Startups & Funding

Ex-BitGo and Galaxy Execs Raise $11M to Finance AI Compute

Marble launched Nov 10, 2026 with an $11M Series Seed led by Polychain to build private-credit infrastructure for mid-market GPU operators and a compute-linked trading platform.

Marble Raises $11M Series Seed to Build AI Compute Credit Platform
Marble Raises $11M Series Seed to Build AI Compute Credit Platformblickpixel / pixabay
By James Calloway5 min read

Why it matters

  • Marble launched November 10, 2026 with an $11 million Series Seed led by Polychain Capital.
  • Co-founders Matt Ballensweig (ex-BitGo) and Leon Marshall (ex-Galaxy Digital Europe) serve as co-CEOs.
  • Marble Compute expects first loan syndications for mid-market GPU operators in early Q1; Marble Markets aims to start digital-asset trading in January.
  • The platform plans overcollateralized lending and OTC compute forwards linking credit and trading businesses.
  • No products are live yet; rollout is subject to regulatory approvals.

Marble launched on November 10, 2026 with an $11 million Series Seed round led by Polychain Capital to build private-credit infrastructure for AI compute operators. The company, founded by former BitGo and Galaxy Digital executives, is also planning an institutional trading platform spanning digital assets, prediction markets and compute.

Maven 11, Ripple, Susquehanna Crypto, Lattice and Blockwall Ventures participated in the round, alongside other venture and strategic investors and a group of AI-focused angels. Marble plans to spend the proceeds on hiring across compute financing, trading, risk, legal and operations. It also intends to establish the trading, custody and banking connectivity behind its Markets platform.

Co-founders Matt Ballensweig and Leon Marshall serve as co-CEOs. Ballensweig was most recently BitGo's global head of trading and previously co-headed trading and lending at Genesis. Marshall was CEO of Galaxy Digital's European business and previously worked at Genesis, Apollo and Trafigura.

The launch matters because it targets a gap in how AI infrastructure gets funded. The largest neoclouds — independent AI cloud providers — and hyperscalers can finance growth through public markets and major banks, according to the founders. Smaller operators have fewer and less flexible options even when customers have already committed to purchasing their capacity. In plain words: Marble wants to lend institutional money against GPU contracts that banks will not touch.

What is Marble Compute?

Marble Compute is designed as an asset-management platform connecting institutional investors with neoclouds and GPU operators. It is not designed to lend from its own balance sheet. Instead, it will source opportunities, conduct diligence and underwriting, and structure each financing for investors.

Its underwriting process will examine three things: the hardware itself, the creditworthiness of contracted customers, and the operator's ability to service debt. Marble says it will also seek diversified customer commitments, so that financing is supported by contracted cash flows rather than a single counterparty's solvency.

The staged rollout, subject to regulatory approvals, begins with loans syndicated directly between operators and institutional lenders. Later stages would add single-opportunity off-chain vehicles and on-chain vaults for individual GPU operators. Diversified multi-opportunity funds would follow as deal flow and a track record develop. The takeaway: this is an intermediation business, and its economics depend on whether institutional lenders actually show up for syndicated GPU debt.

How does Marble Markets work?

Marble Markets is planned around principal trading, meaning clients would face Marble directly while Marble trades from its own capital. Client assets would remain with their existing custodians, brokerages and banks. A high-touch desk would handle trading, backed by a client platform showing positions, collateral and exposure in one place.

The planned product set includes:

  • digital-asset spot trading, borrowing, lending and derivatives
  • large-block bilateral prediction-market trading
  • over-the-counter compute forwards
  • tokenized equities and equity swaps, over the longer term

Rather than requiring clients to prefund each counterparty separately, the design would let them use one pool of eligible digital and traditional collateral across markets. Marble says its Markets lending would be overcollateralized, meaning backed by collateral exceeding the loan's value.

The shared-collateral model is the platform's clearest differentiator on paper. It remains a design, not a shipped service, and Marble has not disclosed expected collateral eligibility criteria or margin levels.

Why compute forwards connect the two businesses

Compute forwards link the credit and trading arms. They would let an operator fix a rental price for capacity delivered in the future. Marble says this gives lenders clearer cash flows to underwrite and gives operators a better basis for determining borrowing capacity and equity requirements.

In the launch release, Marshall said: "Two GPU operators can buy identical hardware and raise very different amounts of debt against it. A lender wants to know who's buying the compute, at what price and for how long."

That quote cuts to the core of the underwriting problem. Identical GPUs do not produce identical credit risk; the contracted customer base and revenue duration do. Marble's answer is to standardize that information through forward contracts that make future cash flows legible to lenders. Marble Compute would originate the operator relationships, with Marble Markets deepening them through hedging. The founders also describe Compute as a structured route for digital-asset capital to finance AI infrastructure and earn returns from it.

What is actually live today?

Nothing, by the company's own account. The products remain development plans rather than live services, and their rollout is subject to applicable regulatory approvals. The timeline Marble has set is specific: Marble Compute expects to close its first loan syndications for mid-market GPU operators in early Q1, while Marble Markets aims to begin digital-asset trading in January.

Those dates are targets, not commitments, and the release does not name any committed lenders, operators or trading clients. An $11 million seed also constrains how quickly a principal-trading desk can scale; Marble has not disclosed its capitalization plans for Markets beyond the round. The caveat is straightforward: a platform announced with two planned business lines and no live products has more design risk than execution risk at this stage.

Why this launch matters

The launch links two requirements for expanding independent AI compute: financing for the hardware and a mechanism for hedging the revenue it generates. Neither exists today at institutional scale for mid-market operators, and hyperscalers' access to cheap public-market capital keeps widening the cost gap. Marble's combined credit-plus-hedging design is an attempt to close that gap with private-credit structuring borrowed from digital-asset markets.

The first syndications and the January trading launch will test whether Marble can turn that combined design into an operating business and broaden institutional financing beyond the largest cloud providers.

Source: unite.ai

Share this article:

More from James Calloway

James Calloway

Show full bio

News editor covering people, launches and day-to-day practice in artificial intelligence and machine learning at AI In Context.

533 articles

Related articles

  1. AMD Buys Fei-Fei Li's World Labs for $8.2 Billion in All-Stock Deal
  2. Microsoft Cuts Its Neocloud Roster as AI Buildout Shows Strain
  3. Three of Four Hyperscalers Now Burn Cash as AI Capex Hits Legacy Businesses
  4. Nous Research hits $1.5B valuation, launches enterprise AI agents
  5. Samsung Puts $1 Billion Into KKR- and Nvidia-Backed AI Infrastructure Firm Helix

« Previous articleNext article »