Infinigence

What Infinigence does

Infinigence (无问芯穹) is a Beijing-based AI infrastructure company founded in 2023, affiliated with Tsinghua University's Department of Electronic Engineering. It provides AGI computing power solutions centered on large model energy efficiency optimization. Its core product, the Agentic MaaS Platform, operates as a neutral infrastructure layer between chip manufacturers and AI model developers, enabling AI inference workloads with prefill-decode separation technology that delivers a stated 5-10x cost-performance improvement. The platform also facilitates deployment of AI models on domestic Chinese chips, addressing hardware compatibility constraints in the Chinese AI ecosystem.

The company runs an API-first, token-based revenue model, generating income from token consumption on its inference platform. Its primary customer segments are enterprise AI agent operators and AI model developers seeking scalable, cost-efficient inference infrastructure. The company is headquartered in Beijing with 51-100 employees and has raised approximately $244 million across four funding rounds since December 2023, with backers including Legend Capital, Qiming Venture Partners, Shunwei Capital, Xiaomi, Shanghai Guotou, and Alibaba Entrepreneurs Fund.

Token call volume grew over 20x from December 2025 to April 2026 and doubled roughly every two weeks since late January 2026, indicating rapid adoption of the platform's inference services. Infinigence operates exclusively in China, with no disclosed international expansion.

Name

Infinigence

Legal name

Infinigence

Website

https://cloud.infini-ai.com

Company type

Private

Founded year

2023

Operating status

Operating

Headcount range

51–100 employees

Short description

Infinigence (无问芯穹) is a Beijing-based AI infrastructure company founded in 2023 that provides an Agentic MaaS Platform using prefill-decode separation to deliver cost-efficient inference for AI model developers and AI agent operators in China.

Ownership category

akta.pro rank

Product category

AI Inference Infrastructure

NAICS

Computing Infrastructure Providers, Data Processing, Web Hosting, and Related Services (51821), Computer Systems Design and Related Services (5415)

SIC

Services-Prepackaged Software (7372), Services-Computer Integrated Systems Design (7373)

akta.pro primary industry

Model Serving, Inference & Deployment Platforms (APIs, Edge/On-Prem) (HDAEANAC)

akta.pro secondary industries

Model Deployment, Serving & Inference Platforms (HDAAABAF), End-to-End Enterprise AI Platforms (MLOps & Model Lifecycle Management) (HDAEANAA), AI Compiler, Runtime & Kernel Optimization Software (CUDA/ROCm/XLA, graph compilers) (HDAAAAAI)

Keywords

• AI inference infrastructure
• Model-as-a-service
• GPU compute optimization
• Token-based API
• Heterogeneous chip deployment

Headquarters

HQ city

Beijing

HQ country

China

HQ region

Asia

Offices1 record

Markets served

GTM type

B2B

Offering type

Software

Cost components

Technology or R&D, Infrastructure, Personnel, Operations, Marketing or Sales

Revenue model

1. AI Inference Compute (Token-Based): Revenue is generated through token-based API usage on the Agentic MaaS platform. The company acts as a neutral 'token factory', providing AI inference compute services to model developers and enterprise customers. Token consumption has shown explosive growth with 20x volume increase over four months, and the broader market saw 15x token revenue growth for major providers since early 2026. The model is usage-based with consumption scaling with enterprise AI agent adoption.

Go-to-market motion2 records

Distribution channels1 record

Marketing channels1 record

Core offering

Infinigence operates an Agentic MaaS (Model-as-a-Service) platform that provides scalable AI inference compute infrastructure for AI agents and enterprise workloads. The platform functions as a neutral 'token factory' middleware between chip manufacturers and model developers, leveraging prefill-decode separation technology to achieve 5-10x cost-performance improvements while enabling AI model deployment on heterogeneous hardware, including domestic Chinese chips.

Product overview

Infinigence is an AI infrastructure company affiliated with Tsinghua University's Department of Electronic Engineering, offering a single core product—the Agentic MaaS Platform—that positions itself as a neutral infrastructure layer between chip manufacturers and AI model developers. The platform facilitates AI inference compute workloads and token generation services, leveraging prefill-decode separation technology to deliver 5-10x cost-performance improvements while providing a deployment pathway for domestic Chinese chips.

Differentiator

Problem solved

Functional benefit

Products and services

• Agentic MaaS Platform

Quantifiable outcome

• 20x token call volume growth from December 2025 to April 2026
• +2 more outcomes

Segments2 records

Ideal customer profiles2 records

Technology focussed Yes

API detail

Has API

No

API docs

API detail

Core technology

AI maturity

App detail

Feature3 records

Scale indicators5 records

Recent moves5 records

Expansion highlights4 records

Direct peers

• Replicate: Replicate provides a cloud API for running and deploying machine learning models, emphasizing simple API-based access to inference infrastructure. It serves the same developer- and enterprise-customer base seeking scalable, on-demand AI inference as Infinigence's Agentic MaaS platform.
• Anyscale: Anyscale provides AI infrastructure built on Ray for distributed model training and serving, targeting enterprise AI workloads with cost and performance optimization. Its serving and inference capabilities directly overlap with Infinigence's positioning as an enterprise AI compute platform.
• SiliconFlow (硅基流动): SiliconFlow is a Chinese AI inference infrastructure startup offering model serving and inference optimization across heterogeneous hardware including domestic Chinese chips. It competes directly with Infinigence's Agentic MaaS platform in the same 'token factory' positioning for Chinese AI model developers and enterprise customers.
• Fireworks AI: Fireworks AI is a global AI inference platform specializing in fast, cost-efficient model serving and deployment, with proprietary optimization techniques similar to Infinigence's prefill-decode separation. It serves the same customer base of model developers and enterprise AI builders seeking production-grade inference infrastructure.
• Modal Labs: Modal Labs provides serverless cloud infrastructure for AI inference and compute workloads, with developer-friendly APIs and a focus on scalable model serving. Its API-first, usage-based model mirrors Infinigence's Agentic MaaS approach for AI model deployment and inference.
• Together AI: Together AI operates a cloud platform for training, fine-tuning, and serving open-source AI models, with significant focus on inference optimization and cost-efficient deployment. It overlaps with Infinigence in providing scalable inference infrastructure to AI developers and enterprises, though serving global rather than Chinese markets.

Broad incumbents

• Baidu AI Cloud (Qianfan): Baidu AI Cloud offers the Qianfan model platform and broader AI infrastructure including inference services to enterprise and developer customers in China. It is a broad incumbent competing with Infinigence for the same Chinese AI inference workloads, with the added dimension of Baidu being an investor in the company.
• Tencent Cloud: Tencent Cloud provides AI infrastructure services including model serving and inference for enterprise customers in China. As a broad hyperscale incumbent, it competes with Infinigence for enterprise AI agent workloads while offering a much wider portfolio of services.
• Alibaba Cloud (Model Studio / PAI): Alibaba Cloud is a hyperscale cloud provider offering Model Studio (Bailian) and PAI for AI training and inference at massive scale. While serving a broader portfolio, its inference capabilities compete directly with Infinigence for Chinese enterprise AI customers and the same underlying compute workloads.
• Volcano Engine (ByteDance): Volcano Engine is ByteDance's cloud and AI platform offering model serving, inference infrastructure, and AI development services to enterprise customers. It competes with Infinigence in the Chinese enterprise AI infrastructure market while leveraging ByteDance's internal AI scale and capabilities.

Market position

Strengths5 records

Weaknesses5 records

Competitive moat4 records

Key risks6 records

Key highlights7 records

Customer concentration

• Xx.com/Infinigence

Revenue estimate

Valuation estimate

Number of profiles

Funding overview

Funding rounds4 records

Investors43 records

Funding detail is available on the Subscription and Enterprise plan.Contact sales →

M&A

Investments

M&A and investment is available on the Subscription and Enterprise plan.Contact sales →

Frequently asked questions about Infinigence