You write a line of code to run a quantum circuit, but you don't know if the result is real or noise. You need to benchmark across three different hardware backends, but the logs live in three different places. You want to see if a GPU could have solved the problem faster than the quantum processing unit you just spent $5,000 to access. This is the daily friction of hybrid compute, a world where quantum, AI, and high-performance classical systems are supposed to work together but mostly just talk past each other. It's the kind of problem that makes a researcher close their laptop and go for a very long walk.
Marqov, a San Francisco startup founded in 2025, is building the guardrails for that walk. Its product, TESTBED, is an orchestration layer that attempts to corral the chaos of heterogeneous compute. It promises to schedule, monitor, and, most crucially, reproduce workloads that span QPUs, GPUs, CPUs, and simulators as a single, observable workflow [Marqov, “About”]. The company is in a quiet, invitation-only private beta, offering its open-source Python SDK for free while it courts its first enterprise customers [Marqov, “About”]. It’s a bet on infrastructure for a future that is still being invented, one circuit at a time.
The Wedge Is Reproducibility
In a field where a single result can be groundbreaking, the inability to reliably reproduce an experiment is more than an annoyance. It's a fundamental blocker. Marqov’s initial wedge isn't raw speed or a novel algorithm. It’s the mundane, critical work of verification and observability. The TESTBED platform is positioned as a reproducible "qOps" environment, designed for designing, benchmarking, and scheduling these complex hybrid jobs [Marqov PhD Residency posting]. Its SDK speaks to major providers like Amazon Braket and Azure Quantum out of the box, allowing users to swap hardware backends without rewriting their core logic [unitaryHACK 2026].
The value proposition is a unified pane of glass for a fragmented stack. According to team posts, TESTBED already supports integrations with 10 quantum processors, 9 simulators, and 14 AI foundation models, providing a single way to pull usage metrics from systems that normally don't speak the same language [Alberto Montagnese, LinkedIn]. For the target customer,research labs, quantum hardware teams, and enterprise AI groups working at this intersection,the product sells order.
A Founder Built for the Frontier
The company’s credibility rests heavily on the profile of its CEO, Anastasia Marchenkova. She is not a newcomer to the quantum scene. A quantum physicist and three-time founder, she has over 15 years in the field, serves as vice chair of the IEEE Quantum Computing Standards Project, and runs a media platform demystifying the technology [Heavybit, September 2026] [RSAC Conference, “Anastasia Marchenkova”]. She has also won top honors in IBM’s Qiskit Camps for her work on quantum programming performance [RSAC Conference, “Anastasia Marchenkova”]. This isn't just technical depth; it's a deep network and a practiced voice in a community that values both.
Her co-founder, David Ryan, brings the infrastructure engineering chops. As CPO, his background at Quantum Brilliance and Red Hat focuses on the orchestration layer itself,the pipes and plumbing that make heterogeneous scientific computing possible [Marqov, “About”]. The division of labor is clear: one founder maps the quantum landscape, the other builds the road.
| Role | Name | Key Background |
|---|---|---|
| Co-Founder & CEO | Anastasia Marchenkova | Quantum physicist, IEEE standards vice chair, three-time founder, investor via Strange Quark Capital [Heavybit, September 2026] [Anastasia Marchenkova, retrieved 2026] |
| Co-Founder & CPO | David Ryan | Infrastructure engineer, former roles at Quantum Brilliance and Red Hat [Marqov, “About”] [David Ryan, LinkedIn] |
The Early Traction and the Silent Round
Marqov operates with a notable lack of fanfare for a venture-backed startup. Its only publicly named investor is Generationship, though the size and terms of any funding remain undisclosed [Fundraising Fox]. In a September 2025 post, Ryan alluded to a pre-seed round intended to fund research with a national lab and a major quantum vendor, alongside open-source benchmark development [David Ryan, LinkedIn post, September 2025]. The company has announced a plan to double its team in 2026 [Alberto Montagnese, LinkedIn].
Its technical integrations are its most tangible proof points. A primary partnership is with quantum computing company Quantinuum, and the platform’s broad backend support suggests active development and engagement with hardware providers [Giuseppe Barbalinardo, LinkedIn]. The traction, for now, is measured in connectors built and researchers invited into the beta, not in public revenue announcements.
Navigating a Nascent Market
The risks for Marqov are inextricable from the market it serves. Quantum computing, for all its promise, remains largely experimental. The pool of customers running serious, production-scale hybrid quantum-classical workloads is small and elite. The company must grow in lockstep with the adoption of the underlying technologies it orchestrates.
- Market timing. Commercial quantum advantage for broad applications is still years away. Marqov’s success is tied to the pace of progress in the core quantum hardware and algorithm space, a timeline it does not control.
- The open-source play. By offering a free, open-source SDK, Marqov is betting on developer adoption and community contribution to build its moat. This is a classic playbook, but monetizing that community with an enterprise product is a separate challenge, especially when the SDK can be used without a Marqov account [Marqov, “About”].
- Future competition. As the hybrid compute market matures, cloud hyperscalers with their own quantum services (AWS Braket, Azure Quantum) or classical workload orchestrators (Kubernetes-focused platforms) could move to own this orchestration layer themselves.
The company’s answer appears to be a focus on deep, early partnerships with labs and hardware providers, embedding itself in the workflow before the workflow is fully defined. Its involvement in standards bodies through Marchenkova is a strategic effort to shape the very environment it will later commercialize.
The Next Twelve Months
The coming year will be about moving from a promising toolkit to a relied-upon system. The plan to grow the team will likely focus on engineering and developer relations to support the open-source community and beta users. The key milestone to watch will be the transition from a free, invitation-only beta to a publicly available product with clear commercial tiers. Landing and announcing a first named enterprise customer,particularly from its target segments of HPC centers or advanced engineering organizations,would provide a crucial signal of market validation.
The cultural question Marqov is implicitly answering is one of trust in a deeply uncertain technical frontier. It is not selling computational supremacy. It is selling a lab notebook that won't catch fire, a control panel for a machine that is still being assembled. In a domain where progress is measured in reproducible results, the company is betting that the most valuable thing to build first is not a faster engine, but a reliable logbook.
Sources
- [Marqov, Unknown] About | https://marqov.ai/about
- [Marqov, Unknown] Marqov, An orchestration layer for heterogeneous compute | https://marqov.ai/
- [LinkedIn Jobs, December 2025] PhD Residency Program at Marqov | https://www.linkedin.com/jobs/view/phd-residency-program-at-marqov-4342083871
- [Heavybit, September 2026] Quantum Uncertainty with Anastasia Marchenkova | https://www.heavybit.com/library/podcasts/generationship/ep-58-quantum-uncertainty-with-anastasia-marchenkova
- [LinkedIn, Unknown] David Ryan, LinkedIn | https://www.linkedin.com/in/hellodavidryan
- [Fundraising Fox, Unknown] Marqov, Investors & Founders | https://fundraisingfox.com/companies/marqov
- [LinkedIn, September 2025] David Ryan, LinkedIn post, September 2025 | https://www.linkedin.com/posts/hellodavidryan_how-did-marqov-get-the-first-investor-in-activity-7371002887293956096-5H81
- [RSAC Conference, Unknown] Anastasia Marchenkova | https://www.rsaconference.com/experts/anastasia-marchenkova
- [unitaryHACK 2026] unitaryHACK 2026 | https://unitaryhack.dev/
- [Giuseppe Barbalinardo, LinkedIn, retrieved 2026] LinkedIn post | https://www.linkedin.com/in/giuseppe-barbalinardo
- [Alberto Montagnese, LinkedIn, retrieved 2026] LinkedIn post | https://www.linkedin.com/in/albertomontagnese
- [Anastasia Marchenkova, retrieved 2026] Press & Speaking | Anastasia Marchenkova | https://www.amarchenkova.com/about-me