Allergic to impossible since day one.
A studio of scientists and engineers who collect the briefs other vendors politely decline — and ship them anyway.
Viceron started with a single brief that three vendors had refused. It shipped. Word got around that there was a team in Copenhagen that didn't flinch — and the refused briefs have been finding their way here ever since.
The first “can’t”
A dataset nobody could read, a deadline nobody believed. We took it because nobody else would — and discovered that impossible mostly means unpriced.
The pattern
Moonshots, needles, velocity. Every refused brief fell into one of three shapes — so we built a studio around exactly those three muscles.
The arsenal
NVIDIA Inception for compute, Danoffice IT for reach. We stopped renting our bottlenecks and started owning them.
Today
A growing file of delivered impossibilities — and a standing invitation: bring us the one everyone says can't be done.
Three rules. Zero exceptions.
Evidence over opinion
Every claim we make ships with the data that proves it. If we can't measure it, we don't say it.
Speed is respect
Your deadline is real. Moving fast isn't recklessness — it's taking your problem seriously.
No black boxes
A system you can't explain is a system you can't defend. Everything we ship shows its reasoning.
Meet our team.
Peter Schwarz
Chairman of the BoardHans Albertsen
Vice Chairman of the BoardJoined by founders
Julius & Sofus Lund-Jacobsen.
Along with an unnamed board member and advisor.
Nobody does the impossible alone.
As NVIDIA Inception members we train on accelerated computing that turns week-long experiments into overnight runs.
Inception ProgramHow we collaborate →
Hardware and licensing, delivered globally — from a single workstation to a GPU cluster in a hurry.
Hardware & licensingHow we collaborate →
And in the field — global pharma · hospitals · public authorities · research institutes · startups
NVIDIA × Viceron
As members of the NVIDIA Inception Program, we build and train on accelerated computing — which quietly redraws the line between possible and impossible.
How we collaborate. Inception gives our engineers early access to NVIDIA's AI stack, technical enablement and preferred access to GPU capacity. Our discovery models, forecasting systems and language pipelines are engineered CUDA-first, so client workloads run on anything from a single workstation to a DGX-class cluster.
What it means for you. Experiments that would take a week converge overnight. A rare-signal model that needs fifty training iterations is a non-starter at one iteration per week — at five per night, it's a Tuesday. Compute budgets define the boundary of the possible; through this partnership, ours is generous.
Member of the NVIDIA Inception Program for companies advancing AI and data science.
HPE × Viceron
Some data can never leave the building — patient records, state casework, trade-sensitive operations. With Hewlett Packard Enterprise, we bring the AI to the data instead.
How we collaborate. We design and deliver Viceron systems on HPE infrastructure inside the client's own perimeter: GPU-dense compute for training, hardened environments for regulated workloads, hybrid setups where only the non-sensitive tiers touch cloud. Reference architectures are pre-agreed, so infrastructure is never the long pole in a six-week delivery.
What it means for you. Sovereignty without sacrifice. Public healthcare regions, authorities and pharma companies get the same models and the same velocity as cloud-native clients — with every record exactly where your auditors, your lawyers and your citizens expect it to be.
First joint deployments live in Danish public healthcare.
Danoffice IT × Viceron
The unglamorous truth about AI programs: someone has to get the right hardware, with the right licenses, to the right place — sometimes a hospital basement, sometimes another continent.
How we collaborate. Danoffice IT handles global hardware logistics and software licensing for our deliveries: GPU workstations for embedded research teams, cluster components for on-premise builds, and the licensing paperwork enterprises require before anything powers on. We specify; they deliver — wherever the project lives.
What it means for you. One accountable chain from architecture to loading dock. No procurement dead-ends three weeks into a six-week plan, no gray-market hardware surprises, and global reach when your project isn't in Copenhagen.
Hardware and licensing delivered worldwide.