1. Executive summary

The proposition

Capacify is, in our assessment, the single best-positioned operator in North America to enter the institutional compute derivatives market. You already sit on both sides of the price-discovery problem: routing demand across heterogeneous supply, with line-of-sight to forward workload curves on the buy side and forward capacity commitments on the sell side. What you do not yet have is the financial-markets architecture to monetise that information asymmetry: a forward curve methodology, a benchmark approach, contract specifications, a regulatory pathway, a settlement and clearing design, and an operating model.

Capacity Derivatives has developed comprehensive intellectual IP across exactly this architecture: contract design, distributed ledger settlement (Canton), benchmark methodology, regulatory analysis, operating model, and financial framework. The work is documented at capacityderivatives.com and is currently advancing toward operational launch with support sought from the Canton accelerator programme. At this stage the value of the work is intellectual: a coherent, internally consistent body of design across all the domains Capacify would otherwise have to develop from scratch.

Outline proposal

This proposal is for an 8-week engagement, fixed-fee at £125,000, in which Capacity Derivatives delivers to Capacify a complete operational blueprint for a full-scope compute derivatives brokerage, contextualised to Capacify's specific strategic situation, to equip Capacify's board to make a go / no-go decision on entering the market.

A full-scope blueprint, in this context, means everything needed to stand up the business: market structure and counterparty universe; product architecture and contract specifications; benchmark, reference data, and pricing methodology (curve construction, volatility model, calibration); the technology, settlement, and clearing stack; the regulatory pathway and legal framework (ISDA architecture, jurisdictional agreements, client documentation, KYC/AML); organisation design (roles, headcount, skills, governance); the brokerage's own pricing model (take rates, market-making spreads, fee structure); a driver-based financial model with scenarios; the full risk framework; and a sequenced implementation roadmap to first cleared trade.

Exclusions

The blueprint is built on pre-existing intellectual IP of the Capacity Derivatives team, developed in advance of and independent of this engagement, currently advancing toward operational launch with support sought from the Canton accelerator programme. The IP being shared is not the bespoke output of a research commission but mature work that Capacity Derivatives has already done for its own venture.

Capacify does not acquire ownership of the IP, exclusive access to it, proprietary rights over it, or preferential treatment in any subsequent commercial relationship with Capacity Derivatives. Capacity Derivatives is and will remain free to share the same IP, methodology, and architecture with other counterparties in the compute ecosystem, including direct and indirect competitors of Capacify, and to operate its own venture in parallel.

Future considerations

If Capacify, having received the blueprint, wishes to negotiate any of the following: exclusivity, operational partnership or integration with Capacity Derivatives' venture, a licence to operationalise the blueprint in a market-facing offering, white-label or branded versions, deeper advisory support through a build phase, or a joint venture, these are explicitly available and would be negotiated under a separate, subsequent commission with separately determined commercial terms. None of these are pre-committed by this engagement.

Relative value of this proposition

The deliverable will be the equivalent of a report commissioned by McKinsey, Bain or BCG (MBB). Because the Capacity Derivatives founders have been building this business, not researching it from the outside, the work will have superior grounding in the reality of setting up and running a derivatives brokerage.

An MBB (or specialist financial-services firm such as Oliver Wyman or EY-Parthenon) engagement of this scope would typically run £400,000 to £1.2m over 12 to 16 weeks, delivered by three to five consultants who had likely never traded a derivative before mobilisation. The output would be comprehensive at the level of what to consider but predictably thinner at the level of what to build: contract specifications, ISDA architecture, settlement design, benchmark methodology, and pricing model would be deferred to "specialist counsel in implementation". That specialist counsel would be the Capacity Derivatives team.

This £125k engagement covers the same scope at greater depth, in less time, because the work an MBB team would have to do from scratch has already been done by Capacity Derivatives for its own venture.

2. Our understanding of Capacify

Strategic positioning

Capacify is building a full-stack, accelerator-native AI cloud anchored on Canadian land and power, with three commercial surfaces in market today:

• Compute sourcing and optimisation: multi-supplier execution for customers who want to avoid lock-in and capacity guesswork.
• Workload-aware routing: matching workloads to infrastructure on cost, latency, reliability, and performance.
• Data centre evaluation and readiness: site-level assessment for buyers and suppliers.

The strategic direction is upward integration from bare-metal infrastructure into orchestration, managed services, and consumption-based products. The customer base spans AI natives, top-tier startups, and the Global 1000. The supplier and partner network spans hyperscalers, NeoClouds, data centre operators, OEMs, system integrators, and colocation brokers.

Four observations frame this proposal

What Capacify does not yet have, and what would otherwise take 18 to 24 months to build internally, is the capital-markets architecture: contract specifications, reference-rate methodology, clearing and settlement design, regulatory perimeter, capital requirements, ISDA framework, and market-making operating model.

3. The opportunity: compute derivatives

Structural analogue: oil

The compute market today has the structural characteristics of every commodity market on the eve of financialisation. The closest and most instructive analogue is oil.

Oil derivatives are the deepest commodity derivatives market in the world, with a derivatives-to-physical ratio of approximately 30x. The structural mapping onto compute is near line-by-line:

The broader commodity-financialisation arc, coal, iron, cotton in the late 1800s; oil and gas through the 20th century; electricity, freight (Baltic indices), iron ore, LNG, and carbon in the modern era, shows the same pattern in each case. The operator that owned the reference price and the counterparty network captured a disproportionate share of the financial layer that grew on top of the physical one.

The compute opportunity is structurally larger and arriving faster than any of these. The current market is $200 billion+ with triple-digit growth. H100 one-year rental prices rose 40% in five months (from $1.70/hr to $2.35/hr) without an institutional venue to hedge. On-demand capacity is sold out across all major types, with Blackwell-generation capacity fully booked through late 2026. Even a 5% derivatives-to-physical ratio (electricity is over 10x, oil over 30x) implies a multi-tens-of-billions notional market by the late 2020s.

Instruments and unit of trade

The unit of trade is GPU-time (GPU-hours) by accelerator class and region, not physical chip delivery. This is structurally identical to oil (barrels of a defined grade at a defined delivery point), electricity (MWh at a defined node), and freight (tonne-miles on a defined route). The physical hardware sits underneath the financial layer; the derivatives reference the service it produces.

The contract suite, in the order conventional commodity-derivatives markets adopt them (OTC to exchange-listed):

• Fixed-for-floating swaps. Buyer pays a fixed GPU-hour rate; seller pays the floating index rate. The standard entry point in commodity OTC markets and the typical first liquid instrument.
• Cash-settled forwards. Agreement to buy or sell GPU capacity at a fixed price on a future date, cash-settled against the published benchmark at expiry.
• Asian-style monthly settlement. Settlement on the arithmetic average of daily index values over the contract period, matching how compute is consumed as a continuous flow.
• Exchange-listed futures. Standardised contracts once OTC convention is set; the second phase of every commodity market.
• Options and structured products. Capacity swaps, range accruals, price floors, and products linking compute price to power, PPAs, or model-specific FLOP requirements.

The technical and regulatory architecture is now buildable on commercially available infrastructure: Canton-based DLT settlement with Daml smart contracts modelling the full ISDA lifecycle, voice brokerage at launch following the model that seeded every major commodity OTC market, and clearing via an existing CCP or a purpose-built mutualisation structure. The constraint is no longer technical; it is commercial and regulatory.

4. Why now

The convention-setting window

Four forces converge in the next 12 months:

The structural opportunity remains large enough to support multiple operators, as oil, electricity, and freight all do. But it is no longer vacant, and the operator that lands the second-credible offering against entrenched first-mover convention faces a materially harder commercial path than the operator that lands an offering during the convention-setting window.

Capacify's structural advantages, bilateral routing data, supplier and buyer network, Canadian jurisdiction, NCP-tier Nvidia relationship, are precisely the advantages that current first movers do not have. The window to convert those advantages into a defensible derivatives operation is immediate.

5. Why Capacify, why Capacity Derivatives

Fit on both sides

6. Engagement objectives

What the board can decide

At the end of the engagement, Capacify's board will be equipped to make decisions on the following:

7. Scope

Ten domains

The blueprint comprises ten domains. The depth in each is calibrated to operational quality, not to a slim feasibility-report standard. A Capacify board could in principle vote to execute against the package; Capacify operators could in principle build the business from it.

The work in each domain comprises two streams. Pre-existing blueprint: Capacity Derivatives' intellectual IP across the domain. This is the bulk of the substance and the basis for the £125k fee. The IP remains the property of Rubytech LLC; Capacify receives a perpetual licence for internal decision-making use (see section 11). Contextualisation: targeted work specific to Capacify's situation, their routing data, their existing team, their Canadian operating venue, their customer and supplier network, their strategic ambition.

7.1 Market, demand, counterparty universe

• Compute price volatility decomposition by accelerator class, region, and time horizon.
• Hedging demand: bottom-up segmentation of buyers (AI labs, enterprise, sovereign, financial) and sellers (hyperscalers, NeoClouds, DC operators, sovereign-aligned capacity).
• Willingness-to-pay and notional capacity by segment.
• Adjacent market analogues (oil, electricity, freight, iron ore, LNG, carbon).
• TAM, SAM, and SOM with explicit assumptions.
• Market-maker requirements and likely initial liquidity providers.
• Convention-setting strategy for the first 18 months.
• Onboarding, KYC, and credit-risk framework.

7.2 Product architecture

• Contract specifications for forwards, futures, options, and swaps, to specification-grade detail.
• Underlying definitions by accelerator class, region, contract month.
• Settlement design (cash-settled against benchmark, Asian-style monthly).
• Contract lifecycle: creation, novation, exercise, settlement, margining.
• ISDA-aligned documentation architecture.

7.3 Quantitative methodology and pricing

• Benchmark and reference-rate construction (transaction-based, manipulation-resistant).
• Forward curve construction across tenor and accelerator class.
• Volatility model design, drawing on Joel Smalley's Volman / ITO33 calibration background.
• Option pricing methodology for the structured-products phase.
• Backtesting against available historical compute price data.
• Index governance and oversight design.

7.4 Brokerage commercial model

• Take rate and fee structure across product types and counterparty tiers.
• Market-making spread design and inventory model.
• Voice-brokerage commercial economics.
• Volume and revenue ramp by phase.
• Unit economics per trade, per counterparty, per product.
• Pricing posture against competitors (Ornn and emerging entrants).

7.5 Technology, data, clearing

• Reference architecture for matching, clearing, settlement, custody, reporting, and data distribution.
• Canton DLT design with Daml smart contracts modelling the full ISDA contract lifecycle (creation, novation, exercise, settlement, margining), advanced through the Canton accelerator application, adapted for the Capacify operating context.
• Leverage of the existing Canton institutional consortium (Goldman Sachs, BNP Paribas, Deloitte, Microsoft) where their prior technical, legal, and compliance assessment of the infrastructure shortens Capacify's approval path.
• Data architecture: transaction data, reference data, benchmark data, regulatory reporting data, client data.
• Build, partner, or acquire across each architectural layer.
• Integration with Capacify's existing routing, telemetry, and customer infrastructure.
• Operational resilience and cyber posture (DORA-adjacent, OSFI E-21).

7.6 Regulatory and legal framework

• Canadian (CSA, OSC, CIRO), US (CFTC, SEC), UK (FCA), and EU (MiCA, MiFIR-adjacent) analysis.
• Recommendation on jurisdiction of incorporation and primary registration.
• Capital and conduct requirements under the recommended structure.
• Cross-border passporting and recognition pathways.
• Timeline and cost of regulatory permissioning.
• ISDA framework: master agreements, schedules, credit support annexes, definitions.
• Client documentation: onboarding, terms of business, risk disclosure, KYC/AML, suitability.
• Trade documentation: confirmations, term sheets, novation deeds.
• Inter-affiliate, brokerage, and clearing agreements.

7.7 Organisation and people

• Functional design: front office (sales, market making, structuring), middle office (risk, surveillance), operations, technology, finance, compliance, legal.
• Org chart and reporting lines.
• Headcount and skills profile, year 1 through year 3, with role descriptions for the first ten hires.
• Build, hire, or partner across each function, including which Capacify roles can extend, which need fresh hires, and which can be partnered or outsourced.
• Compensation framework appropriate to financial-markets recruitment.
• Governance: board, risk committee, oversight bodies, committee mandates.

7.8 Financial model

• Driver-based 5-year P&L, balance sheet, capital, and cash flow model.
• Base, downside, and upside scenarios.
• Capital requirement: regulatory minimum plus working capital plus mutualisation buffer.
• Returns: NPV, IRR, payback, run-rate margin.
• Sensitivity analysis on volume, take rate, capital cost, and regulatory timeline.

7.9 Risk framework

• Risk taxonomy across strategic, financial, operational, regulatory, conduct, reputational, technology, and model dimensions.
• Risk register with likelihood, impact, mitigation, residual rating.
• Compute-specific risks: reference-rate manipulation, capacity gaming, supplier concentration, Nvidia allocation shocks.
• Risk appetite framework and limits structure.

7.10 Implementation roadmap

• 6 to 18-month plan to first cleared trade.
• Critical path with decision gates and exit ramps.
• Phasing of regulatory, technology, commercial, and hiring workstreams.
• Investment profile by phase.
• Quick wins and early proof points.

8. Approach and methodology

Four parallel workstreams

The engagement is structured as a compressed blueprint-delivery and contextualisation cycle, run as four parallel workstreams over eight weeks. Because the underlying IP is pre-existing and comprehensive, the emphasis is on applying and adapting the blueprint to Capacify's situation, not on building it from scratch.

Methodology blends:

• Blueprint delivery: structured walk-throughs of Capacity Derivatives' pre-existing work product across each domain, with Capacify principals and nominated team members. Substance, not summary.
• Contextualisation: targeted application of the blueprint to Capacify's specific routing data (anonymised), customer and supplier network, Canadian operating venue, and strategic ambition.
• Targeted primary research: limited, high-leverage interviews (typically 5 to 10) where Capacify's network provides access that materially sharpens a specific decision.
• Scenario modelling: Capacify-specific financial scenarios applied to the existing financial framework.

Four steering committee gates (end of weeks 1, 3, 5, and 7) bring Capacify principals into the engagement at the points where their input most shapes the final recommendation.

9. Workplan and timeline

Eight weeks

Steering committee attendance: the four Capacify principals plus any nominated advisors. Format and duration adjustable. Weekly written progress notes between steering committees.

10. Team and credentials

Four founders, no bench

The engagement is delivered by the four Capacity Derivatives founders, working directly. There is no consulting bench, no analyst layer, and no escalation chain.

11. Deliverables and licence terms

What you receive, on what terms

11.1 Deliverables

11.2 Licence terms

All deliverables, and the underlying IP they reflect, remain the intellectual property of Capacity Derivatives. On final payment, Capacify is granted a licence to the deliverables on the following terms:

• Permitted use: internal strategic decision-making by Capacify regarding entry into the compute derivatives market, including circulation to Capacify directors, officers, employees, and named professional advisors under standard confidentiality obligations.
• Term: perpetual. The licence does not lapse.
• Exclusivity: none. Capacity Derivatives is free to share, licence, or operationalise the same IP with any other counterparty, including direct and indirect competitors of Capacify.
• Preference: none. Capacify acquires no right of first refusal, no most-favoured-nation provision, no preferred-partner status, and no preferential commercial terms in any subsequent Capacity Derivatives engagement.
• Operational use: not permitted. The licence does not extend to operationalising the IP in a market-facing venture, white-labelling, public distribution, or any commercial application beyond Capacify's internal decision-making.
• Onward transfer: not permitted. The deliverables may not be transferred, sold, sub-licensed, or assigned to any third party without written consent.

11.3 Subsequent engagement options

Each of the following, if Capacify wishes to negotiate it, would be the subject of a separate, subsequently commissioned engagement with separately determined commercial terms: exclusivity over a defined scope and duration; a licence to operationalise the IP in a Capacify-branded market-facing venture; a white-label or co-branded version of Capacity Derivatives' venture; operational partnership or integration between Capacify and Capacity Derivatives' venture; joint venture, equity participation, or co-investment structures; continuing advisory support through a Capacify build phase; bespoke development of additional IP specific to Capacify's situation. None of the above is pre-committed by this engagement, and none is foreclosed.

12. Commercial terms

Fixed fee, milestone payment

13. Disclosures, assumptions, dependencies, exclusions

Stated plainly

Disclosures

Capacity Derivatives is a venture under development in the compute derivatives market. The team is advancing toward operational launch of an OTC voice-brokered compute derivatives offering, settling on Canton Network, partnering with existing benchmark providers, with support sought from the Canton accelerator programme. The venture is at the design and intellectual IP stage; it is not yet operational. This is disclosed transparently and is the principal credential for this engagement: the IP being shared is the same IP underpinning a serious venture being built by the four founders, not the speculation of pure advisors.

The IP being shared is pre-existing. All architecture, methodology, contract specifications, settlement design, regulatory analysis, organisation design, quantitative methodology, and financial modelling that forms the substance of this engagement was developed by Capacity Derivatives prior to Capacify's commission and would have been developed regardless of whether Capacify engaged. The engagement is a contextualised delivery of that pre-existing work to Capacify, not a commissioned research programme with bespoke IP transfer. Licence terms in section 11.

Capacity Derivatives is and will remain engaged with multiple counterparties across the compute derivatives ecosystem, including potentially competing AI infrastructure operators, hyperscalers, NeoClouds, data centre operators, financial market infrastructure providers, exchanges, and clearing venues. Capacify acquires no right to constrain those engagements. Should Capacify wish to negotiate exclusivity covering a defined scope and duration, that is available as a subsequent commission (section 11.3).

Good faith. For the duration of the engagement and a 12-month tail thereafter, Capacity Derivatives commits to operating in good faith on Capacify's interests in respect of the engagement's subject matter. Deliverables will reflect our best current understanding; we will flag material developments in the market or in our own venture that would change the recommendation; and we will not use Capacify's confidential information for any purpose other than this engagement.

Further exclusions

• Legal drafting of definitive contracts, ISDA negotiation, or regulatory filings.
• Vendor selection, technology procurement, or build execution.
• Capital raising, investor pitching, or M&A advisory.
• Customer acquisition or supplier negotiation.
• Implementation phase support (available under a subsequent commission per section 11.3).
• Any operational use, white-label, exclusivity, or partnership rights (available under a subsequent commission per section 11.3).

14. Next steps

To engagement letter

Signatures

The undersigned founders of Capacity Derivatives propose the engagement set out in this document on the terms herein. Execution of the engagement letter by Capacify will confirm acceptance. Signature blocks for Joel Smalley, Mark Abbott, David Gibbs and Nick Andrews, with acceptance by Capacify Inc. through Imtiaz Jaffer, Anik Nagpal, Saif Ajani and Minaz Abdulla, are on the final page of the PDF.

What the market has done

The compute derivatives market now exists on the record, with dates against it, as set out below.

This is validation that compute would end up hedged like a commodity, and the largest exchanges in the world have now said so in public. It is a better position for us. What is left to settle is where we sit now that they are here.

The layer that is open

There are four layers in this market, and only one of them is unoccupied.

• The index is the price. Compute Desk runs the only GPU compute indices distributed on Bloomberg, across 350,000 terminals. Ornn and Silicon Data are the other two index franchises.
• The exchange is the listed, cleared contract. Standardised sizes and tenors, and you have to be a clearing member to touch it.
• Physical delivery turns a futures position into actual GPUs.

This matters because of what a hedger actually needs. Listed contracts come in fixed sizes and fixed tenors, and almost nobody's real exposure lands neatly on one of them. A neocloud wants seven months rather than a calendar quarter, a lab wants a structure rather than a flat future, and someone else wants a size that is nowhere near a round lot. None of that trades on an exchange. It trades over the counter, arranged by a broker working off the published index, and in every commodity market that has ever developed, the OTC book ends up larger than the listed one.

To say it directly, this is not a claim to be better than CME. The claim is about the layer they do not serve, and about who arranges it.

Who builds the desk

David Gibbs sets the desk up. That is his role at Capacity Derivatives and has been from the start: the structure, the people, and the counterparty relationships that let it start trading in London. He brings in a market maker when the book is ready for one. Once the desk runs, the people he has brought in run it. Where and how that team sits is one of the questions the discovery work settles.

The desk opens as a broking operation: matching a buyer to a seller off the published index, earning on the spread or a fee, with no position taken on. That is what builds the market up to a place where there is enough liquidity to warrant a market maker, and a market maker is one of the first hires after that, once the flow justifies the capital and the risk that sit behind it.

This is the part that cannot be bought in. Capital can be raised and permissions can be applied for, both on a known timetable and at a known cost. Setting a desk up is the other thing, and it is done by somebody who has set one up before. David ran equity derivatives trading at Barclays and at Dresdner Kleinwort, and was managing partner of a brokerage he later sold, so he has built this before and he holds the relationships across the dealer and brokerage community that a new desk needs from day one. That is the scarcest aspect in the whole proposition, and we already have it. The desk today is David, and the team is hired with the money being raised. Every venture works that way.

Capacify.ai is the other side of that. It brings the supply and the buy side, private routing data on where GPU demand actually goes, and access to Nvidia hardware at cloud-provider tier. Today it earns on the flow it moves, deal by deal, at a price somebody else sets. A desk changes what that flow is worth to it. The same customers become positions that can be hedged and carried, not bookings that clear and are gone. The routing data becomes what the price is struck from instead of a by-product of running the business, and the margin sits with the party pricing the risk rather than the party passing it on. That is worth more to Capacify.ai than the flow on its own. What Capacify.ai takes for it is a commercial question, settled by a term sheet alongside the discovery work, and it is not settled today.

Our proposed settlement rails

An OTC desk is only as good as the rails it settles on. The ledger is Canton, a blockchain built by Digital Asset for institutional finance. Goldman Sachs was a founding participant at its launch in May 2023 and joined the Canton Foundation, its governance body, in March 2025. BNP Paribas was also a founding participant and joined the Foundation in September 2025. DTCC and Euroclear co-chair the Foundation.

The strongest point is a property of the protocol itself, set out in the documentation. Canton does delivery against payment with atomic settlement: both assets move in a single transaction that either commits in full or does not happen at all. There is no state in which one party has paid and not received.

Privacy comes second. The design gives what Canton calls sub-transaction privacy: each party sees only the parts of a transaction it is party to, and the operators of the network's timing service see only encrypted messages. For a desk that is not a nicety. On a public chain your entire flow is readable by anyone, including the counterparty on the other side of the trade you are arranging.

Operating 24/7/365 comes third. Compute prices move continuously, and so does anyone's exposure to them. Clearing houses run margin cycles and settlement windows, so a move against a counterparty on a Sunday morning is dealt with on Monday. Continuous marking and margining is what a desk needs the day it starts carrying positions of its own, and it is what lets it hold counterparties it would otherwise refuse over a weekend. That is what our design and proposal looks to build and deploy. It is not a service running today.

Named executives at Digital Asset, Franklin Templeton and Virtu Financial have spoken publicly about always-on infrastructure and moving free of traditional market hours, so people in the market are pushing that way, but no derivatives margining service runs 24/7/365 on Canton at present.

So the rails the institutions are moving to do not yet carry a derivatives book. That is the position: building on those rails before the market catches up.

What discovery is now, and what it costs

Discovery is the blueprint for the desk. It is a defined piece of work with a fixed scope and a fixed end, and it is £60,000.

What the £60,000 buys, in deliverables:

• The contract specifications. What the desk arranges, the tenors, the sizes, and settlement against which index.
• The collateral and margin design, including what can be accepted from counterparties whose assets are hardware and revenue rather than treasury.
• The settlement architecture on a privacy-preserving institutional ledger, so trade terms stay between the two parties with disclosure to a regulator on demand.
• The regulatory pathway. Which permissions the desk needs, in which jurisdictions, and how long each one takes.
• The team plan. Who is hired, in what order, at what cost.
• The counterparty map. Who actually wants these hedges and cannot get them today.

Next steps

The next step is a call between the Capacity Derivatives team, Capacify.ai (Anik) and Gino DeMichele. The purpose is to discuss funding and agreement for the discovery work.

Prepared by Mark Abbott and David Gibbs. Capacity Derivatives is a venture of Rubytech LLC, a systems engineering, data analysis and technology consultancy, and the technology work described here is delivered by Rubytech LLC.

Every third-party statement above is attributed to the company or agency that announced it, with the date of that announcement, and is drawn from those published sources as at 25 August 2026. It is included for context and is not a representation by Rubytech. Nothing in this document is an offer of securities, an invitation to invest, or investment advice, and nothing in it should be relied on as such.