Dapple Raises $30M Seed to Sell Private AI Cloud as Owned Infrastructure

Why it matters

Dapple is testing whether regulated enterprises will commit to private AI infrastructure before public-cloud capacity catches up. The demand signal is strong, but the economics remain undisclosed.

Tricia Martinez-Saab and Salam Al-Mosawi's Dapple has raised a $30 million seed round from Raptor Group and Ion Pacific, according to a funding post on X and a June 10 PR Newswire release.

Dapple says the round came five months after launch and after more than $100 million in signed customer contracts. That contract figure is the center of the story, but it is not the same as recognized revenue. Dapple has not disclosed customer names, contract duration, cancellation terms, gross margin, annual recurring revenue, valuation, or whether either investor led the round.

That matters because Dapple is not selling another software dashboard with cloud margins. Dapple is selling dedicated AI infrastructure: single-tenant, in-country GPU capacity wrapped with orchestration, governance, compliance, observability, private connectivity, and developer access through what Dapple calls the Enterprise OS Cloud. The pitch is that regulated enterprises and governments should not have to choose between waiting on hyperscaler capacity, renting bare compute, or spending years building private data centers.

Martinez-Saab framed the raise as capital following proof, not capital funding a slide deck. In a Dapple blog post on the Enterprise OS Cloud, she wrote that Dapple did not start with a thesis and then search for customers. Dapple built the platform, put it into production, and used demand to test whether the category was real.

A founder taking another run at regulated infrastructure

Martinez-Saab is not a first-time founder trying to learn regulated markets from the outside. Dapple's company page describes her as Dapple's founder and CEO, and says she previously founded and scaled a digital bank in Africa, later worked as a White House Presidential Innovation Fellow at the U.S. Department of Energy across AI, energy, and national infrastructure initiatives, and led Techstars programs with partners including JPMorgan, NVIDIA, and the Department of Energy.

That earlier fintech chapter carries useful context. Martinez-Saab was the founder of Wala, a blockchain-enabled financial services platform for underbanked consumers. Techstars wrote in 2021 that she had been part of the 2016 Barclays Accelerator with Wala. CoinDesk later reported that Wala shut down in 2019 after running out of money and that user numbers promoted by the startup were disputed. For Dapple, that history makes the operating question sharper, not softer: Martinez-Saab is again building for institutions and users constrained by infrastructure, but this time the bottleneck is sovereign AI compute rather than payments rails.

Al-Mosawi brings the commercial infrastructure half of that bet. Dapple says he is its COO and leads data center strategy, GPU deployment, go-to-market, category design, and partnerships. Dapple's biography says he previously served as chief commercial officer at FlexAI and SVP of growth and partnerships at Nscale, and that teams he led sold tens of thousands of GPUs globally.

That pairing explains why the seed round is unusually large and unusually operational for a young company. Dapple's own materials say its team has collectively deployed GPU infrastructure at hyperscaler scale. Raptor Group, the family-office-backed private investment firm associated with Jim Pallotta, said in the release that Dapple had proved the category in production before raising to scale. Ion Pacific founding partner Michael Joseph said in the same release that demand is global and that enterprises want to run AI inside their own borders on infrastructure they control.

What Dapple is actually selling

Dapple's product sits between public cloud and a private data center. Its platform page describes one control plane across GPU orchestration, connectivity, governance and security, observability, and developer experience. Dapple says its system connects an enterprise's existing cloud environment to dedicated, single-tenant compute, with workloads including private LLM training, inference, fine-tuning, and digital twins.

The technical claims are specific, but still Dapple-sourced. Dapple's platform page cites 91% to 94% sustained GPU utilization, 256 GPUs as maximum contiguous capacity on an example cluster, and a target of more than 40,000 GPUs deployed by the end of 2026. Its infrastructure page says deployments can move from contract to production in 3 to 9 months, run 100% in-country and single-tenant, and include latest-generation NVIDIA accelerators, including GB300 NVL72, in high-density liquid-cooled configurations.

Those claims, if delivered, map directly to the procurement problem Dapple is attacking. Banks, insurers, energy companies, defense contractors, healthcare systems, and governments do not only need GPUs. They need residency controls, audit trails, identity integration, private networking, predictable throughput, and someone accountable for operating the stack after the contract is signed. Dapple's introductory post put the thesis plainly: regulated organizations need AI infrastructure they fully control, with compliance enforced before workloads run.

That is a different business from selling on-demand GPU instances to developers. It is closer to a managed AI factory contract: Dapple aggregates supply, deploys dedicated capacity against enterprise commitments, and operates the environment under compliance and residency constraints.

The contract number is the headline, but fulfillment is the test

The $100 million-plus contract figure gives Dapple a strong demand signal for a five-month-old operating business. It also raises the hard question: how much capital will Dapple need to turn signed commitments into production capacity?

GPU infrastructure is capital hungry even before adding in-country deployment, private networking, power planning, liquid cooling, compliance evidence, and 24/7 operations. Dapple's infrastructure page says capacity is deployed against real enterprise commitments rather than speculative demand. That is a disciplined framing, but it still means Dapple has to finance and execute physical deployments on enterprise timelines.

The seed round answers only the first part of that problem. A $30 million seed gives Dapple working capital, investor validation, and a larger balance sheet for procurement conversations. It does not, by itself, prove that the contracts are profitable, that customers will expand, or that Dapple can secure enough GPU supply and data-center capacity to hit its own 2026 deployment target.

Dapple's legal and geographic footprint is also worth reading carefully. Dapple's website says Dapple is headquartered in Europe with global operations. Business Barometer lists Dapple AI Networks Limited as an Irish company set up on December 19, 2025, with a Dublin partial address and normal company status. The PR Newswire release was datelined New York, but Dapple's own materials point to a Europe-headquartered operating model.

That Europe-first posture fits the market Dapple is chasing. Sovereign AI is no longer an abstract policy phrase. Enterprises and governments increasingly want model training and inference to happen under local data-residency, governance, and audit regimes. Dapple's bet is that those buyers will commit before public cloud capacity and compliance packaging catch up.

The warm read is that Martinez-Saab and Al-Mosawi have found a narrow but valuable wedge: regulated AI workloads that are important enough to justify dedicated infrastructure but not so large that the buyer wants to build and operate an AI data center alone. The skeptical read is that $100 million in signed contracts can create as much pressure as momentum if delivery requires expensive, country-specific capacity before revenue is fully realized.

For now, Dapple has the right kind of early signal: named investors, production claims, and enterprise contracts large enough to stand out in a crowded AI infrastructure market. The next proof point is not another round. It is whether Dapple can turn those contracts into running, repeatable deployments with economics that hold after the racks are powered on.