Balanced +39%; Agave Harmonic +36%SyndicaMarch 2026Ch 5 / Ch 7 anchor
N9Jito block engines (mainnet): Amsterdam, Dublin, Frankfurt, London, NY, Salt Lake City, Singapore, Tokyo. Optimal latency <50ms; <100ms operates wellJito docs2026New: corrects prompt's "Pittsburgh"
N10Solana stake by city: Frankfurt 19%, Amsterdam 16%, London 12%, Vilnius 6%, Tokyo 4%, Ashburn 4% (top 6 ≈ 61%); West Europe + N. America ≈ 70%Helius Decentralization2025–2026New chapter-level data
N11BAM stake adoption ~25% by late February 2026; launch BAM Node operators: Helius, Triton One, SOL Strategies, FigmentChainflow; Jito BAM launchFeb 2026New stake-adoption number
N12BAM Node expansion: new sites Dublin + Dallas; Lithuania next; Brazil/Hong Kong/Washington DC/Miami/South Africa under evaluation; open-source target mid-Q2 2026ChainflowFeb 2026New
N13Jito tips late 2024: ~$2.5M/day average; single-day high $14.7M on 17 Nov 2024Tokenomics.comlate 2024Indirect anchor
N14Post-JIP-24: entire 6% Block Engine + BAM cut routes to Jito DAO treasury; projected $15–50M annual run-rateJito Foundation — JIP-242025Already cited Ch 6
N15Solana 2025 REV ~$1.4B; Jito tips 41.6%–66% of REV by month; Jan 2025 peak $551.7M / $56.8M single dayBlockworks REV dashboard; News bit2me2025Strong
N16Q1 2026 Solana REV ~$89.5M (down 68% YoY); rank #2 behind HyperliquidPANewsQ1 2026New ranking fact
N17Rakurai stake share: ~2% Jan 2026 → ~6% March 2026; closed-source Agave fork; "no timing-game" positioning explicit in Figment migration report (2 March 2026)Syndica; Figment2026Already cited Ch 5
N18Figment Rakurai migration (2 Mar 2026) outcomes: SRR 6.85% → 7.17% (+32 bps); priority fees +60%; ~5× MEV capture; tip capture median +158%, p95 +55%FigmentMar 2026Already cited Ch 5
N19Chorus One timing-games-plus-scheduler-optimisation: ~+3.0% total rewards (~27 bps annualised); timing games alone ~+1.19%; scheduler alone ~+1.4%Chorus OneAug 2025Already cited Chs 5, 7
N20Solana validator profit Gini ~0.93Placeholder VCSept 2025Already cited Ch 7
N21Alpenglow: testnet 11 May 2026; mainnet target late Q3 / early Q4 2026; SIMD-0326 approved with 98.27% validator vote (Aug 2025)CoinDesk; SIMD-03262025–2026Already cited Chs 1, 5
N22VAT ~1.6 SOL/epoch entirely burned; profitable-validator threshold drops from ~4,850 SOL to ~450 SOL (~90% reduction)SIMD-0357; Helius — Alpenglow2025–2026Already cited Ch 5
N23SFDP May 2026 rules: <25% ASN/hosting-provider stake-share rule; <15% data-centre concentration rule; ~150 validators projected to lose foundation stakeSolana Foundation — Delegation Criteria; PhemexMay 2026New chapter-level enforcement framing
N24Solana Foundation June 2024 enforcement: ~32 validators removed, ~1.5M SOL / ~0.5% of delegation-program stake; Tim Garcia "enforcement ongoing"; no published itemised list of subsequent removalsCoinDeskJune 2024Already cited Chs 3, 7
N25DeezNode / Vpe: 1.55M sandwich tx / 30 days; 65,880 SOL ($13.43M) profit; 50% revenue-share offer; not publicly removed from SFDP between June 2024 and May 2026Helius MEV report; Anarcaze MediumQ4 2024–Q1 2025Already cited Chs 3, 4, 7
N268 April 2026 Solana protocol-level patch further compresses sandwich surface; specific SIMD / validator-client release tag / consensus mechanism not publicly documentedEdgenApril 2026Flag — Ch 3 already flagged; unresolved mechanism citation
N27Colocation costs: bare-metal AMD EPYC 9354P + 512GB RAM + 3×NVMe one-time $8–15K; rack+power $200–500/mo; premium $400–800/moThe Good Shell; RPC Fast2026Strong; new chapter-level data
N28Solana 2025 MEV ~$720M (Jito tips); first year MEV > priority fees as largest REV componentHelius MEV report2025Already implicit in Ch 6

Deliberately not pinned:

  • Specific 8 April 2026 patch's consensus mechanism (Ch 3 already flagged)
  • Aggregate annual Helius MEV-rebate flow on Solana (not publicly disclosed, Ch 7)
  • Specific BAM-Node operator economics (revenue-share to operators not publicly disclosed beyond the 6% pool routing to DAO)
  • Specific operator attributions for ZeroFi / Goonfi (not publicly disclosed)
  • Whether DeezNode has been removed from SFDP post-June 2024 (no public list)

3. Contested or evolving claims ​

  • Prop-AMM ranking volatility. Jan–Mar 2026: HumidiFi → BisonFi → both collapsed. Chapter should frame the segment as competitively contested, not single-firm-dominant. Structural point (prop-AMMs collectively >50% of spot DEX volume) is durable; the leaderboard is not.

  • The 8 April 2026 patch. Ch 3 flagged that the SIMD identifier, validator-client release tag, and consensus mechanism are not in a canonical primary source. Ch 8 inherits the flag; cite the effect without committing to mechanism. Anza release note or SIMD merge would resolve before publication.

  • Jupiter Beam share of total DEX volume. Cited at ~95% of aggregator-routed and ~50% of total DEX volume (Jupiter's own materials, not independently audited); directionality uncontested.

  • The "validator = slot leader = block constructor" framing. Under BAM, sequencing is delegated to BAM Nodes inside TEEs while execution stays with the validator — a partial separation. But unlike Ethereum PBS, the validator and BAM Node run coordinated software in one operational substrate rather than bidding across a relay layer. Frame carefully: Solana is integrated relative to Ethereum, less integrated than Hyperliquid's in-consensus matching. The three-chain stratification matters.

  • Alpenglow timing. Activates between the chapter's writing date and book publication. If publication slips past mainnet activation, the "what changes when…" recasts as present-tense. Ch 12 carries the full architectural treatment regardless.

  • Tessera labeling. "Tessera V" is the Wintermute-operated dark AMM; "Tessera" routing through Titan is the aggregator-side share. Be consistent.

  • The "block engine ↔ stake distribution" causal claim. Striking correlation, but causation is circular — engines locate where validators are; validators colocate where engines are; both follow data-centre infrastructure that predates either. Frame as equilibrium-shape, not one-way causation.

4. Characters introduced ​

No new "Meet the actor" sidebar. All major actors are returnees. Per Ch 7's pattern, Ch 8's job is to develop existing characters' Solana-specific roles in a single integrated frame.

Returning institutions, with Solana-specific roles assembled into one picture:

  • Jupiter (Chs 2, 3): aggregator-as-retail-interface; Beam as default protected routing; 93.6% of aggregator volume; 2025 cumulative ~$716B routed
  • Jito Labs (Chs 3, 5, 6): auction operator; block-engine geography; JitoBAM client; post-JIP-24 6%-to-DAO routing
  • Helius / Triton One / SOL Strategies / Figment (Chs 3, 5, 6): the four launch BAM Node operators — all also institutional validators or RPC providers
  • Harmonic (Chs 3, 5, 7): block-building marketplace; four SFDP-compliant scheduling strategies; Ben Coverston CEO-overlap with Temporal (Ch 6); 17% of stake; +101%/+39%/+36% gaps
  • Anza (Chs 5, 6): reference-client maintainer
  • Jump Crypto (Chs 5, 6): Frankendancer + Firedancer maintainer; BAM Maker Priority plugin operator
  • Rakurai (Ch 5): closed-source "no timing games" alternative; Figment migration anchor
  • Temporal (Chs 2, 6): HumidiFi operator; Nozomi (transaction submission)
  • Ellipsis Labs (Ch 2): SolFi operator; Phoenix → Phoenix Perpetuals pivot
  • Wintermute (Chs 2, 6, 7): Tessera V operator; Hyperliquid 76-market market maker; the firm whose business spans prop-AMM and CLOB architectures
  • Forward Industries / BisonFi: the one new-to-Ch-8 named institutional adopter — a US-listed company launched a Solana dark AMM in Dec 2025 that briefly led prop-AMM volume in Jan 2026. Structural significance: the regulatory-disclosure gap that prop-AMMs previously enjoyed is now compressed by public-company participation.
  • Solana Foundation (Chs 3, 5, 7): SFDP operator; soft-governance lever; ASN / data-centre concentration enforcer
  • DeezNode + Vpe (Chs 3, 4, 7): persistent validator-side grey-market case; not publicly removed from SFDP between June 2024 and May 2026

Named individuals returning (no new): Mert Mumtaz (Helius), Lucas Bruder (Jito), Anatoly Yakovenko (Solana Labs co-founder), Ben Coverston (Temporal + Harmonic), Tim Garcia (SF validator-relations), Ali Rizvi (Rakurai), Evgeny Gaevoy (Wintermute), Jakob Povšič (Harmonic + Temporal co-founder).

5. Worked example candidates ​

Three options for Phase 3's anchoring scenario. The chapter is a synthesis chapter and benefits from a worked example that integrates the layers rather than dissecting any single one.

Candidate A — "$10,000 USDC → SOL through the full Solana stack" ​

Alice (Chs 1, 3, 4) makes her $10,000 swap one more time; the chapter follows the dollar through every layer. Phantom wallet → Helius RPC → Jupiter aggregator → Jupiter Beam → Jito block engine in Frankfurt → Frankendancer Harmonic Performance validator co-located in Equinix FR5 → Solana settlement.

LayerActorTakeCumulative
WalletPhantom (Helius rebate share)~$0 to user; fraction of $0.50–$1 rebate flows to Phantom$0
RPCHelius (50/50 with wallet)~$0.50–$1 (split opaque)$1
AggregatorJupiter~5–10 bps slippage embedded via prop-AMM internalisation$6
RoutingJupiter Beam → Jito$0 to user (Beam is free in Ultra); upside is slippage avoidance$6
BundleJito tip (any backrun)6% to Jito DAOnominal
ValidatorFrankendancer Harmonic Performance+101% priority fees per block~$8 priority fee
Total$8–$15 vs Alice's $73 in 2024

The dollar shrunk vs Ch 3 (the formal-sandwich path closed), but much of Alice's spend still flows to firms she doesn't see.

  • Pros: integrates every layer; mirrors the Ch 7 Banana Gun + Titan dollar trace for Solana; lands both moves (Solana compressed the extraction surface; the structural take still runs through a handful of firms).
  • Cons: per-layer estimates are partially illustrative (Helius rebate flow not disclosed; Beam pricing not published as a take). Chapter must say so explicitly.

Candidate B — Figment's 2 March 2026 Rakurai migration ​

The empirical version of "client choice is the moat": SRR 6.85% → 7.17%, MEV ~5×, priority fees +60%, tip capture +158% at median — one software switch, no other changes. Combine with Ch 7's Harmonic gap (+33–101%) and Chorus One's operational ceiling (+3%) to land the access-vs-operational argument at chain-architecture level.

  • Pros: clean; sourced; cleanest published institutional migration of 2026.
  • Cons: not narratively vivid; risks duplicating Ch 5.

Candidate C — Geographic case study ​

A single Frankfurt-based validator (Helius's Frankfurt validator, named) across one slot: bundle arrival from the Jito Frankfurt engine (~1–2ms LAN-local), scheduler-mode choice (Harmonic FBA + Priority Fee for SFDP compliance), BAM-node sequencing (TEE-encrypted), priority-fee capture (+36% Agave Harmonic, more for Frankendancer variants). Contrast with a hypothetical stock-Agave-default validator in São Paulo with no SFDP delegation, earning at-or-below network median.

  • Pros: most differentiating new material; concrete vendors and data centres; ties +101% Performance to the 70%-Europe-plus-NA concentration.
  • Cons: per-slot dollar figures not fully public; risks technical density.

Recommendation ​

Candidate A as the primary worked example, with Candidate C's geographic frame threaded through the mechanics section. A gives the chapter a follow-the-dollar spine that mirrors the book's. C lives best as a one-page passage with a small map/diagram of block-engine-to-stake-city. B's Figment data is best deployed as a callout or dense footnote, not as the primary narrative engine — Ch 5 already developed it.

6. Open questions for Nick ​

Q1 — Word budget. Chs 5/6/7 ran ~6,500–7,500 words inc. footnotes. Ch 8 is synthesis (no new concepts), so it can be tighter. Recommend 5,500–6,500. Acceptable?

Q2 — "Pittsburgh." Not a Jito block-engine site. Actual NA mainnet: NY + Salt Lake City. Chapter will name actual sites. Flag if "Pittsburgh" came from a source I missed.

Q3 — Prop-AMM ranking update. Ch 2's HumidiFi-65% framing was mid-2025; March 2026 is competitively contested with BisonFi briefly leading and both declining >55%. Update framing to "contested." Confirm.

Q4 — Forward Industries / BisonFi as a new named character. Only Ch 8-new institutional name. Recommend short paragraph cameo — a publicly-listed US firm operating a Solana dark AMM compresses the disclosure gap. Or skip?

Q5 — Geographic treatment depth. Rich new material (Frankfurt 19% / Amsterdam 16% / London 12%; Jito engines in those cities; $200–800/mo colocation; Equinix FR5 as anchor). Recommend full subsection in mechanics. Alternative: paragraph inside the verdict.

Q6 — Alpenglow treatment. SPEC says light touch. Recommend ~300-word subsection at the end of mechanics landing the three Solana-specific shifts (vote-tx removal, VAT, ~10× threshold drop). Full architectural treatment is Ch 12.

Q7 — DeezNode / Vpe. Returning case from Chs 3, 4, 7. No new material; recommend forward-reference, not re-litigation.

Q8 — "Solana #2 by REV behind Hyperliquid in Q1 2026." Cameo in Ch 8 verdict, full treatment in Ch 9? It's one of the cleanest single anchors for Part III's chain-comparison thesis.

Q9 — Worked example. Candidate A (full-stack dollar trace) is my recommendation. Confirm, or prefer B (Figment) / C (geographic)?

Q10 — Sidebar. Per SPEC convention, no new sidebar needed (no new actor type). Alternative: a "Meet the Slot Leader" sidebar integrating validator + block constructor + matching-engine-host into one Solana-specific role definition. Recommend skip; chapter is synthesis.

Sources cited ​

All URLs accessed 2026-05-14. URLs cited inline in claims/numbers above; this is the consolidated bibliography.

Primary research and protocol documentation:

  • Syndica — Onchain Activity (March 2026, January 2026), DeFi (March 2026): blog.syndica.io
  • Helius — MEV Report, BAM, Proprietary AMM Revolution, Ecosystem Report H1 2025, Decentralization, Alpenglow: helius.dev/blog
  • Chorus One — Timing Games on Solana; Market Making, propAMMs, and Solana Execution Quality: chorus.one/reports-research
  • Figment — Rakurai migration, Q1 2026 Validator Report, Rise of Proprietary Market Makers: figment.io/insights
  • Placeholder VC — Leveling the Stakes on Solana: placeholder.vc/blog/2025/9/15/leveling-the-stakes-on-solana
  • Jito Labs / Foundation — Low Latency Transaction Send (docs.jito.wtf); Introducing BAM (bam.dev); JIP-24 / JIP-28 / JIP-31 (forum.jito.network)
  • Solana Foundation — Delegation Criteria + Delegation Program (solana.org)
  • SIMDs — 0326 (Alpenglow), 0357 (VAT): github.com/solana-foundation/solana-improvement-documents
  • Harmonic — Scheduling Strategies (docs.harmonic.gg); Jupiter — Ultra docs (developers.jup.ag)

News and analytical coverage:

  • SolanaFloor — Jupiter 93.6%; 70%-aggregator-routed; BisonFi vs HumidiFi; Forward Industries BisonFi: solanafloor.com
  • DL News — Temporal/HumidiFi; Solana dark AMMs / Tessera V / Wintermute confirmation; HumidiFi $34bn record: dlnews.com
  • CoinDesk — June 2024 Foundation enforcement; Jito launches BAM (Jul 2025); Alpenglow testnet (11 May 2026): coindesk.com
  • PRNewswire — BAM launches; Rakurai $3M seed: prnewswire.com
  • Edgen — 8 April 2026 patch: edgen.tech
  • Blockworks — Foundation pruning; REV dashboard (blockworks.com analytics): blockworks.co / blockworks.com
  • Phemex — Solana Foundation May 2026 requirements: phemex.com
  • PANews — Solana Q1 2026 report: panewslab.com
  • Chainflow — Validator Discussions Feb 2026; Constellation: chainflowsol.substack.com / chainflow.io
  • RPC Fast — Co-location Strategies; MEV Infrastructure: rpcfast.com
  • The Good Shell — Solana Validator Cost: thegoodshell.com
  • Tokenomics.com — Jito Tokenomics
  • Anarcaze (Medium) — Spotting Shady Validators
  • bit2me — REV metric explainer
  • Blocmates — BAM article
  • DefiLlama — Phoenix, Manifest, Jito

Pulled through from prior chapters' RESEARCH.md (URLs in those files, not re-listed): Wu et al. arXiv 2405.01329 (Ethereum builder concentration, cross-chain comp); Wahrstätter et al. on private order flow (cross-chain comp); Ellipsis Labs — Phoenix Perpetuals; Anza central scheduler docs; June 2024 enforcement framing (Ch 3); Helius wallet customer roster + 50/50 rebate (Chs 6, 7); Bitwise Hyperliquid ETF amendment (Ch 5, 7); relayscan.io builder/relay shares (Chs 5, 6, 7).

Phase 1 is complete. Per the user's compressed-review pattern, Phase 2 (OUTLINE.md) follows immediately after Nick's review of this RESEARCH.md.