Problem

AI runs on physical infrastructure. That infrastructure costs trillions to build, and equity markets alone cannot absorb that issuance pace. Roughly 30% has to be debt financed.

Hyperscalers issue bonds. Smaller GPU operators and independent datacenter companies can get bank loans, but it is harder, slower, and the terms are worse , banks do not have a framework for underwriting GPU collateral. So these companies turn to private credit and pay a steep spread.

That spread is real, durable, and tied to physical demand. It exists for one simple reason: the infrastructure has to get built.

Opportunity

Ethena proved the playbook. Take a large, structural yield source that TradFi keeps to itself, make it accessible on-chain, and build around it before anyone else does. One durable yield category was enough to build a multi-billion dollar protocol.

Compute financing is the same shape, with a bigger underlying market, a cleaner economic driver, and far fewer people who understand both the structured finance side and the Solana distribution side well enough to execute.

Ethena was built on perp funding rates, a yield source that lives entirely inside crypto and disappears when funding flips. Compute financing is built on the debt funding the AI build-out, trillions in issuance over the next few years, paid by companies building real assets.

The pool is orders of magnitude larger, and the yield is paid by physical demand, not by traders on the other side of a basis trade.

Solution

Deposit USDC. Foundation routes that capital into verified AI infrastructure credit and gives you fcyUSD in return, a vault token that appreciates as the underlying loans pay out.

How compute financing compares to everything else available on-chain today:

US T-bills sit around 4%, driven purely by the sovereign risk-free rate
Solana money markets and lending protocols sit around 3 to 6%, driven by DeFi rate arbitrage
Ethena-style basis trades can reach 8 to 15% but are reflexive and collapse when funding rates flip
AI compute financing sits structurally above all of these, backed by credit spreads tied to real infrastructure demand that does not disappear in a bear market

Market

Today there is only about $80 million of active compute-financing loans on-chain, set against over a trillion dollars of AI capex this year alone. The category is barely touched. Cumulative AI capex is projected to reach somewhere between three and seven trillion dollars by 2030.

To picture what on-chain capture looks like: if even 1% of that issuance moves on-chain, that is tens of billions in TVL. The ceiling is set by how fast the build-out happens, and the build-out is already happening.

Foundation is here at the moment the asset class is being defined. The protocol that sets the standard first becomes the benchmark. Benchmarks attract liquidity, and liquidity, once it consolidates, does not move.

Traction

The vault infrastructure is live on Solana mainnet. We have roughly 950 waitlist signups, a verbal commitment from GAIB with an active proposal under review, and Turtle Club committed to running an LP campaign at launch.

Kamino is in discussion as our exclusive launch venue, where users will be able to use fcyUSD as collateral, borrow against it, and loop it for amplified exposure. We are part of the Circle Alliance Program, and in active conversations with the Solana Foundation across their DeFi, AI, and RWA teams.

You can join the waitlist at beta.fdnusd.com.

Defensibility

We are not betting on a tech moat. We are betting on three things competitors cannot replicate by writing code.

Originator relationships. The hard part is not building a vault. It is getting credit-quality issuers to allocate to you. Every issuer we onboard makes the next one easier and makes a competitor's catch-up harder. Supply-side scarcity compounds.

Vertical focus. Generalist credit protocols can add AI exposure as a checkbox. They will not win the category. When a yield category becomes large enough to deserve its own brand, the dedicated product takes the share.

Roadmap

Month 1 (June) : Legal structure finalized (BVI, tax, compliance). Audit underway. GAIB onboarding confirmed. Internal testnet live.

Month 1.5 (mid-July) : fcyUSD vault launches on Solana mainnet with GAIB as first yield source. $500 seeded at launch from ecosystem LPs and team. Kamino and Loopscale integrations live from day one. Turtle Club LP campaign goes live alongside the launch.

Month 2 (late July) : fcyUSD token listing. Yield becomes accessible to retail. $3 to 5M TVL through LP networks, since every dollar of fcyUSD can be levered, which multiplies effective yield exposure and pulls in deposits.

Parallel track begins: CLO and private credit onboarding (Turtle Club and other LP networks, commitments already in place). Securitize, Centrifuge, Fasanara onboarding alongside.

Month 3 (September): Alternative yield assets onboarded as second track (CLOs, private credit).

Q4 2026 : Structured RWA assets from Securitize and Centrifuge go live. A multi-constituent vault gives diversified compute and structured credit exposure. $15 to 25M TVL. Foundation is the default AI-yield product on Solana.

2027 : Foundation transitions from aggregator to originator. Direct datacenter capex and power grid financing. Institutional instrument suite. Category benchmark for AI yield on Solana.

Team

Eugene Bochkov, CEO. Three-time founder with twelve years across product, fintech, afinance. Previously PM at Spenmo (YC S20) and KPMG. Most recently scaled Synnax to $20M+ AUM on EVM.

Vivek Pal, CTO. Protocol engineer with six-plus years shipping production infrastructure across Solana, Filecoin, and Ethereum.

We have been operating at the intersection of stablecoins, RWA, and fintech long enough to know where the gaps are. This is not a pivot. It is the application of accumulated context to the largest new yield category on the table.

Previously backed by

Antler and SEI Grant. Accepted into Colosseum Accelerator. Won the local track at Superteam Malaysia's Frontier Hackathon. Active support from Solana Foundation.

Links