Finnish entrepreneur Peter Sarlin is back with a new bet on enterprise AI for the quantum age. Eighteen months after selling Silo AI to AMD for $665 million and leading the unit now known as AMD Silo AI, Sarlin has emerged as chairman of QuTwo, an “AI lab for the quantum era” that aims to put companies on quantum-ready rails before large-scale quantum hardware arrives.
Funded by Sarlin’s family office PostScriptum, QuTwo is already working with paying customers. Early projects include “lifestyle agents” with European fashion platform Zalando—systems that go beyond search to anticipate style intent—and a joint quantum-AI research initiative with OP Pohjola, one of Finland’s largest financial services groups.
Building Now For A Hybrid Quantum Future
QuTwo’s thesis is pragmatic: useful quantum advantage will land unevenly across workloads and will initially live in hybrid stacks. Its QuTwo OS is pitched as a routing and orchestration layer that can push or pull AI workloads across classical accelerators and, when viable, quantum chips—while also exploiting “quantum-inspired” algorithms that run on today’s CPUs and GPUs.
Quantum-inspired optimization and sampling methods have been shipping for years—Microsoft has offered a quantum-inspired optimization suite in Azure, and D-Wave customers regularly mix annealing with classical solvers. QuTwo is leaning into this middle ground to capture immediate gains in combinatorial optimization, personalization, and planning, without waiting on error-corrected qubits.
That hardware agnosticism is reflected in the team. QuTwo brings together IQM cofounder Kuan Yen Tan and board member Antti Vasara, who also chairs Finnish quantum chip venture SemiQon, alongside enterprise veterans Sarlin and Kaj‑Mikael Björk, a cofounder from Silo AI. Former Nokia CEO Pekka Lundmark has joined the board, and the company says it has more than 30 quantum and AI scientists on staff.
Early Use Cases With Measurable Business Lift
Why would an online retailer care about quantum-inspired tooling? Recommendation engines, sizing guidance, delivery routing, and inventory placement are all optimization-heavy. Even on classical hardware, techniques inspired by quantum annealing and tensor networks can search large solution spaces faster or with better heuristic quality, which can translate into higher basket size or lower returns.
In finance, OP Pohjola’s collaboration points to risk, fraud, and portfolio optimization—domains where Monte Carlo sampling and combinatorial selection dominate. Banks already test quantum methods with providers like IBM and AWS’s Braket; a fabric that lets models flip between classical solvers and quantum backends as they mature is an attractive hedge against vendor and technology risk.
The timing aligns with industry estimates. McKinsey has projected that quantum technologies could unlock up to $1 trillion in value by 2035 across sectors such as automotive, chemicals, and finance, with early wins in optimization and materials simulation. Building software interfaces, data pipelines, and MLOps practices now reduces the time-to-impact once hardware crosses reliability thresholds.
Energy Efficiency As A Strategic Driver For AI
There’s also an energy story. The International Energy Agency has warned that global data center electricity demand could roughly double between 2022 and 2026, with AI a major driver. While broad efficiency gains from quantum remain a research question, certain workloads—optimization, sampling, and linear algebra primitives—could see order-of-magnitude improvements on future quantum machines, lowering energy per solved task.
QuTwo’s pitch is that a scheduler able to dispatch to the most energy- and cost-efficient pathway—classical, quantum-inspired, or quantum—will appeal to CIOs under sustainability and budget pressure. Even before fault-tolerant systems exist, routing to the right classical heuristic at the right time can shave compute hours and cloud spend.
Design Partnerships And Revenue From Day One
Unlike many research-heavy quantum plays, QuTwo is starting with enterprise co-development. The company says it has secured large design partnerships worth tens of millions, a structure that lets customers influence product roadmaps in exchange for early access. It’s a playbook seen in quantum programs at firms like JPMorgan and Volkswagen, which have worked closely with hardware and platform vendors to shape use cases.
For buyers, the bet is twofold: capture incremental gains today from quantum-inspired techniques, and shorten the integration gap to true quantum acceleration later. For QuTwo, design partners provide grounded datasets, clear KPIs, and hardening paths for an OS that must straddle GPUs, specialty accelerators, and, eventually, cryogenic quantum processors.
A Finnish Beachhead With European Scale Ambitions
Finland has quietly assembled a quantum stack from chip to application. IQM co-built a quantum computer with VTT, the national research center; SemiQon is pursuing silicon-based qubits; and the European Union’s Quantum Flagship initiative has poured sustained funding into the region’s research. QuTwo slots into this ecosystem as the software and AI abstraction layer designed for real enterprise traffic.
The company’s runway is currently secured by PostScriptum, and Sarlin is also chairing NestAI, a physical AI lab, hinting at cross-pollination between embodied intelligence and quantum-inspired planning. The near-term question is execution: Can QuTwo convert pilots into scaled deployments while quantum hardware roadmaps remain volatile?
If it can, the payoff is leverage. Enterprises that standardize on a hybrid AI fabric now will be positioned to flick on quantum acceleration when and where it helps—without rewiring their entire stack. That promise, more than speculative speedups, explains why some customers are choosing to run before quantum fully walks.
