Inferact
Inferact commercializes vLLM, the leading open-source LLM inference engine, delivering 2-4x faster inference across 500+ model architectures and 200+ accelerator types. The San Francisco startup serves AI companies and enterprises deploying frontier models at scale, with a planned paid serverless platform for organizations lacking dedicated inference infrastructure teams.
- Company typePrivate
- Founded2026
- HeadquartersSan Francisco, United States
- Headcount11–50
- GTM typeB2B
- OfferingSoftware
What Inferact does
Inferact is a San Francisco-based AI infrastructure startup founded in January 2026 by the creators and core maintainers of vLLM, the open-source large language model inference engine originally developed at UC Berkeley's Sky Computing Lab. The company's core technology, vLLM, accelerates LLM inference by 2-4x and supports more than 500 model architectures across 200+ accelerator types, currently running on over 400,000 GPUs globally with contributions from more than 2,000 open-source developers. vLLM is used in production by major technology companies including Meta, Google, and Character.ai, and the project has day-zero support relationships with model vendors and hardware vendors, positioning the company at the intersection of model and hardware ecosystems.
The company pursues a dual-product strategy: continuing to maintain and enhance the free open-source vLLM engine while developing a planned paid serverless inference platform targeting enterprises that need to deploy frontier AI models at scale without dedicated infrastructure teams. Optimizations developed commercially flow back to the open-source project. The six-person founding team comprises Simon Mo (CEO), Woosuk Kwon, Kaichao You (Chief Scientist, Tsinghua University PhD), Roger Wang, Joseph Gonzalez, and Ion Stoica (Databricks co-founder). Inferact's customer segmentation is horizontal, serving both AI-native companies deploying frontier models at scale (frontier labs, hyperscalers, AI startups serving millions of users) and mid-market companies that lack the engineering resources to build custom inference infrastructure.
Inferact is venture-backed following a $150 million seed round in January 2026 at an $800 million post-money valuation, co-led by Andreessen Horowitz and Lightspeed Venture Partners, with participation from Sequoia Capital, Altimeter Capital, Redpoint Ventures, ZhenFund, Databricks Ventures, UC Berkeley Chancellor's Fund, and other investors. The company currently has 11-50 employees and has not yet disclosed revenue, as the commercial serverless platform remains in development.
Inferact firmographics
Firmographics- Name
- Inferact
- Legal name
- Inferact Inc.
- Website
- https://inferact.ai
- Company type
- Private
- Founded year
- 2026
- Operating status
- Operating
- Headcount range
- 11–50 employees
- Short description
- Inferact commercializes vLLM, the leading open-source LLM inference engine, delivering 2-4x faster inference across 500+ model architectures and 200+ accelerator types. The San Francisco startup serves AI companies and enterprises deploying frontier models at scale, with a planned paid serverless platform for organizations lacking dedicated inference infrastructure teams.
- Ownership category
- akta.pro rank
Inferact industry classification
Industry- Product category
- AI Inference Infrastructure
- NAICS
- Computing Infrastructure Providers, Data Processing, Web Hosting, and Related Services (51821), Computer Systems Design and Related Services (5415), Custom Computer Programming Services (541511)
- SIC
- Services-Prepackaged Software (7372), Services-Computer Programming, Data Processing, Etc. (7370)
- akta.pro industry
- Model Hosting, Serving & Inference PlatformsHardware, Technology & Telecom/Artificial Intelligence & Machine Learning/Foundation Models & Generative AI (LLMs, multimodal)Model Deployment, Serving & Inference PlatformsHardware, Technology & Telecom/Artificial Intelligence & Machine Learning/ML Platforms, MLOps & Model Lifecycle ManagementLLMOps & Generative AI Platforms (Prompt/Agent Orchestration, RAG)Hardware, Technology & Telecom/Enterprise Platforms & Databases/Artificial Intelligence Platforms for Enterprise
Keywords
Where Inferact is headquartered
LocationHeadquarters
- HQ city
- San Francisco
- HQ country
- United States
- HQ region
- North America
Markets served
Inferact business model
Business model- GTM type
- B2B
- Offering type
- Software
- Cost components
- Personnel, Technology or R&D, Infrastructure, Marketing or Sales, Operations
Revenue model
- Serverless Inference Platform: Commercial paid serverless version of vLLM targeting enterprises that need to deploy frontier AI models at scale without dedicated infrastructure teams. The open-source vLLM remains free while the managed/cloud service provides additional convenience and support.
Go-to-market motion1 record
Distribution channels2 records
Marketing channels3 records
Inferact product offering
Product offeringCore offering
Inferact develops and commercializes vLLM, the open-source LLM inference engine originally created at UC Berkeley's Sky Computing Lab. The engine delivers 2-4x faster LLM inference, supports 500+ model architectures across 200+ accelerator types, and runs on 400,000+ GPUs globally. Inferact plans to layer a paid serverless version of vLLM on top of the open-source project so enterprises can deploy frontier AI models without dedicated infrastructure teams.
Product overview
Inferact operates a dual-product strategy centered on vLLM. The company maintains and develops vLLM as an open-source inference engine (the core open-source product), while also building a commercial serverless offering (Inferact Serverless) on top of it. The open-source vLLM provides the foundational technology and community, while the commercial product adds managed infrastructure capabilities for enterprise customers who want serverless deployment without building custom infrastructure.
Differentiator
Problem solved
Functional benefit
Products and services
- vLLM (Open Source)
- Inferact Serverless
Quantifiable outcome
- 2-4x LLM inference speed improvement
Companies that use Inferact
Customer profileSegments2 records
Ideal customer profiles2 records
Inferact technology and API
TechnologyTechnology focussed Yes
API detail
- Has API
- No
- API docs
- API detail
Core technology
AI maturity
App detail
AI capability2 records
Feature2 records
Inferact partnerships and signals
Strategic signalScale indicators6 records
Recent moves7 records
Expansion highlights5 records
Inferact competitors and assessment
Company assessmentDirect peers
- Fireworks AI: Fireworks AI provides a low-latency LLM inference and fine-tuning platform optimized for production deployment. It targets the same developer and enterprise buyers looking to serve frontier models at scale.
- Anyscale: Anyscale commercializes Ray and offers Anyscale Endpoints for LLM serving and inference at scale. It targets the same enterprise developer persona and competes directly for managed LLM inference workloads.
- OctoAI (OctoML): OctoAI builds an AI compute platform optimized for running and fine-tuning open-source models. It is comparable to Inferact as a managed serving layer for production AI inference.
- Modal Labs: Modal offers serverless compute infrastructure for AI and data workloads, including LLM inference. Its 'serverless GPU' positioning competes head-on with Inferact's planned serverless inference offering.
- Replicate: Replicate runs a cloud that lets developers deploy and scale AI models via an API, with a focus on open-source models. It is a direct competitor in serving open-source LLMs to developers and small enterprises.
- DeepInfra: DeepInfra provides low-cost, low-latency inference-as-a-service for open-source LLMs and other models. It targets a similar cost-sensitive developer and enterprise customer base.
- Together AI: Together AI operates a cloud platform for training, serving, and inference of open-source and frontier LLMs. It is the closest direct competitor in the open-source-rooted, paid-inference serverless model.
- Baseten: Baseten offers managed infrastructure and serving for ML models, with a strong focus on LLM inference workloads. Its developer- and enterprise-focused serving platform directly overlaps with Inferact's planned offering.
Broad incumbents
- Hugging Face: Hugging Face is a broad platform for open-source AI models and includes Inference Endpoints for serving models. While not specialized in inference, it overlaps with Inferact's open-source-rooted go-to-market and offers managed serving.
- NVIDIA (TensorRT-LLM / NIM): NVIDIA's TensorRT-LLM and NIM microservices provide optimized LLM inference on NVIDIA hardware, bundled with the dominant accelerator stack. It competes for the same workloads but as part of NVIDIA's broader AI platform portfolio.
Market position
Strengths4 records
Weaknesses4 records
Competitive moat5 records
Key risks5 records
Key highlights6 records
Customer concentration
Inferact social profiles
Digital presenceInferact financial estimates
Financial estimateRevenue estimate
Valuation estimate
Inferact leadership team
Management profileNumber of profiles
Profiles6 records
Inferact funding detail
Funding detailFunding overview
Funding rounds1 record
Investors12 records
Funding detail is available on the Subscription and Enterprise plan.Contact sales →
Inferact 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 Inferact
What does Inferact do?
Inferact develops and commercializes vLLM, the open-source LLM inference engine originally created at UC Berkeley's Sky Computing Lab. The engine delivers 2-4x faster LLM inference, supports 500+ model architectures across 200+ accelerator types, and runs on 400,000+ GPUs globally. Inferact plans to layer a paid serverless version of vLLM on top of the open-source project so enterprises can deploy frontier AI models without dedicated infrastructure teams.
Is Inferact a public or private company?
Inferact is a private company. It is classified as venture growth investor backed and is currently operating.
When was Inferact founded?
Inferact was founded in 2026. It employs 11 to 50 people.
Where is Inferact based?
Inferact is headquartered in San Francisco, United States, in the North America region.
How does Inferact make money?
One revenue line is on record: serverless Inference Platform.
Who are Inferact's main competitors?
Direct peers on record are Fireworks AI, Anyscale, OctoAI (OctoML), Modal Labs, Replicate, DeepInfra, Together AI and Baseten. Broad incumbents are Hugging Face and NVIDIA (TensorRT-LLM / NIM).
Does Inferact have an API?
No public API is recorded for Inferact.
What industry is Inferact in?
Inferact's product category is AI Inference Infrastructure. Its primary akta.pro industry code is HDAAACAB, Model Hosting, Serving & Inference Platforms, with a secondary code of HDAAABAF, Model Deployment, Serving & Inference Platforms. Its NAICS code is 51821 and its SIC code is 7372.