Headline & intro
Enterprises are only just getting comfortable running AI at scale, and already another wave of complexity is looming: quantum computing. While most CIOs still file quantum under "interesting, but later," Finnish startup QuTwo is arguing that "later" is exactly when you are too late. By offering an orchestration layer that can push AI workloads across classical, quantum-inspired and, one day, real quantum hardware, QuTwo wants large companies to treat quantum not as a moonshot, but as a gradual infrastructure upgrade. If that strategy works, it could quietly define how businesses actually consume quantum power long before the qubits go mainstream.
The news in brief
According to TechCrunch, Finnish entrepreneur Peter Sarlin, who sold his previous startup Silo AI to AMD for $665 million around a year and a half ago, has left his role as CEO of the acquired unit (now AMD Silo AI) and re-emerged as chairman of two new ventures. One is NestAI, a physical AI lab. The other is QuTwo, an "AI lab for the quantum era," currently financed via Sarlin’s family office PostScriptum.
QuTwo is already working with enterprises. With German-based fashion platform Zalando, it is co-developing so‑called "lifestyle agents" – AI systems that go beyond search to proactively recommend products and experiences. QuTwo is building QuTwo OS, an orchestration layer meant to route AI workloads across classical, quantum-inspired and, eventually, quantum hardware. The company has assembled a team of more than 30 AI and quantum scientists, including former IQM cofounder Kuan Yen Tan, and ex-Nokia CEO Pekka Lundmark on the board. TechCrunch reports that QuTwo has signed design partnerships with enterprise customers worth "tens of millions" of dollars, including a quantum AI research initiative with Finnish financial group OP Pohjola.
Why this matters
QuTwo is going after the most painful part of today’s AI boom: the compute bottleneck. Training and running state-of-the-art models is colliding with limits on available GPUs, rising latency, and an uncomfortable energy bill. The bet behind QuTwo is that enterprises do not actually care whether that compute comes from GPUs, quantum-inspired algorithms, or real quantum processors — as long as someone abstracts the complexity away and reduces cost per useful operation.
If QuTwo OS works as advertised, it becomes a kind of Kubernetes or VMware for the quantum era: a control plane that decides which hardware to use for which workload. That is strategically powerful. Instead of each enterprise building its own quantum experiments and integration layers, they could simply treat quantum as another back-end target selected by policy, not by hand-coded experiments in a lab.
The immediate winners are large enterprises that know they will eventually need more efficient optimization and simulation but cannot bet on one quantum hardware vendor or timeline. For them, a hybrid orchestration layer is a form of insurance policy: it lets them be early without committing to one stack.
The potential losers are cloud incumbents that hope quantum will be another proprietary service inside their walled gardens. If an independent layer like QuTwo becomes the default control point, hyperscalers risk being reduced to interchangeable commodity execution engines for whichever algorithms QuTwo routes their way.
There is also a more subtle shift: QuTwo frames quantum not as a separate R&D project, but as a continuum of algorithmic choices to solve business problems. That pragmatic packaging could be exactly what moves quantum from whiteboards and conference talks into actual P&Ls.
The bigger picture
QuTwo slots into a broader trend: the realization that quantum’s first commercial impact will likely be hybrid. IBM’s work on Qiskit Runtime, Microsoft’s Azure Quantum and Amazon Braket have already shown that the industry expects quantum workloads to be orchestrated alongside classical ones. What QuTwo is doing is pushing that idea up the stack and centering it specifically on AI workloads.
We have seen this movie before. In the early cloud days, most enterprises did not migrate directly from on-premise to a single hyperscaler. They tiptoed via abstraction layers, multi-cloud tools and vendors that promised portability: RightScale, Cloud Foundry, VMware, OpenStack. Many of those names faded, but they bought time for enterprises and helped shape today’s norms around containers and APIs.
Similarly, "quantum-inspired" algorithms running on classical hardware are gaining real traction, particularly in optimization, finance and logistics. They borrow ideas from quantum annealing and other techniques but execute on GPUs and CPUs available today. That is where QuTwo becomes very interesting: even if large-scale fault-tolerant quantum computers slip another decade, the company can still route workloads to these quantum-inspired implementations and deliver value now.
Competitively, QuTwo sits somewhere between hardware providers (IQM, Rigetti, IonQ, Pasqal), cloud abstractions from hyperscalers, and specialist quantum software houses. Its differentiation is its focus on AI and on enterprise co-design via big-ticket design partnerships. Rather than publishing toolkits and hoping developers show up, it is embedding itself inside corporations’ roadmaps in fashion, finance and beyond.
The underlying message is clear: the industry is shifting from "Can we build more qubits?" to "Who controls the interface between messy business problems and whatever exotic hardware solves them?" QuTwo wants to own that interface.
The European angle
For Europe, QuTwo is more than just another deeptech startup from the Nordics; it is a proof point for the EU’s ambition to be more than a research lab for American platforms. The company is Finnish, its investors and several team members are tied into Europe’s quantum ecosystem (IQM, SemiQon), and its flagship early customers — Zalando and OP — are European brands.
That matters in a regulatory environment shaped by GDPR, the Digital Services Act and the coming EU AI Act. European enterprises are under pressure to run AI workloads that are not only powerful but also energy-efficient, auditable and compliant with strict data protections. Quantum and quantum-inspired optimization could, in theory, cut energy use for some classes of problems. If QuTwo can wrap that in a controllable, policy-aware orchestration layer, it becomes much easier for a European bank or retailer to justify pilot projects to auditors and boards.
There is also a sovereignty dimension. The EU has poured billions into its Quantum Flagship and Chips Act programs, but much of the public narrative still orbits around qubit counts at IBM, Google and the US cloud giants. Seeing a Finnish company position itself as a future gatekeeper for quantum-accelerated AI is strategically comforting for policymakers who worry about repeating the cloud and social media playbook, where Europe became a rules-maker but not a platform-maker.
For European enterprises, the timing is good. AI adoption is accelerating, but infrastructure decisions for the next five to ten years are being made now. A vendor-neutral way to experiment with quantum-style acceleration without committing to one non‑EU hyperscaler fits very neatly into European CIOs’ long-term thinking.
Looking ahead
In the next three to five years, most of QuTwo’s value will likely come from two areas: quantum-inspired algorithms running on classical hardware and smarter routing of heavy AI workloads across existing accelerators. Fully error-corrected quantum machines able to give clear, repeatable advantages in mainstream AI remain a longer-term prospect.
For readers in enterprise roles, the key signals to watch are not glossy demos, but whether any of QuTwo’s design partners can publicly show business outcomes: better recommendation quality, improved risk modeling, lower energy bills, shorter training cycles. If those metrics move meaningfully, the argument for a quantum-ready orchestration layer becomes much stronger.
Strategically, two questions loom. First, will hyperscalers let an independent control plane thrive, or will they respond with their own quantum-first AI orchestration offerings, bundling them tightly into their clouds? If that happens, QuTwo must prove that its neutrality, European roots and deep specialization justify a place in the stack. Second, can QuTwo avoid being trapped in endless co-development loops with a few big customers, and instead generalise QuTwo OS into a product that mid‑market enterprises can adopt without custom integration marathons?
On timeline, expect more joint research announcements and design partnerships before broad commercial availability. Quantum-related enterprise procurement moves slowly, especially under European regulation. But that slowness cuts both ways: it gives QuTwo a window to entrench itself before the hardware ecosystem consolidates.
The bottom line
QuTwo is making a smart, almost contrarian bet: stop arguing about when quantum computing will change everything, and start selling enterprises a way to be ready regardless of the date. If it can turn quantum and quantum-inspired resources into just another target for AI workloads, it may quietly shape how Europe — and perhaps the world — consumes quantum power. The real question for readers is simple: do you want to meet quantum as a future migration crisis, or as a capability you have been incrementally testing for years?
