Legal name: Quanscient Oy
Quanscient Deep Dive: Cloud-Native and Quantum-Powered Multiphysics Simulation as the Data Engine for AI-Native Hardware Engineering
Quanscient (Finland) powers hardware engineering with cloud-native multiphysics Allsolve and quantum algorithms; a EUR 10 million Series A in May 2026 targets physics-aware AI training data.
Executive Summary
Quanscient is a multiphysics simulation company founded in 2021 in Tampere, Finland, by Juha Riippi, now CEO, along with Alexandre Halbach, Asser Lahdemaki and Valtteri Lahtinen, with Andrew Tweedie joining as a fifth founder in 2024. The team numbers 40 people from 15 nationalities. Its product is Allsolve, a cloud-native multiphysics solver platform with fully programmatic Python control, positioned as a multiphysics simulation platform for the AI era.
On May 26, 2026 the company announced a 10 million euro Series A, roughly 11.6 million dollars, led by Danish quantum fund 55 North and Austrian industrial investor B and C Group, with existing investors Maki.vc, Crowberry Capital, QAI Ventures and First Fellow Partners all participating. Proceeds fund international expansion and a platform unifying simulation, quantum algorithms and AI integration.
There are three technology threads with completely different natures. The first is cloud-native solving: splitting simulation work into thousands of parallel cloud jobs with a full Python interface, which the company says delivers up to 100 times speedup and up to 99 percent runtime reduction. The second is quantum algorithms: Quanscient Quantum Labs focuses on the quantum lattice Boltzmann method for computational fluid dynamics, partners with IBM, IQM, Airbus, Embraer, Quantinuum and IonQ, employs eight quantum algorithm researchers including six PhDs and three professors, has filed seven patents and executed five world-first demonstrations on physical quantum hardware. The third is AI as a data engine: the company argues that simulation value is not only a result but the large-scale multiphysics data needed to train physics-aware AI.
The company cites its own study finding that 89 percent of engineers routinely simplify physical models to fit runtime and memory budgets. That number is the starting point of the commercial argument: if accurate, a platform that runs larger models faster addresses clear demand, and generating high-quality physical data is incremental value in the AI era.
Customer evidence is comparatively checkable. Albis Electronics runs 70 gigahertz radio frequency simulation in the cloud. skyTran cut runtime from three weeks to eight hours while improving accuracy from 10 to 3 percent. Infineon uses it for MEMS speaker design, UKAEA for high-temperature superconducting cables and coils, Proxima Fusion for stellarator superconducting magnets, and Tampere University for three-dimensional superconductor simulation. These customers concentrate in radio frequency and microwave, micro-electromechanical systems, superconductivity and fusion, which shows the actual foothold is not general CAE but problems desktop tools cannot handle: high frequency, multiscale and strongly coupled.
One framing point matters. This is a company where the technology narrative runs ahead of commercial scale. The 10 million euro Series A is its first disclosed round, the team is 40 people, revenue is undisclosed and customers are mainly research-intensive small firms and institutes. Its value lies in demonstrating two paths, the engineering usability of a cloud-native solver and early exploration of quantum algorithms for fluid dynamics, but in the near term it does not threaten Ansys, COMSOL or Siemens.
Company Overview
Origins and Team
Quanscient was founded in Tampere in 2021 with close ties to Tampere University, which appears among its simulation customers. The founding group spans commercial and numerical backgrounds: Riippi as CEO, Halbach with a background in electromagnetic and superconducting numerical computation, Lahdemaki and Lahtinen in engineering and product, and Tweedie joining in 2024 to fill a capability the company has not detailed publicly.
Forty people across 15 nationalities is micro-scale for simulation software. That scale determines product strategy: it cannot build dozens of physics modules as Ansys does, so it must go deep in a few scenarios that legacy tools cannot handle. The customer distribution confirms the choice of radio frequency and microwave, superconducting magnets, MEMS and fusion devices, all sharing high frequency, multiscale geometry and strong coupling that strain desktop memory and time budgets.
Funding Structure
The Series A investor structure is notable. Lead investor 55 North is a Danish quantum technology fund whose chief science officer Helmut Katzgraber is a well-known researcher in quantum annealing and optimization, while B and C Group is an Austrian industrial investor. Existing investors Maki.vc of Finland, Crowberry Capital of Iceland, QAI Ventures of Switzerland and First Fellow Partners all participated in full.
Two things follow. First, the narrative is quantum plus industry rather than cloud CAE alone, and a quantum specialist leading means investors are buying an entry point for quantum computing in engineering simulation rather than merely a faster simulation tool. Second, full participation by existing investors indicates the technical progress since earlier rounds was accepted.
Valuation, revenue and prior round sizes are undisclosed. A third-party database records total funding of 10 million euros across one round, which contradicts the reference to existing investors, since their participation implies earlier financing. That is a clear data gap.
Evolution of Positioning
The self-description shifted noticeably within two years. The earlier positioning was cloud multiphysics simulation, while after the Series A in 2026 it became cloud and quantum powered multiphysics plus AI-native hardware engineering. The ordering reflects strategic weight: cloud is the current product, quantum is a medium-term option and AI is narrative framing.
The shift is not a problem, since startup positioning commonly follows financing narratives, but it complicates assessment. If resources across three directions are unclear, the risk is doing all three without depth in any. Public information suggests quantum is not pure narrative, since eight of 40 people, or 20 percent, work on quantum algorithms.
Technology
What Cloud-Native Solving Implies Architecturally
The first capability is moving finite element solving to the cloud with large-scale parallelism. This sounds like installing desktop software on a server, but the architectural difference is substantial.
Desktop CAE assumes single-machine memory and cores, so model size is limited by workstation memory, typically 64 to 512 gigabytes, solvers are optimized for shared-memory parallelism, and pre-processing, solving and post-processing are coupled in one process. A cloud-native solver must address three new problems. Distributed memory parallelism with load balancing across thousands of cores requires mesh partitioning that scales. Task scheduling must decompose a parameter sweep into thousands of independent jobs allocated dynamically to available capacity. Data movement must be efficient between object storage and compute nodes so that it does not become the bottleneck.
The claim of thousands of parallel cloud simulations, if it refers to concurrent jobs in a parameter sweep, describes task-level throughput rather than latency reduction for a single large model. The 100 times speedup and 99 percent runtime reduction need that distinction. For sweep-type workloads such as frequency sweeps in antenna design or structural parameter optimization in MEMS, task-level parallelism approaches linear scaling and 100 times is attainable. For a single very large model, speedup is limited by parallel efficiency and communication overhead and rarely reaches two orders of magnitude.
The distinction matters because the customer cases illustrate it. skyTran's runtime fell from three weeks to eight hours, about 63 times, with accuracy improving from 10 to 3 percent. Part of the gain came from parallel throughput across design points and part from better algorithms and mesh quality, since the accuracy change means the model itself changed. Attributing it simply to faster cloud hardware would misread the case.
The Value of Full Python Programmability
The second feature is fully native Python, which matters more than it sounds in engineering practice.
Secondary development in traditional CAE typically relies on proprietary scripting languages or recorded GUI actions. The problem is separation from the mainstream scientific computing ecosystem of NumPy, SciPy, PyTorch and JAX, so custom constitutive models, coupling algorithms and optimization loops require reinventing infrastructure. Native Python means users define partial differential equations, material models, boundary conditions and coupling terms as scientific computing scripts, call libraries directly and embed simulation in machine learning pipelines.
Two user groups benefit most. Researchers doing multiphysics coupling need to couple electromagnetic, thermal and structural equations at the weak-form level rather than exchanging files between tools. AI for science teams need the solver as a differentiable layer or data generator embedded in training loops.
The GSS Audio testimonial states that Quanscient provides a pure Python implementation without layers of legacy code, allowing complex physics to be coupled from the core up, which points exactly at this capability. skyTran's engineer said that after recognizing improvements in speed and convenience the team decided to move completely to Allsolve.
Actual Multiphysics Coverage
The company says the platform unifies structural mechanics, fluid dynamics, electromagnetics and thermal equations in one composable environment. From customer cases the strongest areas are three: high-frequency electromagnetics such as 70 gigahertz radio frequency work at Albis and MEMS speakers at Infineon; superconducting electromagnetic-thermal coupling such as high-temperature superconducting cables and coils at UKAEA and stellarator magnets at Proxima Fusion; and acoustic-structural coupling such as audio devices at GSS Audio and MEMS speakers at Myvox.
Their shared technical signature is strong coupling plus multiple scales. Superconducting magnet current distribution depends on the magnetic field which depends on current distribution, while heat conduction and quench propagation must also be solved. MEMS speaker diaphragm vibration couples to the air acoustic field across micron to millimetre scales. A 70 gigahertz radio frequency structure has millimetre wavelengths with micron detail. All of these either exceed desktop capacity or require heavy simplification.
Conversely, Quanscient has no clear advantage in conventional structural strength, conventional fluid flow or conventional heat transfer, where Ansys, COMSOL and Siemens tools are already sufficient with complete ecosystems and extensive validation. Its opportunity is in new scenarios, and new scenarios are inherently smaller markets.
Quantum Lattice Boltzmann: Real Progress and Timescale
Quantum Labs is the most distinctive and most easily misread part. It works on the quantum lattice Boltzmann method, encoding the distribution functions of classical lattice Boltzmann into quantum states and implementing collision and streaming through unitary evolution, which theoretically offers exponential speedup for fluid dynamics on fault-tolerant quantum computers.
Disclosed progress includes partnerships with IBM, IQM, Airbus, Embraer, Quantinuum and IonQ, eight quantum algorithm researchers including six PhDs and three professors, seven patent filings, and five world-first demonstrations on physical quantum hardware. These are checkable facts showing it is not an empty research group.
The timescale must be stated clearly. Current hardware is in the noisy intermediate-scale quantum era, with qubit counts and coherence times far from running engineering-relevant fluid dynamics. Practical QLBM depends on fault-tolerant quantum computers, generally expected in the 2030s. Quantum will therefore generate no revenue for five to ten years, and its value is an option: if hardware matures, Quanscient holds algorithm and patent incumbency.
The company's own framing is honest in inviting partners to co-develop quantum-native algorithms today so that research is ready when hardware matures, which acknowledges immaturity. Assessment should treat quantum as research spending rather than a business line.
AI as a Data Engine: A Correct Argument
The most insightful claim is that simulation is the data source for AI. The logic runs that physics-aware AI needs large volumes of high-quality multiphysics data, real experimental data is scarce and expensive, traditional simulation tools are slow and require manual simplification so they cannot generate enough data, and a cloud-native solver lowers the cost of data generation to support physical AI training.
The direction matches NVIDIA PhysicsNeMo, Ansys SimAI and the broader narrative that simulation data is an asset, so it is industry consensus. Quanscient's difference is entering from the cost of data generation rather than from model architecture.
Delivering on the argument requires one proof: that the data it generates can train useful models. Public information contains no such evidence, with no customer having trained a usable surrogate on Quanscient-generated data and no published benchmark. That remains an unverified assumption.
Technical Shortfalls
First, physics coverage is narrow. The company emphasizes structures, fluids, electromagnetics and thermal, but actual cases concentrate in electromagnetics and superconductivity, with no public cases in turbulence, combustion, multiphase flow or contact nonlinearity.
Second, the verification and validation framework is unclear. Industrial buyers care most about result credibility, which requires standard benchmark comparisons, mesh convergence studies and correlation with experimental data. Ansys and COMSOL have decades of accumulation, and a company founded in 2021 cannot match that documentation depth, a real barrier in aviation and nuclear where regulation is strict.
Third, performance claims lack third-party benchmarks. The 100 times speedup and 99 percent runtime reduction do not specify the comparison target, model size or hardware configuration and have not been reproduced independently. The skyTran case at 63 times is a single data point.
Fourth, the quantum timescale and capital needs do not match. Quantum algorithm research is long-cycle and capital intensive, and a 10 million euro Series A cannot carry it to fault-tolerant hardware. That work depends either on external partnership funding from IBM or IQM or on dilution in a later round.
Product and Commercial Model
Allsolve Product Form
Allsolve is a cloud platform with a Python interface and a web interface. Core capabilities include geometry and meshing, weak-form definition for multiple physics, parallel cloud solving, visualization and post-processing, and integration with the Python scientific computing ecosystem. Pricing is usage-based on compute consumption with no upfront license fee, and trials are available including free compute hours or personal demos.
Allsolve Agent, released 3 September 2026 and open to all Allsolve customers, is the most material product change in this period. It is an assistant built directly into the platform: a plain-language description is turned into geometry, materials, physics, meshing and the solve itself, with no code to write, no separate environment to set up and no external LLM subscription required. The vendor describes five capabilities. First, building a simulation by describing it, including reasoning about which physics to couple and where boundary conditions belong, with an adjustable depth of reasoning from quick lookups to full builds. Second, catching mistakes before they cost time, by inspecting the model against the design intent when a result looks wrong and by reading solver logs, explaining the failure in plain language and proposing the fix, including in customised script sections. Third, turning a single unit cell into a full array with one grid command, which is what the PMUT and CMUT array work depends on. Fourth, a documentation tutor that searches the documentation, replicates a tutorial end to end from geometry through solve, or walks through it stage by stage and answers questions. Fifth, background execution: the agent asks before changing anything by default, can be put on auto-accept, and keeps track of long-running jobs while the user works on something else.
The vendor's own case study names a customer engineer: Jack Massaad, a senior development engineer at KROHNE, approached the agent as a skeptic and reports that building a complex parametric 3D ultrasonic flow-metering model with coupled acoustic-piezoelectric physics took less than a week, against the better part of a month by hand. This is a company-published customer case, not an independently verified measurement, and no method or baseline definition is published.
A second signal of the same direction is the Hack for Humanity: Finland edition, held on 3 October 2026 at Tampere University and hosted by TRES and AI Collective. About 80 people registered across three challenges, and ten teams of two to four people, roughly 30 participants in total and none of them previous Allsolve users, took the Allsolve challenge. Between 11:00 and 17:00 each team had to pick an engineering problem, build a web app that calls Allsolve through the SDK, run a real simulation and return an actionable result. All ten delivered a working app. The winner, HolkkiLab, a two-person team with a medical background, built a tool that designs patient-specific prosthetic socket liners: in its 80 kg demo case the app ran 96 parallel 3D FEM simulations in 66 seconds and the best liner cut the load at the limb tip by 31 percent, from 29.4 to 20.3 kPa, for under 1 mm of extra sink, with runtime staying nearly flat from 27 to 216 liners. The team states the model's own limits, namely a linear static tissue model, simplified limb geometry and no friction between limb and socket. Second place, ModalForge, generated footstep sounds by simulating floor vibration in 3D, at about two minutes per draft-quality solve. Third, Touchstone, designed PMUT cells for mid-air haptics, ran 228 simulations across 12 designs in parallel in 2.5 minutes and matched a published chip's resonance within 2 percent, while finding that combining single-cell results drifts 10 to 19 percent from a full array solve beyond two by two cells. The judging was performed by the company on five criteria, with the top three teams separated by one point, and the numbers above are the teams' own figures compiled by the vendor.
Read together, the agent release and the hackathon show the same thing from two directions: the time from a problem statement to a running simulation is falling, and the application layer is moving from the vendor's own interface into customer-built tools. Neither event changes the underlying constraints, which remain an undisclosed validation framework, no third-party benchmark for any performance claim, and an income base the company does not publish.
The pricing suits research-intensive customers, since small teams can validate value on a few cases before committing. For large enterprises, usage-based billing requires procurement alignment across budget lines, consumption monitoring and cost allocation, which is more complex than annual licenses.
Reading the Customer Cases
Each of the five public cases points somewhere different.
Albis Electronics runs 70 gigahertz radio frequency simulation against complex parasitic effects and expensive high-frequency hardware iterations, using cloud parallelism to iterate faster and reduce prototyping. The case shows usability in high-frequency electromagnetics.
GSS Audio works on complex physical coupling in audio equipment, valuing the pure Python implementation and core-level coupling. The case shows the value of programmability for unusual coupling problems.
skyTran works on maglev personal rapid transit, cutting runtime from three weeks to eight hours and improving accuracy from 10 to 3 percent. This is the most complete quantification and best illustrates the combined effect of cloud plus improved models.
Infineon uses Allsolve for MEMS speaker design and values effortless compute scaling for exploring larger design spaces. The case shows a large company willing to trial a new tool on specific projects.
UKAEA, the United Kingdom Atomic Energy Authority, and Proxima Fusion use it for high-temperature superconducting cables, coils and stellarator magnets. These cases show the technical threshold most clearly, since electromagnetic-thermal-mechanical coupling, quench propagation and alternating current loss in superconducting magnets are recognized hard problems that legacy tools handle at high cost.
The Reality of Commercial Scale
Revenue is undisclosed. With 40 people and European software cost structures, annual operating expense plausibly runs between five and eight million euros, so the 10 million euro Series A covers 18 to 24 months. Revenue is therefore likely in the low single-digit millions of euros, consistent with early commercialization.
The customer mix confirms this: research-driven small firms, institutes, and innovation projects at large companies rather than enterprise-wide deployments. Such customers bring limited revenue per account but decide quickly and accept technology risk.
Ecosystem and Partnerships
The quantum partnership network is among the most valuable intangible assets: IBM, IQM, Airbus, Embraer, Quantinuum and IonQ. The specific forms, whether joint research, project funding or technical validation, are undisclosed, but working with multiple quantum hardware vendors indicates Quanscient is regarded as a valuable algorithm partner in the QLBM niche.
The simultaneous presence of Airbus and Embraer is worth noting, since it places QLBM applications in aerodynamics and fluid dynamics, the most promising direction for quantum CFD.
Competitive Landscape
Cloud CAE Competitors
Cloud-native CAE is not new. SimScale is the most direct rival, a German company offering cloud computational fluid dynamics, structures, thermal and acoustics with subscription pricing, broad customer coverage and higher brand recognition. Rescale provides a cloud high performance computing platform running third-party commercial solvers including Ansys and COMSOL, competing on orchestration rather than owning solvers. OnScale built cloud multiphysics and was later acquired by Ansys.
Quanscient differs from all three through owning the solver, being programmable, and targeting strongly coupled scenarios. SimScale leads on usability and market coverage with less solver depth and weaker programmability, while Rescale can run any solver but owns none and cannot optimize at the architecture level. Quanscient chose the harder but more differentiated path.
Relationship With Desktop Incumbents
Ansys, COMSOL, Siemens and Dassault all offer cloud deployment, but their core products were designed for desktops, so cloud is hosting rather than rewriting for cloud. The difference is invisible on small models and material on very large models and large parameter sweeps.
The real incumbent advantage lies elsewhere: completeness of physical models with dozens of modules and thousands of material models, decades of validation data and experimental correlation, and embedding in customer workflows from pre-processing through solving, post-processing, reporting and data management. Quanscient is at the beginning on all three.
The realistic relationship is therefore complementary rather than substitutive: customers use incumbents for mature scenarios and Quanscient for new ones such as high frequency, superconductivity and strong coupling. That position can survive but has a limited ceiling.
Structure of the Quantum Simulation Field
Three groups participate in quantum computing for engineering simulation. Hardware vendors including IBM, IQM, Quantinuum, IonQ and PsiQuantum explore applications themselves. Algorithm and software companies including Classiq, QC Ware, Zapata and Quanscient build algorithms and compilation. Industry users including Airbus, Embraer, aero engine makers and energy companies validate scenarios.
Quanscient occupies the algorithm plus scenario intersection: it builds no hardware but has taken one specific algorithm to the point of validation with multiple hardware vendors. The value of that position depends on whether QLBM becomes the mainstream route for quantum CFD, and it would weaken if the field shifts toward quantum linear solvers or other methods.
Financing and Capital Structure
Structure and Meaning of the Series A
The 10 million euro Series A of May 26, 2026 was co-led by 55 North and B and C Group with full participation from Maki.vc, Crowberry Capital, QAI Ventures and First Fellow Partners. One lead is a quantum specialist and the other an industrial investor, reflecting a dual bet on quantum technology and industrial deployment.
Katzgraber of 55 North framed the logic clearly, describing the cloud-native multiphysics platform combined with forward-looking quantum algorithms and AI tools as giving innovators in fusion, advanced electronics and quantum technologies a throughput step change without sacrificing accuracy. The purchase is throughput, meaning the ability to explore larger design spaces at equal accuracy.
Capital Efficiency and Funding Needs
Ten million euros supports 40 people for 18 to 24 months, which is moderate capital efficiency. Compared with cloud CAE peers such as SimScale with over 100 million dollars raised and Rescale with over 300 million, Quanscient's scale is an order of magnitude smaller, which forces a strategy of differentiated scenarios and high value per account rather than broad coverage.
Quantum research funding is a separate question. Eight researchers including three professors cost one to two million euros annually, and quantum hardware access through IBM and IonQ cloud services adds per-use fees, so that work consumes two to three million euros per year. Within a 10 million euro total that is a substantial share for a line with zero revenue for five to ten years.
Key Undisclosed Data
Valuation, revenue, gross margin, retention and prior round sizes are all undisclosed. That is normal for an early-stage company but means unit economics cannot be judged externally, particularly because under usage-based pricing the share of cloud infrastructure cost in revenue sets the gross margin ceiling. If customers run large sweeps at prices near cost, growth would deepen losses.
DISCREPANCIES
1. Currency and amount of the financing. The official figure is 10 million euros, with some coverage converting to about 11.6 million dollars. Both refer to the same round with exchange rate differences. This report uses 10 million euros, roughly 11.6 million dollars. Source tier: OFFICIAL (company announcement).
2. Cumulative funding and round count. A third-party database records 10 million euros total across one round, yet the Series A announcement names Maki.vc, Crowberry Capital, QAI Ventures and First Fellow Partners as existing investors participating in full, which implies at least one earlier round of undisclosed size. This report states that the first disclosed round is the 10 million euro Series A and earlier rounds are undisclosed. Source tier: OFFICIAL (announcement) over third-party database.
3. Founding date. A third-party database records incorporation on May 26, 2026, which is actually the Series A announcement date, while official and multiple sources agree on 2021. This report uses 2021. Source tier: OFFICIAL.
4. Quantum-inspired versus quantum algorithms. Third-party material says Allsolve uses quantum-inspired algorithms, while the official description is that Quantum Labs researches quantum algorithms running on quantum hardware, specifically QLBM. The former refers to classical algorithms mimicking quantum ideas and the latter to algorithms executing on quantum processors, which are technically different. This report follows the official distinction. Source tier: OFFICIAL over third-party.
5. Framing and scope of the speedup claim. The company says up to 100 times faster and up to 99 percent runtime reduction, which are mathematically equivalent statements of the same fact. The claim does not specify the comparison target, model size or hardware, nor does it separate task-level parallel throughput from single-model latency. The published skyTran case at roughly 63 times is below 100. This report separates the two kinds of speedup in the technology section and labels 100 times as company framing. Source tier: OFFICIAL, company claim without third-party benchmark.
6. Source of the 89 percent survey. The company states its research shows 89 percent of engineers routinely simplify physical models to fit runtime and memory budgets, without disclosing sample size, respondent composition, methodology or timing. This is company research rather than independent study. This report retains the figure with its source character noted. Source tier: company research, UNVERIFIED.
7. Exclusivity of the first cloud-and-quantum claim. The company describes itself as the first cloud-and-quantum-powered multiphysics simulation software provider. The claim cannot be verified and depends on defining cloud and quantum powered, since it is rare to combine cloud solving with quantum algorithm research but other participants exist in quantum simulation. This report does not use the exclusivity claim. Source tier: company marketing, UNVERIFIED.
8. Composition of the team figure. The company states a team of 40 from 15 nationalities and separately that Quantum Labs has eight quantum algorithm researchers including six PhDs and three professors. Three professors among eight usually indicates academic affiliation or advisory roles rather than full-time employment. This report keeps the original wording without inferring full-time ratios. Source tier: OFFICIAL (company statement) with composition detail undisclosed.
9. Vendor-reported scope of the agent and hackathon results. The company's blog states that Allsolve Agent is available to all Allsolve customers, that a KROHNE engineer built a complex parametric 3D ultrasonic flow-metering model in less than a week against a hand-built baseline of the better part of a month, and that the winning hackathon team ran 96 parallel 3D FEM simulations in 66 seconds while reducing limb-tip load by 31 percent from 29.4 to 20.3 kPa. Adjudicated: retained as company-reported figures, with the KROHNE statement attributed to a named customer engineer and the hackathon numbers labelled as team-reported results rather than audited benchmarks. The KROHNE figure comes from a case study published by the vendor and the hackathon figures were compiled by the vendor from the teams' own submissions, so neither has independent verification, and the winning team itself lists a linear static tissue model, simplified geometry and no friction as the model's limits. The three-way gap between the hackathon figures, the vendor's own evaluation of the model and an independent reproduction is recorded here rather than smoothed over.
Risks and Diligence Gaps
First, revenue and unit economics are entirely opaque. Revenue, gross margin, retention and average contract value are undisclosed. Under usage-based pricing, cloud infrastructure is the main variable cost, and if prices are close to cost, scale does not improve profit. This is the core gap in judging whether the model works.
Second, the validation framework is untested. Industrial buyers treat verification and validation documentation as a hard threshold. With only five years of history, Quanscient cannot match incumbents in benchmark libraries, mesh convergence studies and experimental correlation, which may block entry in aviation, nuclear and medical markets.
Third, quantum timelines and capital are mismatched. Practical QLBM depends on fault-tolerant hardware expected in the 2030s, and a 10 million euro Series A cannot carry the company there. The work needs continuing external funding from partnerships or government grants, or later rounds. Slower hardware progress would extend the payback period further.
Fourth, customers concentrate in niche scenarios. Radio frequency and microwave, superconducting magnets, MEMS speakers and fusion devices have high technical barriers but limited market size. Without expanding from these beachheads into broader industrial scenarios such as mainstream structural and fluid analysis in automotive and aviation, the revenue ceiling stays low.
Fifth, coexistence with cloud providers. Allsolve runs on public cloud, with the provider undisclosed, while cloud vendors themselves promote simulation services. If a cloud provider builds its own capability or acquires a competitor, Quanscient's bargaining power falls.
Sixth, talent dependency. Eight of 40 people work on quantum algorithms, and loss of core researchers including the professors would directly affect that line. Finland's quantum talent pool is limited and competes with local companies such as IQM.
Seventh, geopolitics and export controls. Customers include UKAEA in fusion, Infineon in semiconductors and aerospace partners in the quantum work. These domains face export controls and investment screening, which can constrain cross-border business.
Composite Assessment
Technology scores 7.5. Cloud-native architecture and full Python programmability are genuine technical differences, strongly coupled scenarios in superconductivity, high frequency and MEMS have verifiable customer cases, and QLBM research has patents, hardware demonstrations and multiple vendor partnerships rather than being an empty shell. Deductions for narrow physics coverage, undisclosed validation framework and performance claims without third-party benchmarks.
Business scores 6.5. The customer list is respectable with Infineon, UKAEA, Proxima Fusion and aerospace partners, scenario selection is precise and usage-based pricing lowers trial thresholds. Deductions for very small revenue, concentration in niche scenarios and opaque unit economics.
Finance scores 6.0. The 10 million euro Series A with a sensible investor structure combining a quantum specialist and an industrial investor shows moderate capital efficiency. Deductions for total funding an order of magnitude below peers, quantum work consuming capital without near-term revenue, and undisclosed valuation and revenue.
Product scores 7.0. The Allsolve architecture of cloud parallelism, native Python and composable multiphysics is directionally right, interfaces and workflows are well received by early customers, and trial mechanics are friendly, and since 3 September 2026 the platform has shipped Allsolve Agent for all customers, with the October 2026 hackathon showing ten non-users building working simulation applications in one afternoon. Deductions for maturity gaps against incumbents, undescribed post-processing and data management depth, missing industry templates and compliance modules, and the fact that the agent's evidence is vendor-published rather than independently verified.
Market scores 6.5. Cloud CAE is a clear trend, quantum simulation is an imaginative long-term direction and demand in strongly coupled scenarios is real. Deductions for limited market size, SimScale's lead in cloud CAE awareness, and incumbent cloud moves compressing differentiation.
Overall 6.7. This is a company with leading architecture, minimal scale and a correct path: it made cloud-native and programmability engineering-usable, built beachheads in superconductivity and high frequency where legacy tools strain, and invested substantively in quantum CFD. Yet revenue scale, the validation framework and the quantum timescale mean it remains a niche player for three to five years. The judgment on physical AI is that the simulation-as-data-source argument is correct and cloud-native architecture is indeed a prerequisite for generating large-scale physical data, but turning the argument into a business requires proving that the generated data trains useful surrogates and that the platform gains validation acceptance in mainstream industrial scenarios.
References
- Quanscient Series A announcement, May 26, 2026, Tampere: 10 million euro round led by 55 North and B and C Group with Maki.vc, Crowberry Capital, QAI Ventures and First Fellow Partners participating in full; founded 2021 by Juha Riippi, Alexandre Halbach, Asser Lahdemaki and Valtteri Lahtinen with Andrew Tweedie joining as fifth founder in 2024; team of 40 from 15 nationalities; platform up to 100 times faster than traditional desktop tools with runtime reduced up to 99 percent; company research finding that 89 percent of engineers simplify physics models to fit runtime and memory budgets
- Quanscient website, accessed 2026: Allsolve positioned as a multiphysics simulation platform for the AI era supporting thousands of parallel cloud simulations and fully Python-native programmatic control; Quantum Labs researching the quantum lattice Boltzmann method for computational fluid dynamics with partners IBM, IQM, Airbus, Embraer, Quantinuum and IonQ; eight quantum algorithm researchers including six PhDs and three professors, seven patent filings and five world-first demonstrations on physical quantum hardware
- Quanscient customer cases from the website: Albis Electronics running 70 gigahertz radio frequency simulation in the cloud against complex parasitic effects; GSS Audio using a pure Python implementation to couple complex physics from the core up; skyTran cutting runtime from three weeks to eight hours and improving accuracy from 10 to 3 percent; Infineon using it for MEMS speaker design and larger design space exploration; kiutra reducing hardware iterations; IXI finding a working design on the first iteration and saving three months; UKAEA modelling high-temperature superconducting cables and coils in three-dimensional multiphysics; Proxima Fusion simulating stellarator superconducting magnets; Tampere University for three-dimensional superconductor simulation; Myvox for MEMS speakers
- Helmut Katzgraber, chief science officer of 55 North, in the Series A announcement: describing the cloud-native multiphysics platform combined with quantum algorithms and AI tools as giving customers a throughput step change without sacrificing accuracy in fusion, advanced electronics and quantum technologies
- Julia Reilinger, managing director of B and C Group, in the Series A announcement: describing Quanscient at the intersection of simulation, quantum computing and AI as setting a new standard for physical product development and strengthening Europe's long-term industrial innovation capacity
- Juha Riippi, CEO, in the Series A announcement: AI will not transform hardware engineering unless simulation itself is rebuilt for it; making multiphysics code-driven and cloud-scalable generates the volume of physics data AI needs, turning simulation from a bottleneck into the engine of data-driven design; the architecture built for cloud and quantum is also the foundation for physics-aware AI models
- Finnish first-half 2026 funding statistics from EU Information Service: Quanscient raised a 10 million euro Series A in the first half of 2026 for international expansion and development of its cloud simulation platform, AI capabilities and quantum computing technology for hardware engineering; other Finnish rounds in the period included Algorithmiq at 19.7 million euros, Kelluu at 15 million, Capalo AI at 11 million and Vexlum at 10 million
- Chinese coverage, 2026: the 10 million euro Series A equals about 11.6 million dollars, funding international expansion and the cloud and quantum powered multiphysics platform; the platform unifies structural mechanics, fluid dynamics, electromagnetics and thermal equations in a single composable environment; customers span Fortune 100 firms and industrial manufacturers in Europe, North America and Japan
- Third-party funding database, startup.eu: records Quanscient as a Finnish company with 10 million euros raised across one round from six investors and undisclosed valuation; records incorporation on May 26, 2026, which conflicts with the official 2021 founding; describes differentiation as cloud-native API-first architecture, massively parallel simulation, usage-based pricing with no upfront license fee, and integration of quantum algorithms and AI workflows
- Third-party technology media description of the platform: cloud-native architecture delivering up to 100 times faster simulation than traditional desktop tools with overall compute runtime reduced up to 99 percent; treating simulation data as a foundational data source for generating the large-scale multi-domain physics datasets needed to train generative AI design models
- Competitive information from public sources: SimScale offering cloud computational fluid dynamics, structures, thermal and acoustics with subscription pricing; Rescale providing a cloud high performance computing platform hosting third-party commercial solvers; OnScale having built cloud multiphysics before acquisition by Ansys; Ansys, COMSOL, Siemens and Dassault offering cloud deployment of desktop-architected products
- Industry background on quantum computing for engineering simulation: current hardware in the noisy intermediate-scale quantum era with qubit counts and coherence times insufficient for engineering-relevant fluid dynamics; the quantum lattice Boltzmann method encoding distribution functions in quantum states and implementing collision and streaming through unitary evolution, theoretically offering exponential speedup with practical use dependent on fault-tolerant hardware
- Quanscient company blog, "The Allsolve Agent is here: 5 ways it can make your life easier", 3 September 2026, by Juha Riippi, for the Allsolve Agent being available to all Allsolve customers, its five capabilities covering plain-language modelling, result troubleshooting, array generation, documentation tutoring and background execution, and for the KROHNE case in which senior development engineer Jack Massaad reports a complex parametric 3D model in less than a week against the better part of a month by hand
- Quanscient company blog, "What 30 first-time users built with the Allsolve SDK in six hours", 6 October 2026, by Jukka Knuutinen, for the Hack for Humanity: Finland edition at Tampere University on 3 October 2026 hosted by TRES and AI Collective with about 80 registered and ten teams of two to four people, all ten delivering working applications by the 17:00 deadline, and for the winning team HolkkiLab running 96 parallel 3D FEM simulations in 66 seconds to cut limb-tip load by 31 percent from 29.4 to 20.3 kPa, for second place ModalForge on floor-vibration footstep synthesis and third place Touchstone on PMUT array design with 228 simulations across 12 designs in 2.5 minutes
This report is compiled from public sources for informational purposes only and is not investment advice. The company is private and revenue and financial data are undisclosed; performance claims and survey figures are company-reported without independent third-party verification, and related judgments are labeled as uncertain. Data is current as of 2 November 2026.
Related tools
Related cases
Related reports66
- 51World Deep Dive: A Listed Digital Twin Company Where 51Sim Becomes the First Growth Engine While 51Earth Nearly Vanishes, and Two Irreconcilable Financing Naming Conventions Must Both Be Kept
- AMI Labs: A USD 1.03B Seed Bet on World Models, LeCun's Non-Generative Path and Its Blank Scorecard
- Anshi Pacific (PERA) Deep Dive: A Beijing Domestic-CAE Leader Founded in 2003 Whose PERA SIM General Simulation, AIROM Reduced-Order Models, AgentX Multi-Agent and Jingzhi iGPT/Whale Span from Industrial Simulation to Physics AI
- AspenTech Deep Dive: The Process-Simulation Incumbent Acquired by Emerson and Its Hybrid Industrial-AI Route
- Autodesk Deep Dive: Neural CAD Foundation Models and a $200M World Labs Bet on Physical-World AI, with NAVASTO Adding CFD Surrogate Capability
- Bentley Systems Deep Dive: iTwin Digital Twin and Generative AI Empower Full-Lifecycle Infrastructure Engineering
- BeyondMath Deep Dive: A Foundation Physics AI on the Neural-Operator Route, a $19M Consortium Budget, and Three Misread Facts
- Cadence Deep Dive: System-Level Multiphysics EDA and the Millennium M2000 Supercomputer Anchor Physical AI, with AuraStack Agentic AI and Fidelity GPU CFD as Key Levers
- COMSOL Deep Dive: Four Decades of Multiphysics Solvers, Entering Physical AI via DNN Surrogate-Model Training and All-Physics GPU Acceleration
- Dassault Systèmes Deep Dive: 3DEXPERIENCE and SIMULIA Virtual Twin Physics Behavior Anchor Industrial AI, with LEO Virtual Companion and GPU Explicit Dynamics as Physical-AI Levers
- DP Technology: From the Gordon Bell Prize to National-Level Unicorn Status and the Door of a Hong Kong Listing, the AI-for-Science Leader's Technical Delivery and Commercial Blank
- Emmi AI: From Linz Spin-out to a EUR 330M Mistral Acquisition in 17 Months
- Fainite Deep Dive: A Zurich Physics-AI Simulation Startup Whose Caltech, ETH Zurich, Cambridge and Google Team Uses Physics-Aware Deep Learning to Cut Simulation from Tens of Hours to Seconds, With Only Venture Kick Grants and No Priced Round
- Fysics AI: The PhysX Founder's Second Act, a Domestic Differentiable Physics Engine and Its Ecosystem Play
- Frontop (301313) Deep Dive: From Digital Exhibition Halls to a Physical-AI Data Foundation - FTE Engine, FunSim Simulation and the Gap Between Positioning and Revenue
- Godela Deep Dive: A YC X25 Physics-AI Simulation Startup Whose Ex-Apple, Google and Intel Team Uses Physics-Constrained Surrogate Models to Cut Simulation from Days to Seconds, at Seed Stage with Pilot-Only Access
- HollyWave (688682) Deep Dive: RDSim AI Electromagnetic Simulation and Engineering Agents Close the Simulation-Measurement-Product Loop
- IAT (Alt) Deep Dive: China's First Listed Independent Automotive Design Company Pivots to AI-native R&D with the Yuanfang, Yufeng and Taiyi 3.0 Product System and an NVIDIA Omniverse Partnership, Reporting RMB 975M Revenue and a Widened RMB 250M Net Loss in 2025
- Inductiva Deep Dive: A Closed Case in Open-Solver Orchestration, Eight Premises That Must Be Corrected, and Where Co-Founder and CTO Hugo Penedones Went Next
- Intesim (INTESIM) Deep Dive: A Home-grown Chinese CAE Leader Founded in 2009 with Self-developed Solvers for Structure, Fluid, Electromagnetics, Thermal and Acoustics and Multiphysics Coupling, a ~RMB 100M B Round from CNII and Shenzhen Capital, and 2025 Revenue and Bookings Both above RMB 100M
- iSoftStone (301236) Deep Dive: A ChiNext-Listed IT Services Giant Whose TianShu Platform Builds a Three-Layer Physical-AI Stack with Self-Developed TwinHub Digital Twin, FSim Industrial Simulation, CAE Multiphysics and Phyxis Embodied Control
- Jua Deep Dive: The Earth Physics Transformer as a Large Physics Model for Energy Trading
- Keysight (with ESI Group) Deep Dive: Virtual Prototyping and Hybrid AI Bridge Measurement and Simulation, with Natural-Language Assistants in EDA
- Lightwheel Deep Dive: A Physical AI Data and Evaluation Infrastructure Startup Founded by a Former NVIDIA Simulation Lead, Seven Rounds Totalling Roughly CNY 2.65 Billion at a Valuation Above USD 2 Billion, Seats on Both the Newton and EgoVerse International Standards Committees, With Two Incompatible Accounts of Q1 Orders
- Luminary Cloud: The Physics AI Model Factory Backed by NVentures
- MathWorks Deep Dive: MATLAB/Simulink and Simscape Multiphysics Power Physical AI, with R2026a Copilots and an Agentic Toolkit for Edge Intelligence
- Maniformer (Mifeng) Deep Dive: An Agibot-Spun Physical-AI Data Infrastructure Firm Whose Bodyless MEgo Capture Hardware Is in Mass Production and Which Closed Three Rounds in Half a Year Led by China Telecom
- Mistral AI: How Physics AI Moved From Language Models to EUV Lithography and Nuclear Power
- Monolith AI: The British Engineering-AI Firm That Cut Physical Test Counts Instead of Accelerating Simulation, Now Inside CoreWeave
- Monumo Deep Dive: A Cambridge Deeptech Company Redesigning Electric Motors with the Anser AI Engine, Tens of Millions of Multiphysics Simulations per Day, GBP 10.5M Raised in Total, and a Series A Being Raised
- NAVASTO: Moving Solver Insight Onto the Designer's Desk, and How Autodesk Bought That Entry Point
- Navier: Handing CFD Workflows to Agents, a Seed-Stage Bet on AI for Hardware Engineering
- Neural Concept: An EPFL Spin-out Betting on Neural Surrogates, with $138M Raised
- NP Co. (NP Company) Deep Dive: A 2025-founded Paris Inria-born AI-native simulation startup that raised a EUR 6M pre-seed led by Partech with Mistral AI co-founders as angels, claiming a pre-trained physics foundation model at 1,000x speedup — most claims unverified
- NVIDIA Deep Dive: Building the Full-Stack Physical-AI Infrastructure with PhysicsNeMo, Warp, Cosmos World Foundation Models and Omniverse
- Physical Intelligence Deep Dive: A $2.1B Bet on the GPT Moment for Robotics, With pi0.7 Self-Reporting 60-80 Percent Success on Unseen Tasks and Neither Pricing nor Revenue Disclosed for the Pi API
- Physical Superintelligence Deep Dive: A $58M Seed Bet on Virtual Physicists, With Emmy Targeting AI Data Center Multiphysics Optimization as Its First Commercial Proof Point
- PhysicsX: Pretrained Geometry and Physics for Industrial Engineering at a $2.4B Mark
- Prometheus: A $41B Mark, $18.2B Raised, and Not Yet a Product
- PTC Deep Dive: Creo/Onshape and the ThingWorx Digital Thread Power Physical AI, with NVIDIA Isaac Sim Bridging Robotics Design to Simulation
- Qianjue Technology Deep Dive: How a Body-Agnostic Embodied Brain Pairing the Predictive World Model Polibrain with Brain-Inspired Partitioning and Polynomial Representation Reached a Hundred Thousand Cleaning and Service Robots
- Rescale Deep Dive: An HPC Operating System and Desktop Engineering Workbench in the Cloud, a $115M Series D, and Five Misread Facts
- Schrödinger Deep Dive: Physics-Based FEP+ Simulation and Machine Learning Power Molecular Discovery, with the 2026 Bunsen Agentic Co-Scientist
- Secondmind Deep Dive: A Cambridge Engineering-AI Firm Whose Bayesian Active Learning Cuts CAE Simulation and Test-Bench Runs by up to 90%, with Roughly USD 73M Raised and Mazda Both Customer and Investor
- Shichuang Technology Deep Dive: SupreCAST Cloud-Native Material-Processing Simulation and Industrial Agents Advance the CAE+AI Strategy
- ShonCloud (Shunyun) Deep Dive: Meshless Particle-Method CFD for EV and Nuclear Fluid Multiphysics
- Shuzhi Maijin Deep Dive: A Shenzhen Fully Self-Developed SNN-plus-LBM CFD Startup Claiming Second-Level Results on Hundreds-of-Millions of Grids on Consumer Hardware, With a Tens-of-Millions-RMB Angel Round From a Ministry-of-Finance-Affiliated Fund and an FPGA Migration Plan
- Siemens (incl. Altair) Deep Dive: The 2025 Altair Acquisition Unites Simcenter and Altair, with PhysicsAI Geometric Deep Learning and PhysicsAI Generate as the Physical-AI Core
- Silurian Deep Dive: A Generative Forecasting Transformer Earth-System Foundation Model for Weather Risk on Energy and Infrastructure Assets
- SimScale Deep Dive: A Cloud CAE Platform Built on Core-Hour Subscriptions, EUR 52M of Equity Kept Separate from EUR 25M of Debt, and Five Misread Facts
- Skild AI Deep Dive: A General Robot Foundation Model Incubated at CMU's Robotics Institute, With a USD 1.4B Series C at a USD 14B Valuation, a Conflicting Series B Size, and an ARR Figure That Is a Run-Rate Rather Than Revenue
- Songying Deep Dive: A Huawei-Cloud-Alumni Startup Building a Domestic Physical AI Operating System, With ORCA OS Positioning Against NVIDIA Omniverse, Three Incompatible Financing Narratives, and a Verifiable Third-Party Anchor in a CNY 20 Million Minority Investment
- Sudo Tech Deep Dive: Zero-Real-Data Simulation Training, the Fully In-House Sudo R1 Stack, and an Atomic-Skill Commercialisation Strategy from a Startup Valued Above CNY 20 Billion in 17 Months
- Suochen Technology: The Pain of Becoming a Physics-AI Platform from a CAE Vendor
- Synopsys (Ansys) Deep Dive: The 2025 Ansys Acquisition Turns the EDA Leader into a Silicon-to-Systems Engineering Powerhouse, with Ansys SimAI and the NVIDIA Physical-AI Partnership as the Core Physics-AI Levers
- TenFong Deep Dive: An Academician-Led Chinese Industrial Simulation Software Vendor Whose Official Timeline and Database Records Use Incompatible Round Names, With an Official Financing PDF and an Investment Database Both on File
- Tianfu (Nanjing Tianfu Software) Deep Dive: A 2011-founded Chinese Intelligent-CAE Veteran That Closed a Several-Hundred-Million-RMB Series C, Pivoted to an Industrial World Model in 2024, With Unreconcilable Financing and Valuation Records and No Disclosed Revenue
- Tongyuan Soft Control Deep Dive: MWORKS (Syslab/Sysplorer) as a Domestic MATLAB/Simulink Alternative, with MWORKS AI and the MOGICK Engineering Agent Advancing Home-Grown Physical AI
- Vinci AI: Turning Physics Simulation from a Checkpoint into a Continuous Constraint
- Westlake Robotics Deep Dive: Westlake University's First Robotics Spin-out Builds a Dual-Pretrain Brain-and-Cerebellum Architecture with WR1 Whole-Body Foundation Model and GAE Behaviour Expert
- World Labs: Fei-Fei Li's Spatial-Intelligence Unicorn, Acquired by AMD for $8.2B
- XVERSE Deep Dive: A Simulation-World Route to Physical AI Built on the XVERSE Model Family, an Edge-Cloud 3D Engine and the Physics-Based MotionGen Model
- SimWE: From In-House Solvers to a Physical-AI Engine, the Domestic CAE Leader and What Comes Next
- Yunjing Zhifang Deep Dive: A Peking-University-Incubated Cloud-Native GPU CAE Company with the Simversus Platform and the YJ-DORGP Small-Sample Intelligent Solver, Reporting 275 Percent Order Growth, 160 Percent Revenue Growth and Breakeven in 2025
- Zhongke Diwiji (Five Ages) Deep Dive: A CAS-Institute-of-Automation and Tsinghua Spin-out Whose FAM Embodied Foundation Model Uses Action Optical Flow for Cross-Embodiment Generalisation, Reaching 97% Success from 3 to 5 Demos, with EraBrain and Jiling Robots Serving Sinopec, Leapmotor and Midea
- ZWSOFT (688083) Deep Dive: Overdrive Kernel and ZW3D Wukong 2027 Underpin an Integrated CAD/CAE/CAM Stack, with a CAx Agent Platform as the Practical Route to Engineering AI
This report is compiled from public sources. It is for reference only and does not constitute investment advice. Data is current as of the publication date; please refer to the company’s official filings.