TensorX
- Company typePrivate
- Founded2026
- HeadquartersDublin, Ireland
- Headcount1–10
- GTM typeB2B
- OfferingSoftware
TensorX firmographics
Firmographics- Name
- TensorX
- Legal name
- TensorX Ltd.
- Website
- https://tensorx.ai
- Company type
- Private
- Founded year
- 2026
- Operating status
- Operating
- Headcount range
- 1–10 employees
- Ownership category
- akta.pro rank
TensorX industry classification
Industry- Product category
- AI Inference Platform
- NAICS
- Computer Systems Design and Related Services (54151), Computer Systems Design and Related Services (5415)
- SIC
- Services-Computer Programming, Data Processing, Etc. (7370), Services-Computer Programming Services (7371)
- akta.pro primary industry
- Model Hosting, Serving & Inference Platforms (HDAAACAB)
- akta.pro secondary industries
- AI Server Systems & HGX/Accelerator Platforms (HDAAAAAB), GPU-Accelerated & AI Training/Inference Servers (HDACABAG)
Keywords
Where TensorX is headquartered
LocationHeadquarters
- HQ city
- Dublin
- HQ country
- Ireland
- HQ region
- Europe
Offices3 records
Markets served
TensorX business model
Business model- GTM type
- B2B
- Offering type
- Software
- Cost components
- Technology or R&D, Infrastructure, Personnel, Marketing or Sales, Supply Chain, Operations
Revenue model
- Pay-per-Token (Usage-Based) Inference: Customers create an account, buy credits, and pay per million tokens for API access across 26+ open-weight models. Pricing is per 1M tokens for input, cache read, and output. No monthly fees, no minimums. This is the primary self-serve revenue stream targeting developers and AI builders.
- Dedicated Sovereign GPU Clusters: Enterprise customers rent physically isolated GPU clusters on TensorX-owned EU infrastructure with fixed monthly pricing. Includes 30+ model access with full customisation, ISO 27001-ready infrastructure, and Enterprise SLA. 1–3 business day setup. Additional GPU capacity on-demand for burst traffic.
- Shared Inference Clusters: Shared GPU clusters with vetted, non-competing partners. Cost-effective GPU access with 30+ managed models, EU-sovereign infrastructure, fully managed by TensorX with zero data retention on every request.
Pricing tiers
| Model | Billing | Price |
|---|---|---|
| Usage-based | Pay-as-you-go | Pay-per-Token: Usage-based API access for developers and AI builders |
| Subscription | Monthly | Dedicated Sovereign Inference: Physically isolated GPU cluster for enterprises with fixed monthly pricing |
| Usage-based | Pay-as-you-go | Shared Inference Clusters: Multi-tenant GPU access with vetted non-competing partners |
Go-to-market motion3 records
Distribution channels4 records
Marketing channels6 records
TensorX product offering
Product offeringCore offering
TensorX provides EU-sovereign AI inference infrastructure, running 26+ open-weight large language models on TensorX-owned NVIDIA Blackwell B300 GPUs hosted in EU data centres (Dublin, Helsinki, Paris). The platform serves regulated industries (financial services, healthcare, legal, government, aviation) through two product lines: a pay-per-token OpenAI-compatible API for developers and dedicated physically isolated GPU clusters for enterprises. All inference operates with zero data retention by design — prompts and completions are processed in ephemeral enclaves and never persisted.
Product overview
TensorX is a sovereign AI inference platform providing EU-hosted private access to 30+ open-weight AI models on NVIDIA Blackwell GPUs with zero data retention. The core offering consists of two tiers: Dedicated Sovereign Inference (physically isolated GPU clusters for enterprises requiring fixed-capacity, custom configuration, and enterprise SLAs) and Pay-per-Token Private Inference (usage-based API access for developers and AI builders). The TensorX Model Library provides transparent access to models including GLM-5.2, MiniMax-M3, DeepSeek-V4, Kimi-K2, and Qwen variants, benchmarked via the Artificial Analysis Intelligence Index. The platform targets regulated industries (finance, healthcare, legal, government) requiring GDPR compliance and EU data sovereignty. An optional Workload-Matched Inference Engineering engagement helps enterprises size and migrate to dedicated infrastructure. All inference runs on TensorX-owned GPUs in Dublin, Helsinki, and Paris with zero data retention by design.
Differentiator
Problem solved
Functional benefit
Products and services
- TensorX Private AI Inference Platform EU-sovereign AI inference platform providing private, zero-data-retention access to 26+ open-weight language models on NVIDIA Blackwell GPUs. Serves regulated industries including finance, healthcare, legal, and government with EU data residency and GDPR compliance.
- Dedicated Sovereign Inference Physically isolated GPU clusters on EU-sovereign infrastructure (Dublin, Helsinki, Paris) with zero data retention, fixed monthly pricing, full model customisation, and enterprise SLAs. Targets mission-critical applications in financial services, healthcare, aviation, and government.
- Pay-per-Token Private Inference Usage-based private inference via OpenAI-compatible API enabling instant account setup and access to 26+ models with zero data retention and 60–97% cost savings versus US providers.
- TensorX Model Library Catalogue of 26+ open-weight AI models available for inference including GLM-5.2, MiniMax-M3, DeepSeek-V4-Pro, Kimi-K2.6/2.7, Qwen3 variants, and others, with transparent benchmark scores and per-token pricing.
- TensorX API OpenAI-compatible REST API providing programmatic access to private AI inference, supporting drop-in replacement for OpenAI applications with a single line of code change and EU data residency.
Quantifiable outcome
- 60–97% lower inference cost vs OpenAI GPT-4o and Anthropic Claude 3.5 Sonnet
- +4 more outcomes
Companies that use TensorX
Customer profileNamed customers3 records
Segments6 records
Ideal customer profiles3 records
TensorX technology and API
TechnologyTechnology focussed Yes
API detail
- Has API
- Yes
- API docs
- API detail
Core technology
AI maturity
App detail
Integration1 record
AI capability7 records
Feature7 records
TensorX partnerships and signals
Strategic signalPartnerships
Six partnerships are on record, tiered core and strategic.
- NVIDIAcoreOfficial member of the NVIDIA Inception Program, receiving accelerated development support, direct access to NVIDIA engineers and latest GPU architectures, and connections to the broader NVIDIA ecosystem. TensorX deploys NVIDIA Blackwell B300 GPUs on its sovereign infrastructure.
- DellcoreHardware sourcing partner for NVIDIA GPU infrastructure. Darius Cubed Ventures committed €6.5 million to Dell for NVIDIA hardware, including Dell PowerEdge XE9780 servers with B300 SXM6 GPUs. Dell is the primary server and GPU procurement partner.
- OpenRoutercoreOpenRouter routes developer demand onto TensorX sovereign GPU capacity, serving as a channel partner that drives inbound API volume through its developer marketplace. TensorX provides the underlying sovereign inference infrastructure for OpenRouter's European users.
- Digital Realty Trust, Inc.coreData centre colocation provider for GPU infrastructure in Dublin, Ireland (Tier III+ data centre with 24/7 physical security, biometric access controls, and CCTV monitoring). Part of the EU-sovereign infrastructure footprint.
- VerdacoreData centre colocation provider for GPU infrastructure in Helsinki, Finland. Tier III+ facility with physical security. Part of the EU-sovereign infrastructure footprint alongside Dublin.
- Deus X Capital (Tim Grant & Stuart Connolly)strategicTim Grant (CEO at Deus X Capital & Executive Chairman at Cor Prime) and Stuart Connolly (CIO at Deus X Capital & CEO at Alpha Lab 40) serve on the TensorX Advisory Board, providing strategic guidance on market development and the sovereign AI opportunity.
Scale indicators10 records
Recent moves9 records
Expansion highlights7 records
TensorX competitors and assessment
Company assessmentBroad incumbents
- Scaleway: French European cloud provider offering GPU instances and AI inference products with sovereignty positioning. Comparable in targeting European regulated workloads with EU-hosted GPU capacity, though operating across a broader cloud portfolio.
- OpenAI: Largest commercial AI inference provider whose API TensorX is positioned as a sovereign drop-in alternative. The reference competitor on cost (60-97% savings claimed) and the developer-experience baseline that TensorX's OpenAI-compatible API explicitly mimics.
- Anthropic: Major frontier AI lab offering Claude API for enterprise inference. Comparable as a high-end inference provider cited by TensorX as a pricing benchmark and as an alternative being displaced by sovereign offerings in regulated European accounts.
- OVHcloud: European cloud provider expanding into AI inference with sovereign cloud positioning. Comparable as a European infrastructure alternative to US hyperscalers, though AI inference is a smaller part of a much broader cloud and hosting portfolio.
Direct peers
- Mistral AI: Paris-based AI company providing open-weight foundation models and a sovereign European inference platform. Most directly comparable European sovereign AI peer, with overlapping model catalogue, developer API offering, and European regulated-industry go-to-market.
- Fireworks AI: AI inference platform serving open-weight and custom models with optimised serving infrastructure. Comparable as a developer-first inference API targeting production AI workloads, competing for the same AI builder and SaaS developer customer segment.
- Together AI: Open-source AI cloud platform offering inference APIs for a wide range of open-weight models on owned GPU infrastructure. Directly comparable product offering (multi-model API, usage-based pricing, open-weight focus), differentiated mainly by geography and sovereignty posture.
- Lambda: GPU cloud provider offering dedicated NVIDIA clusters and on-demand inference for AI workloads. Comparable in operating NVIDIA GPU infrastructure for AI customers with both reserved cluster and usage-based pricing models.
- Aleph Alpha: Heidelberg-based sovereign AI provider offering European-hosted inference of its own models for regulated public-sector and enterprise customers. Comparable in its EU sovereignty positioning, regulated-industry focus, and emphasis on compliance-grade AI infrastructure.
- CoreWeave: US-based GPU cloud provider operating NVIDIA hardware at scale for AI inference and training workloads. Comparable as a GPU-as-a-Service operator serving AI labs and enterprises, though lacking TensorX's EU sovereignty and zero-retention positioning.
Market position
Strengths5 records
Weaknesses5 records
Competitive moat5 records
Key risks5 records
Key highlights7 records
Customer concentration
TensorX social profiles
Digital presenceTensorX compliance and trust
Trust signalCompliance6 records
TensorX financial estimates
Financial estimateRevenue estimate
Valuation estimate
TensorX leadership team
Management profileNumber of profiles
Profiles9 records
TensorX funding detail
Funding detailFunding overview
Funding rounds1 record
Investors1 record
Funding detail is available on the Subscription and Enterprise plan.Contact sales →
TensorX M&A and investment
M&A and investmentM&A
Investments
M&A and investment is available on the Subscription and Enterprise plan.Contact sales →
Frequently asked questions about TensorX
What does TensorX do?
TensorX provides EU-sovereign AI inference infrastructure, running 26+ open-weight large language models on TensorX-owned NVIDIA Blackwell B300 GPUs hosted in EU data centres (Dublin, Helsinki, Paris). The platform serves regulated industries (financial services, healthcare, legal, government, aviation) through two product lines: a pay-per-token OpenAI-compatible API for developers and dedicated physically isolated GPU clusters for enterprises. All inference operates with zero data retention by design — prompts and completions are processed in ephemeral enclaves and never persisted.
Is TensorX a public or private company?
TensorX is a private company. It is classified as venture growth investor backed and is currently operating.
When was TensorX founded?
TensorX was founded in 2026. It employs 1 to 10 people.
Where is TensorX based?
TensorX is headquartered in Dublin, Ireland, in the Europe region.
How does TensorX make money?
Three revenue lines are on record. Pay-per-Token (Usage-Based) Inference is the primary driver. The others are dedicated Sovereign GPU Clusters and shared Inference Clusters.
Who are TensorX's main competitors?
Broad incumbents on record are Scaleway, OpenAI, Anthropic and OVHcloud. Direct peers are Mistral AI, Fireworks AI, Together AI, Lambda, Aleph Alpha and CoreWeave.
Does TensorX have an API?
Yes. TensorX offers a drop-in OpenAI-compatible REST API for private AI inference on EU-sovereign infrastructure. The API accepts standard OpenAI-format requests, returns responses in OpenAI-compatible format, and supports 30+ open-weight models including GLM-5.2, MiniMax-M3, DeepSeek-V4, Kimi-K2, and Qwen variants. Authentication uses API keys prefixed with 'tx-'. The API supports both pay-per-token (shared inference) and dedicated inference endpoints. Developer documentation is at docs.tensorx.ai.
What industry is TensorX in?
TensorX's product category is AI Inference Platform. Its primary akta.pro industry code is HDAAACAB, Model Hosting, Serving & Inference Platforms, with a secondary code of HDAAAAAB, AI Server Systems & HGX/Accelerator Platforms. Its NAICS code is 54151 and its SIC code is 7370.