Connect your agent to akta.pro and build company lists like this across sectors and geos.
Private US-headquartered ai inference & model-serving companies, All stages, ranked by total capital raised. Data compiled .
| Company | Category | Funding | Products & Services | Select Customers | Expansion Signal | Management |
|---|---|---|---|---|---|---|
| Lambda ๐บ๐ธ San Jose, North America Founded 2012 501-1000 employees | GPU Cloud Infrastructure | Latest Round Series E Aug 2026 Total Raised $5.11B total raised | Lambda builds and operates single-tenant AI cloud infrastructure (Superclusters, 1-Click Clusters, and Instances) running on NVIDIA GPU architectures including HGX B200, GB300 NVL72, and VR200 NVL72. Superclusters Core product 1-Click Clusters Core product Instances Core product AI Factories (Contiguous andโฆ Core product | Microsoft Hudson River Trading Orange Kodiak | Funding Signal Lambda raised $926 million in debt funding and signed a reported $35 billion multi-year compute deal with Anthropic (September 2026). | Michel Combes Chief Executive Officer Stephen Balaban Chief Technology Officer Michael Balaban Chief Product Officer John Donovan Chairman of the Board |
| Fluidstack ๐บ๐ธ New York, North America Founded 2017 101-250 employees | AI Cloud Infrastructure | Latest Round Private Equity Sep 2026 Total Raised $3.78B total raised | Fluidstack builds and operates gigawatt-scale AI compute infrastructure for training and inference workloads. The company acquires power, designs and constructs data centers, and runs HPC clusters wit Fluidstack AI Cloud Platform AI cloud infrastructure / neocloud Fluidstack Console Cloud management software platform Data Center Infrastructure Services Data center build-operate services | Anthropic Meta Poolside Black Forest Labs | Funding Signal Fluidstack raised $1.5 billion and is reserving 1.5GW of data center capacity for further AI compute expansion (September 2026). | qw gf Chief Executive Officer and Co-founder Cesar Maklary Co-founder & President Peter Francis Co-founder Jason Kichen Chief Information Security Officer |
| Groq ๐บ๐ธ San Jose, North America Founded 2016 501-1000 employees | Ai Inference Semiconductor Solutions | Latest Round Aug 2026 Total Raised $3.40B total raised | Groq develops and operates the LPU (Language Processing Unit), a custom inference accelerator silicon pioneered in 2016, and delivers it as a vertically integrated AI inference cloud called GroqCloud. GroqCloud AI Inference Platform Groq LPU (Language Processing Unit) AI Inference Silicon GroqCloud Console Developer Platform Console | Dropbox Chevron Volkswagen Canva | Funding Signal Groq is reportedly in talks around a $6 billion valuation as its LPU-based inference hardware enters full production (September 2026). | Adam Winter Interim CEO Matt Eng CFO Alan Rice COO Sinclair Schuller CTO |
| SambaNova ๐บ๐ธ Palo Alto, North America Founded 2017 251-500 employees | Ai Infrastructure Platform | Latest Round Series F Jul 2026 Total Raised $2.49B total raised | SambaNova designs and manufactures proprietary Reconfigurable Dataflow Unit (RDU) AI inference chips (SN40L and SN50 generations) and packages them into air-cooled SambaRack systems. It sells these as SambaCloud AI inference cloud platform SambaStack Enterprise AI inference platform SambaManaged | SoftBank Corp. OVHcloud TEPCO Systems Together AI | Ecosystem Buildout SambaNova partnered with Lawrence Livermore National Laboratory and announced a UK sovereign AI cloud collaboration with Argyll (September 2026). | Rodrigo Liang Co-Founder & CEO Kunle Olukotun Co-Founder & Chief Technologist Christopher Rรฉ Co-Founder Lip-Bu Tan Chairman of the Board |
| Baseten ๐บ๐ธ San Francisco, North America Founded 2019 251-500 employees | AI Inference Infrastructure | Latest Round Series F Jun 2026 Total Raised $2.39B total raised | Baseten provides an AI inference platform that enables businesses to deploy, optimize, and scale open-source, custom, and fine-tuned AI models in production. The Baseten Inference Stack combines custo Baseten Inference Stack AI Inference Platform Dedicated Deployments Dedicated Inference Model APIs Model APIs Training AI Training Infrastructure | Notion Speechify Gamma Writer | M And A Signal Baseten acquired Blaxel to build infrastructure for running AI agents in production (September 2026). | Tuhin Srivastava Co-Founder & CEO Amir Haghighat Co-Founder Phil Howes Co-Founder Pankaj Gupta Co-Founder |
| Fireworks AI ๐บ๐ธ Redwood City, North America Founded 2022 101-250 employees | AI Inference & Model Serving | Latest Round Series D Jul 2026 Total Raised $1.83B total raised | Fireworks AI provides a unified AI inference and training platform that runs and fine-tunes 1000+ open-source generative AI models at production scale. The platform offers serverless pay-per-token inf Fireworks Inference Platform AI Inference Platform Fireworks Training Platform | Cursor Notion Cresta Upwork | Funding Signal Fireworks AI is reportedly exploring a new funding round as AI inference-platform valuations continue to rise (September 2026). | Lin Qiao Co-Founder and CEO |
| Together AI ๐บ๐ธ San Francisco, North America Founded 2022 251-500 employees | AI Cloud Infrastructure | Latest Round Series C Jul 2026 Total Raised $1.33B total raised | Together AI operates an AI Native Cloud platform that provides GPU-accelerated compute, inference APIs, fine-tuning, and proprietary systems research for serving open-source generative AI models at sc Serverless Inference Inference API Batch Inference Batch Processing API Dedicated Model Inference Dedicated Inference Dedicated Container Inference Container Inference | Cursor Decagon The Washington Post Zomato | New Product Line Together AI launched new inference capabilities including Lakebase Cache, preemptible GPUs, and an Agents API (September 2026). | Vipul Ved Prakash Co-founder and Chief Executive Officer Tri Dao Chief Scientist Ce Zhang Co-founder and Chief Technology Officer Chris Re Co-founder |
| Positron ๐บ๐ธ Reno, North America Founded 2023 101-250 employees | AI Inference Hardware | Latest Round Series C Sep 2026 Total Raised $1.18B total raised | Positron designs and supplies purpose-built AI inference hardware and software for transformer-based models. Its product stack includes the shipping Atlas FPGA-based inference appliance (8x Positron A Atlas Transformer Inference Server AI inference appliance (hardware + bundled software) Positron Inference Engine & OpenAIโฆ Inference software platform / API | Funding Signal Positron AI raised an $875 million Series C at a $5 billion valuation to build inference chips positioned to rival Nvidia (September 2026). | Mitesh Agrawal Chief Executive Officer Thomas Sohmers CTO and Founder | |
| QumulusAI ๐บ๐ธ Marietta, North America Founded 2019 11-50 employees | AI Cloud Infrastructure | Latest Round Public Jun 2026 Total Raised $544M total raised | QumulusAI provides GPU-as-a-Service (GPUaaS) for large-scale AI workloads through three product tiers โ QumulusAI Cloud (shared GPUs), Cloud Pro (dedicated GPUs), and Cloud Pure (bare metal clusters) QumulusAI Cloud (Shared GPUs) Cloud GPU Services QumulusAI Cloud Pro (Dedicated GPUs) Cloud GPU Services QumulusAI Cloud Pure (Bare Metalโฆ Cloud GPU Services | vCluster Shadeform Hyperbolic | Funding Signal QumulusAI signed $240.9 million in new AI inference deals and completed contracted NVIDIA B200/B300 Blackwell deployments (September 2026). | Mike Maniscalco Chief Executive Officer and Chairman of the Board Scott Krosnowski Chief Financial Officer Ryan DiRocco Chief Technology Officer Stephen Hunton Chief Marketing Officer |
| TensorWave ๐บ๐ธ Las Vegas, North America Founded 2023 101-250 employees | AI Cloud Infrastructure | Latest Round Series B Jun 2026 Total Raised $497M total raised | TensorWave operates an AMD-exclusive AI cloud platform providing GPU compute infrastructure for AI training, fine-tuning, and inference workloads. The platform is built on AMD Instinct accelerators (M Bare Metal GPU Infrastructure Core Infrastructure AMD Instinct MI300X GPU Accelerator AMD Instinct MI325X GPU Accelerator AMD Instinct MI355X GPU Accelerator | Meta Modular Higgsfield AI Moreh | Leadership Change TensorWave expanded its executive leadership team as it scales AI infrastructure operations under a $6.5 billion deal with Fermi (September 2026). | Darrick Horton Co-founder & Chief Executive Officer Jeff Tatarchuk Co-founder & Chief Growth Officer Piotr Tomasik Co-founder, President & Chief Operating Officer Kathleen Simon Executive Vice President of Finance |
| Modal Labs ๐บ๐ธ New York, North America Founded 2021 51-100 employees | Cloud GPU Infrastructure | Latest Round Series C May 2026 Total Raised $466M total raised | Modal Labs operates an AI-native serverless GPU cloud platform that lets developers run inference, training, batch processing, and isolated code execution on NVIDIA GPUs (H100s, A100s, B200s, H200s, L Inference AI Inference Training AI Training Sandboxes Agent Runtime Batch | Cognition DoorDash Decagon TypeSafe AI | Funding Signal Modal Labs is reportedly nearing a $750 million funding round at a $16 billion valuation (September 2026). | Information not provided in source materials Leadership (specific individuals not disclosed in source) |
| General Compute ๐บ๐ธ West Covina, North America Founded 2025 11-50 employees | AI Inference Cloud Infrastructure | Latest Round Seed Jul 2026 Total Raised $415M total raised | General Compute operates an ASIC-native AI inference cloud (General Compute Cloud) running on SambaNova dataflow accelerators and exposed via an OpenAI-compatible REST API. It sells three commercial t API Access Inference API Custom Deployments Dedicated Infrastructure Bring Your Own Model Private Model Deployment General Compute Cloud AI Inference Cloud Platform | Funding Signal General Compute raised up to $400 million in debt financing, using its AI inference chips as loan collateral (July 2026). | Jason Goodison Co-founder & CTO Finn P. Co-founder & CEO | |
| PaleBlueDot AI ๐บ๐ธ Palo Alto, North America Founded 2024 11-50 employees | AI Compute Platform | Latest Round Series B Jul 2026 Total Raised $415M total raised | PaleBlueDot AI provides an AI compute platform centered on PBD TokenRouter, a unified API layer that consolidates access to multiple frontier AI model providers (OpenAI, Anthropic, Google, Moonshot AI PBD TokenRouter AI API Platform Token Factory GPU Compute Infrastructure | Funding Signal PaleBlueDot AI raised $150 million in Series B funding and closed a $255 million GPU-backed credit financing to expand its agentic AI infrastructure (August 2026). | Stephen Watts Chief Executive Officer Jonathan Zhu Founder Sheldon Huang Co-founder Vincent Li Chief Financial Officer | |
| Gimlet Labs ๐บ๐ธ San Francisco, North America Founded 2023 11-50 employees | AI Inference Infrastructure | Latest Round Series B Sep 2026 Total Raised $392M total raised | Gimlet Labs builds and operates Gimlet Cloud, a serverless multi-silicon AI inference platform that runs agentic workloads (simple agents through multi-agent systems with custom logic and data sources Gimlet Cloud AI Inference Platform kforge Developer Tools / AI Compiler | Funding Signal Gimlet Labs priced a $3 billion round and secured Cerebras as a hardware supplier for its multi-silicon inference cloud (September 2026). | Zain Asgar Co-Founder/CEO Michelle Nguyen Co-Founder Omid Azizi Co-Founder Natalie Serrino Co-Founder | |
| Massed Compute ๐บ๐ธ Kansas, North America Founded 2021 11-50 employees | Cloud GPU Infrastructure | Latest Round Private Equity Aug 2025 Total Raised $300M total raised | Massed Compute provides on-demand access to enterprise NVIDIA GPUs (Blackwell, Hopper, Ada Lovelace, and Ampere architectures) via four product tiers: hourly On-Demand GPU rental starting at $0.43/hr, On-Demand GPU Cloud GPU Compute Bare Metal Servers Dedicated Server Hosting GPU Clusters High Performance Computing API and MCP Programmatic Access Developer API | Ecosystem Buildout Massed Compute partnered with PacketFabric to launch an integrated GPUaaS and network-as-a-service offering for enterprise AI (January 2026). | Sunny Smith Founder Craig Gieringer Senior Vice President | |
| Deepgram ๐บ๐ธ San Francisco, North America Founded 2015 251-500 employees | Voice Recognition Api | Latest Round Series C Aug 2026 Total Raised $218M total raised | Deepgram provides an end-to-end voice AI platform delivered through APIs for real-time speech-to-text (Nova-3), text-to-speech (Aura-2), conversational speech recognition (Flux), and unified voice age Nova-3 Speech-to-Text Speech-to-Text API Aura-2 Text-to-Speech Text-to-Speech API Flux Conversational Speech Recognition Conversational Speech Recognition | Twilio Cloudflare Sierra IBM | Funding Signal Deepgram surpassed $100 million in ARR and landed a $130 million Series C extension, opening an APAC headquarters in Singapore (August 2026). | Scott Stephenson Co-founder and Chief Executive Officer Anoop Dawar Chief Sales Officer (CSO) Praveen Rangnath Chief Marketing Officer (CMO) Adam Sypniewski Chief Technology Officer (CTO) |
| Cartesia ๐บ๐ธ San Francisco, North America Founded 2023 51-100 employees | Real-Time Voice AI Platform | Latest Round Series B Oct 2025 Total Raised $213M total raised | Cartesia provides a unified real-time voice AI platform built on proprietary State Space Model (SSM) architecture, delivering Sonic text-to-speech, Ink speech-to-text, and Line voice agent orchestrati Sonic (Sonic-3.5) Text-to-Speech (AI Voice Generation) Ink (Ink-2) Speech-to-Text (Streaming Transcription) Line Voice Agent Development Platform Narrations Long-Form Audio Content Platform | ServiceNow Retell Fundamento Quora | Karan Goel, PhD CEO & Founder Albert Gu, PhD Chief Scientist & Co-founder Arjun Desai, PhD Co-founder Brandon Yang Team Member & Co-founder | |
| Prime Intellect ๐บ๐ธ San Francisco, North America Founded 2024 51-100 employees | AI Training Infrastructure | Latest Round Series A Jul 2026 Total Raised $200M total raised | Prime Intellect operates an integrated open stack for training, deploying, and continuously improving AI agents. Its hosted Lab platform combines on-demand GPU compute (aggregated from 50+ datacenters Prime Intellect Lab Hosted AI Training Platform Prime Intellect Compute GPU Compute Marketplace Environments Hub RL Environments Marketplace INTELLECT-3 Open Foundation Model | Ramp Zapier Arcee AI Character.ai | New Product Line Prime Intellect released Prime Agent, a self-improving open-source coding harness that outperformed human experts on a key AI benchmark (August 2026). | Will Brown Research Lead Johannes Hagemann Co-founder Matej Sirovatka Research Team Member |
| Deep Infra ๐บ๐ธ Palo Alto, North America Founded 2022 11-50 employees | AI Inference Cloud Infrastructure | Latest Round Series B May 2026 Total Raised $154M total raised | DeepInfra operates a purpose-built AI inference cloud platform that delivers pay-per-token API access to 100+ open-source AI models across text, image, video, audio, code, embeddings, and reranking mo DeepInfra AI Inference Cloud Core Inference Platform DeepCluster Dedicated GPU Infrastructure GPU Instances On-demand GPU Rental DeepStart AI Agent Platform | Abacus.AI Hugging Face interface.ai Salesforce | Funding Signal DeepInfra surpassed $100 million in ARR and 22 trillion tokens served per week (September 2026). | Nikola Borisov Co-founder and CEO Georgios Papoutsis Co-founder and Engineer Yessen Kanapin Co-founder |
| Inferact ๐บ๐ธ San Francisco, North America Founded 2026 11-50 employees | AI Inference Infrastructure | Latest Round Seed Jan 2026 Total Raised $150M total raised | 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+ mode vLLM (Open Source) Open-source inference engine Inferact Serverless Managed inference platform | Funding Signal Inferact raised $150 million in a seed round, backed by Andreessen Horowitz, to commercialize the vLLM inference engine (January 2026). | Simon Mo Co-founder & CEO Woosuk Kwon Co-founder Kaichao You Co-founder & Chief Scientist Roger Wang Co-founder | |
| Inworld AI ๐บ๐ธ Mountain View, North America Founded 2021 101-250 employees | Voice AI Infrastructure | Latest Round Series A Aug 2023 Total Raised $123M total raised | Inworld AI provides a realtime voice AI infrastructure platform comprising text-to-speech (TTS), speech-to-text (STT), end-to-end Realtime API, and an LLM routing layer (Realtime Router) that connects Realtime TTS Voice AI / Text-to-Speech Realtime STT | Talkpal Wishroll Bible Chat | New Product Line Inworld AI launched Realtime TTS-2, a new voice model family for controllable, realtime speech, and partnered with Tencent Cloud on voice AI (September 2026). | Kylan Gibbs Chief Executive Officer Ilya Gelfenbeyn Co-founder Michael Ermolenko |
| Runpod ๐บ๐ธ Mount Laurel, North America Founded 2022 101-250 employees | Cloud GPU Infrastructure | Latest Round Series A Jun 2026 Total Raised $122M total raised | Runpod is a cloud platform purpose-built for GPUs that lets developers and AI teams spin up on-demand GPU instances, autoscaling serverless inference endpoints, multi-node training clusters, and one-c Cloud GPUs (Pods) On-demand GPU compute Serverless Serverless GPU inference endpoints Clusters Multi-node GPU clusters Runpod Hub AI model deployment templates | Rendair Nodalview Faceless.video | Ecosystem Buildout Runpod expanded its enterprise platform for mission-critical AI workloads and completed a 616-GPU deployment with QumulusAI (September 2026). | Zhen Lu Co-founder and CEO Pardeep Singh CTO and Co-founder Luke Piette Head of Growth Greg Wester Director of Product |
| Theta Labs ๐บ๐ธ Cupertino, North America Founded 2017 11-50 employees | Decentralized GPU Cloud Infrastructure | Latest Round Seed Oct 2023 Total Raised $113M total raised | Theta Labs operates Theta EdgeCloud, a hybrid decentralized GPU compute platform that combines 30,000+ distributed community-run edge nodes with enterprise-grade cloud infrastructure (Google Cloud, AW Theta EdgeCloud Decentralized GPU Compute ThetaDrop NFT Platform Theta Video API Video Streaming API AI Agent Platform AI Agent Platform | Seoul National University KAIST Houston Rockets Cloud9 | Mitch Liu CEO Jieyi Long CTO and Co-Founder Ryan Nichols Chief Product Officer and Co-Founder Wes Levitt Head of Strategy | |
| Lightning AI ๐บ๐ธ New York, North America Founded 2019 11-50 employees | AI Cloud Infrastructure Platform | Latest Round Series B Nov 2024 Total Raised $108M total raised | Lightning AI is an AI-native cloud platform that unifies AI model development and deployment with managed GPU infrastructure. It offers collaborative workspaces (AI Studio), persistent GPU notebooks, AI Studio AI Development Workspace AI Notebooks AI Development Workspace GPU Clusters GPU Compute Infrastructure Inference AI Model Serving | LinkedIn NVIDIA Runway Stability AI | Leadership Change Lightning AI expanded its executive team with new finance, marketing, and people-leader appointments (September 2026). | William Falcon Founder and CEO |
| RadixArk ๐บ๐ธ San Francisco, North America Founded 2024 11-50 employees | AI Inference and Training Infrastructure Software | Latest Round Seed Jan 2026 Total Raised $100M total raised | RadixArk develops and operates open-source AI infrastructure for inference and training. Its core open-source products are SGLang, an inference engine for serving modern large language models that sit SGLang AI inference engine Miles AI training framework Managed AI Infrastructure Managed AI infrastructure services | Google Microsoft NVIDIA Cursor | New Product Line RadixArk released Miles, an open-source framework unifying SGLang, NVIDIA Megatron-LM, and Ray for large-scale LLM reinforcement-learning post-training (July 2026). | Ying Sheng Co-founder and CEO |
An AI inference and model-serving company sells the runtime layer that takes a trained AI model and makes it usable in production: GPU cloud capacity, serverless or dedicated inference endpoints, model-hosting APIs, or purpose-built inference silicon paired with a cloud/platform offering. The defining test is that inference or GPU compute delivered as a service is the company's primary product, not a secondary feature bolted onto a broader analytics, MLOps, or application platform.
This run drew a firm line against three adjacent categories that are easy to mistake for the core sector: (1) AI application companies (answer engines, creative tools, agentic SaaS) that consume inference from someone else rather than sell it; (2) foundation-model labs whose primary business is training and licensing proprietary models rather than operating a serving platform for third parties; and (3) inference-chip or accelerator companies that sell hardware alone, with no accompanying cloud or platform-as-a-service offering. A company qualified only when GPU cloud, inference APIs, or model-serving infrastructure was demonstrably its central, revenue-generating product.
The three most-funded companies on this list, Lambda, Fluidstack, and Groq, together account for 47.2% of the $26.1 billion in total disclosed funding across all 25 companies.
Excluded from this list: public companies, AI-application and analytics platforms that merely consume third-party models, foundation-model developers without a third-party serving platform, hardware-only semiconductor companies with no cloud/platform offering, and any company whose independent private/operating status could not be confirmed as of the run date (one qualifying candidate, Anyscale, was excluded after news signals confirmed a pending acquisition by Nscale).
Companies are tagged against standard industry codes and against akta.pro's internal taxonomy, so the same set can be rebuilt by code, by segment, or by a combination of both.
NAICS 5415 Computer Systems Design and Related Services 518210 Computing Infrastructure Providers, Data Processing, Web Hosting, and Related Services 513210 Software Publishers 541512 Computer Systems Design Services 541513 Computer Facilities Management Services SIC 7372 Services-Prepackaged Software 7373 Services-Computer Integrated Systems Design 7370 Services-Computer Programming, Data Processing, Etc. 7374 Services-Computer Processing & Data Preparation 3571 Electronic Computers Internal taxonomy HDAAABAF Model Deployment, Serving & Inference Platforms HDAAACAB Model Hosting, Serving & Inference Platforms HDAEANAC Model Serving, Inference & Deployment Platforms (APIs, Edge/On-Prem) HDAAAAAK AI Compute Cloud & GPU-as-a-Service HDAAAJAB On-Device Inference Runtimes & SDKs (mobile/embedded)Industry codes are provided for cross-referencing against public filings and procurement records. akta.pro classification is independent of them, so a company may carry an internal segment without a matching NAICS or SIC registration.