Ascend tech readiness — two stages to mainnet

Ascend runs end-to-end on a public testnet today. The system was built in-house; OpenZeppelin is the security and audit partner. A working testnet is not yet a release candidate. Two stages separate today from mainnet in early 2027: Stage A — five internal readiness gaps, no external party needed — produces the release candidate; Stage B — three externally procurable verification legs — takes it to mainnet. Each external leg maps one-to-one to a part of the scope-of-work document now going to vendors, so incoming quotes land apples-to-apples against the cost cards on this page. How the system got here — what shipped, when, and by whom — is charted in the development history below.

Ascend has been paying for platform development since about January 2026.^[17] The chart shows what shipped, when, and by whom — from the OpenZeppelin engagement through the June 9 decision to build in-house, to today. The repository record shows the shape plainly: no shipped code in January, OpenZeppelin's first feature commits in early February, and an MVP-grade private beta by late May. The in-house era opens June 8; a full end-to-end stack was live on the Ascend testnet by June 11 and has grown into today's full-featured private beta. Every dated element carries its provenance; single-source and asserted dates are tagged.

Marcelo (CSO) works the commercial and strategy track — the OpenZeppelin commercial relationship and partner engagements. That stream ships no code, so he carries no bar color and is not folded into any other contributor; he is part of the June 9 build-in-house decision milestone. Lane colors identify contributors only — they carry no status or performance meaning.

Bars are feature groups colored by contributor; widths are calendar spans from the repository record, not effort. The right edge of this chart hands off to the 'Now' node of the roadmap in the next section.

Contribution mix — CPO estimate^[20]

The projected row is normalized at the low end of Dennis's 40–50% band; if Dennis lands higher, the difference comes out of Manny's share. Manny grows with the SDD gauntlet (+10 pts) and auth/wallets/onboarding (+10 pts) — the same Stage A items, so the projection inherits Stage A risk.

The release candidate is the hard boundary: everything to its left is Ascend's own work; everything to its right verifies and remediates the candidate — it does not add features. Two calendar facts drive procurement: the Q4 2026 closed beta overlaps Leg 1, and top audit firms book out six to eight weeks^[7] — so Ascend must reserve the Leg 3 audit slot around September, while Stage A is still running (conditional on Stage A effort estimates firming by end-August).

Five gaps stand between today's testnet and the release candidate — all Ascend-internal; none needs an external services firm (wallet infrastructure consumes third-party wallet providers as components, decided and integrated internally). Internal cost is the team's own effort and calendar time, estimated where known and marked pending where not. Rank one is the SDD gauntlet: it turns Leg 1 from open-ended vendor discovery into a defined, quotable engagement. It outranks CI/CD but depends on it in sequence — the diagram shows both.

Stage B is the verification program: three separate engagements, one per part of the outbound scope-of-work document, so every vendor quotes the same scope. The leg cards carry each engagement's full scope; the matrix beneath separates four often-conflated security terms.

Price signals track crypto-native top-tier firms only — the calibre Ascend already works with. Illustrative figures are internal calibration, not quotes; the pipeline below replaces them with real numbers as they land. Leg 3's ~$200K ballpark^[8] presumes mid-band researcher-week staffing; at OpenZeppelin or Trail of Bits list rates the same scope prices at or above the public band's $300K+ top — the equivalent-spend chart shows what the same dollars buy at each firm.

Vendor scoping and delivery verification are covered today: the CPO runs both as interim owner. What a part-time interim owner cannot supply is full-time engineering execution — and that is the hire: a hands-on director of engineering (senior-staff level). The CEO (Dennis) currently carries the engineering build alongside fundraising, and the named Stage A gaps — CI/CD + IaC, the API gateway + SDK — are exactly the workload the hire absorbs. The hire frees the CPO to run product end-to-end (strategy, technical, marketing) rather than splitting into engineering execution, and gives Ascend an engineering skillset that meets external vendors apples-to-apples: rate-card fluency, scope pushback, billed weeks verified against delivered work. It is also the continuity role — one accountable engineering owner through hardening, audit remediation, and post-mainnet operations (incident response, upgrades) — a role institutional counterparties expect to exist. Spend shape matters as much as rate: the current contract bills monthly for as long as it runs; Stage B is a bounded package delivered in about a quarter.

Every firm quotes against the same scope-of-work document — the one whose Parts I–III define the three legs — so quotes land apples-to-apples against the cost cards above.

Every figure on this page carries its provenance: a named document, a public benchmark with its URL, the repository record (commit history pulled July 28–29, 2026), a dated meeting record, "internal technical review, July 27 2026 — single source," or an owner-asserted anchor labeled as such.

Glossary

Footnotes

1. [1]Contract count: the July 1, 2026 scope-of-work document states "70+ smart contracts" (document-verified — a figure matching the repository's June 8 day-one file count: repository size, not audit scope). The July 27, 2026 internal technical review cited 71 — single source. Direct repository count at the remote head, July 29, 2026: 162 source contract files (145 concrete contract declarations), excluding mocks and tests; the earlier 115/101 figure came from a stale June 25 local mirror. Body copy uses "145+"; audit vendors price on nSLOC counted at quote time — that count runs against the current ~162-file contract set, not the "70+" label.^
2. [2]Wallet vendors — decision not made. BitGo is offering services; the CPO's position is that all user and admin wallets (everything except superadmin wallets) require self-custody and heavy smart-contract interaction, so the model to follow is Grvt-like: Privy as primary, Dfns as fallback. The superadmin-class institutional custody layer can sit with BitGo-class custody.^
3. [3]OpenZeppelin public list rate: $25,000 per engineer-week, matched at the top of the market only by Trail of Bits (same rate, same source). Public benchmark: 7blocklabs 2026 audit-cost roundup — https://www.7blocklabs.com/blog/smart-contract-audit-cost-range-2026-and-trail-of-bits-smart-contract-audit-cost-benchmarks^
4. [4]Spearbit (via Cantina): $32.5K–$48K per team-week is the published figure, for teams of 3–5 researchers (per the source page). Straight division yields ≈$6.5K–$16K per researcher-week; the $7K–$10K this brief uses is a rounded mid-band derivation within that division — presented as derived, not quoted. The equivalent-spend bar (20–28 researcher-weeks per $200K) is computed at that $7K–$10K mid-band. Same source as [3].^
5. [5]Quantstamp: $13,000 per audit-week ($325/hour); unused weeks carry over. Public retainer proposal, Venus forum, January 2024 — https://community.venus.io/t/quantstamp-retainer-proposal-for-ongoing-audits/4083^
6. [6]$150K–$300K+ band for large interconnected systems: composite of two Sherlock posts — "routinely exceed $150,000" (https://sherlock.xyz/post/smart-contract-audit-pricing-a-market-reference-for-2026) and "$300,000 or more" (https://sherlock.xyz/post/smart-contract-audit-cost). Cite both if challenged.^
7. [7]"Most reputable auditors book out 6–8 weeks in advance" — Sherlock audit-cost post: https://sherlock.xyz/post/smart-contract-audit-cost (note: this claim lives on the audit-cost post, not the market-reference post).^
8. [8]~$200K audit ballpark — ILLUSTRATIVE: internal technical review, July 27 2026 (single source), corroborated by staffing math against public MID-BAND rates only (Spearbit researcher-weeks[4], Quantstamp audit-weeks[5]). The corroboration does not hold at the top of the list: at OpenZeppelin / Trail of Bits rates ($25K/engineer-week[3]) the same scope prices at or above the $300K+ top of the public band[6]. Not a vendor quote; the pipeline produces the real number. Scale calibration from received proposals (proposal — internal calibration only; firms never named): two independent reviewed proposals for a ~15–17-contract, largely standard ERC-3643 stack each carried a $10K–$30K third-party single-pass audit estimate — one noting complex multi-contract audits with custom compliance modules exceed that range, the other excluding the audit from the development quote and billing it separately, corroborating the audit-carved-out-as-its-own-line quote structure. A small-stack floor showing why nSLOC × complexity drives the price — not a comparable for Ascend's ~115-file protocol.^
9. [9]~$300K–$400K package envelope over ≈ one quarter — ILLUSTRATIVE roll-up of Legs 1–3 plus remediation re-reviews; external Stage B spend only (Stage A internal effort and the dedicated-engineer hire are separate). The $300K floor presumes Leg 3 landing at or below its ~$200K mid-band center with Legs 1–2 near their floors; the spread above the leg-card floors is the remediation-re-review allowance; a top-of-list Leg 3 quote pushes the envelope above the stated range. Superseded by pipeline quotes as they land.^
10. [10]$113K/month current contract rate; $75K/month OZ counter; $900K/year at the counter rate — internal technical review, July 27 2026 — single source; verify against the signed contract.^
11. [11]Internal senior engineer $120K–$200K/year: public compensation data (Cyfrin smart-contract security-auditor salary guide — https://www.cyfrin.io/blog/smart-contract-security-auditor-salary). ≈$2.3K–$3.8K per engineer-week is annual cost ÷ 52 — salary-derived, excluding benefits and overhead; a fully-loaded figure (~1.25–1.4×) would run ≈$2.9K–$5.4K per week. The internal-vs-external comparison holds under either basis.^
12. [12]$5.3K/engineer-week development-team benchmark (proposal — internal calibration only): the floor of the $5.3K–$13K/week benchmark range is a received-proposal rate — a reviewed fixed-scope proposal from an established firm blends a full senior engineering team to ≈$5,335 per FTE-week ($350K ÷ 65.6 FTE-weeks, ~$133/hour); credible fixed-scope proposals from established firms price full senior engineering teams at roughly $5,300 per engineer-week. The received rate calibrates the rate basis only, never a leg-level price. Never attributed to any firm, including firms named in the quote pipeline.^
13. [13]Generalist web-app/API penetration-test band $10K–$30K (Astra public cost guide — https://www.getastra.com/blog/security-audit/penetration-testing-cost/). A generalist FLOOR for the pen-test component only: crypto-native scope with wallet and onchain integrations prices above it, and red teaming is quoted as a separate line above the floor.^
14. [14]Hardening staffing shape, 1–2 senior engineers × 4–6 weeks — internal technical review, July 27 2026 — single source. Effort-weeks and calendar-weeks are different units: with testers in parallel, Leg 1's calendar target is roughly one month. The ≈$25K–$150K illustrative range is this staffing shape priced at $5.3K–$13K/week benchmark rates[12][4][5].^
15. [15]SDD gauntlet effort, 3–4 man-weeks — CPO estimate, July 2026.^
16. [16]Timeline (final engineering handover within 1–2 weeks; Q4 2026 closed beta, late-October window; beta through January; mainnet early 2027) — internal technical review, July 27 2026 — single source. The "RC target ~late Oct 2026" label is inferred from the beta window, not separately stated in the review — tagged "inferred" on the timeline.^
17. [17]Engagement payment start ~January 2026 — asserted (CPO, July 29 2026); consistent with the current $113K/month contract rate[10]. Repository record: zero commits dated January 2026 in either OpenZeppelin repo; first feature commits February 4 and contracts-repo creation February 9, 2026 (commit history pulled July 28, 2026). A single scaffold commit exists on December 8, 2025, before the chart window.^
18. [18]CPO start — asserted late March 2026 (CPO, July 29 2026); the documented record shows internal reviews of OpenZeppelin's output from March 2, a written product review sent ~March 10, and the first direct OZ engineering sync with the CPO attending on March 17, 2026 (meeting record). The chart marks the documented March 17 sync; the product-review bar starts ~March 10 (estimated).^
19. [19]The June 8 – July 2 audit bracket covers OpenZeppelin's audit of ITS OWN beta codebase (code freeze June 8; first-cycle results June 30: 0 critical / 1 high / 1 medium / 5 low / 3 informational, fixes submitted; report delivered July 2, release v0.1 merged July 3 — repository record + meeting records). That codebase was set aside at the June 9 pivot. The in-house stack shipping today has not yet been audited — that audit is Stage B, Leg 3. OZ commit activity tapers after July 3 (last commits July 22–23); the chart renders the taper, with development end at the July 2 report. The bar's span also includes the late-May pre-freeze beta wrap (Sepolia environment prep ahead of the June 8 code freeze), before the audit window.^
20. [20]Contribution mix — CPO estimate, single source (July 29 2026). Commit counts measure volume, not contribution weight (different codebases, squash merges, brand vs protocol not comparable); for reference only: AscendFi org 1,270 default-branch commits through July 28 (Dennis 968, Renée 209, Manny 93), OpenZeppelin repos a separate 415 excluding bots. The projected bar assumes the SDD gauntlet (+10 pts) and auth/wallets/onboarding (+10 pts) land as planned — the same items as Stage A gaps 1 and 3.^
21. [21]Commit split per month, default branches, excluding CI bots (history pulled July 28–29, 2026). OZ series: the two OpenZeppelin repos (ascend-contracts + ascend-platform). Ascend series: all 17 AscendFi-org repos, including pre-consolidation ascend-site history — the May–June brand-site work by Manny/Renée predates the June 22 org consolidation. OZ + Ascend = total: Jan 0+0=0 · Feb 39+0=39 · Mar 63+0=63 · Apr 59+0=59 · May 73+97=170 · Jun 116+518=634 · Jul 23+655=678 (July partial, through July 28). Counts are volume, not contribution weight.^

Links out: the SDD gauntlet explainer (in Stage A). Footnote URLs are citations only; all page assets are inline.