AI Infrastructure · United States

Top 35 AI Infrastructure Companies in US

The top 35 AI infrastructure companies in the US at Series B or later: privately held, US-headquartered, ranked by total capital raised, with most recent round, offering and revenue model for each. Data compiled 18 September 2026.

• — Funding is concentrated at the top. The 35 companies have raised $39.6B between them, and the five largest (Scale, Lambda, Baseten, Reflection AI, Etched.ai) hold 69.9% of that total. The top 10 hold 85.1% .
• — Chip and hardware companies make up a third of the list. 10 of 35 companies (29%) earn revenue from hardware sales, and six of them design custom AI silicon: Etched.ai, MatX, d-Matrix, Mythic, EnCharge AI and Tensordyne, funded at $3.6B combined.
• — The Bay Area still holds most of the sector. San Francisco alone is home to 11 of 35 companies (31%) and the wider Bay Area accounts for 25 of 35 (71%). The remainder sit in New York (4), Seattle (2), and Austin, Las Vegas, Miami and Irvine with one each.

Editorial list Coverage: US Filter: Private, Series B+ Ranked by: Total funding Source: akta.pro Company Data Updated: 18 September 2026

Top 35 US AI infrastructure companies

Top 35 ai infrastructure companies US, ranked by total funding, with HQ, most recent funding round, total funding, product offering, top management, and revenue model for each. HQ is shown beneath each company name.

Company	Category	Most Recent Round	Total Funding	Offering	Management	Revenue Model
Scale 🇺🇸 San Francisco, CA	AI Data Infrastructure	Series F May 2024	Funding $15.90B	End-to-end AI data infrastructure for data labeling, model training/evaluation, and enterprise/defense deployment.	Jason Droege CEO (interim) Clemens Viernickel Head of Product	Subscription Recurring Managed Services Usage Based
Lambda 🇺🇸 San Jose, CA	GPU Cloud Infrastructure	Series E Nov 2025	Funding $5.11B	Single-tenant NVIDIA GPU clusters for AI training/inference across 320+ MW of U.S. data centers.	Michel Combes CEO Stephen Balaban CTO Michael Balaban CPO Charles Fisher CFO	Usage Based Subscription Recurring Managed Services
Baseten 🇺🇸 San Francisco, CA	AI Inference Infrastructure	Series F Jun 2026	Funding $2.39B	Full-stack AI inference platform with dedicated deployments, model APIs, and white-label gateway across 87 clusters globally.	Tuhin Srivastava Co-Founder & CEO	Usage Based Subscription Recurring
Reflection AI 🇺🇸 New York, NY	AI Foundation Models & Training Infrastructure	Series B Oct 2025	Funding $2.13B	Frontier open-weight foundation models and proprietary large-scale LLM/RL training stack.	Misha Laskin CEO Ioannis Antonoglou CTO	Subscription Recurring Managed Services
Etched.ai 🇺🇸 San Jose, CA	AI Inference Semiconductors	Series D Aug 2026	Funding $2.13B	Designs Sohu, a transformer-specific ASIC for AI inference.	Gavin Uberti Co-founder & CEO Robert Wachen President Mark Ross CTO	Hardware Sales
Fireworks AI 🇺🇸 Redwood City, CA	AI Inference Platform	Series D Jul 2026	Funding $1.83B	Unified AI inference/training platform running 1000+ models, processing 30T+ tokens daily.	Lin Qiao Co-Founder & CEO	Usage Based Subscription Recurring
Poolside 🇺🇸 San Francisco, CA	AI Foundation Models	Series B Oct 2024	Funding $1.73B	Foundation models for software engineering with proprietary training platform and on-prem deployment.	Jason Warner Co-CEO Eiso Kant Co-CEO	Subscription Recurring Usage Based Professional Services
Together AI 🇺🇸 San Francisco, CA	AI Cloud Infrastructure	Series C Jul 2026	Funding $1.33B	AI Native Cloud with GPU compute, inference, fine-tuning, and proprietary kernel optimizations for 200+ models.	Vipul Ved Prakash CEO Tri Dao Chief Scientist Ce Zhang CTO Charles Zedlewski CPO	Usage Based Subscription Recurring Managed Services
MatX 🇺🇸 Mountain View, CA	AI Accelerator Chips	Series B Feb 2026	Funding $625M	Custom AI accelerator chip (MatX One) for frontier-scale LLM training/inference.	Reiner Pope CEO Mike Gunter CTO	Hardware Sales
TensorWave 🇺🇸 Las Vegas, NV	AI Cloud Infrastructure	Series B Jun 2026	Funding $497M	AMD-exclusive AI cloud with 30,000+ Instinct GPUs across U.S. data centers.	Darrick Horton CEO Piotr Tomasik President & COO	Usage Based
Armada 🇺🇸 San Francisco, CA	Edge AI Infrastructure	Series B May 2026	Funding $469M	Ruggedized modular data centers and full-stack edge computing platform for AI in remote environments.	Dan Wright CEO Jon Runyan COO Pradeep Nair CTO	Hardware Sales Subscription Recurring Usage Based
Modal Labs 🇺🇸 New York, NY	Cloud GPU Infrastructure	Series C May 2026	Funding $465M	Serverless GPU cloud for AI with sub-second cold starts and 0-to-1000+ GPU autoscaling.		Usage Based Freemium
d-Matrix 🇺🇸 Santa Clara, CA	AI Inference Accelerators	Series C Nov 2025	Funding $429M	Digital In-Memory Compute AI inference accelerators for low-latency generative AI at data center scale.	Sid Sheth CEO Sudeep Bhoja CTO	Hardware Sales
Mythic 🇺🇸 Austin, TX	AI Inference Processors	Series D Dec 2025	Funding $418M	Analog compute-in-memory AI inference chips delivering 100x energy efficiency vs GPUs for edge and data center.	Taner Ozcelik CEO Dave Fick CTO	Hardware Sales Licensing Royalties
WEKA 🇺🇸 Campbell, CA	AI Data Platform	Series E May 2024	Funding $415M	Software-defined AI data platform (NeuralMesh) extending GPU memory 1000x. Trusted by 30% of Fortune 50.	Liran Zvibel CEO Amit Pandey Chairman Jonathan Martin President Shimon Ben-David CTO	Subscription Recurring Hardware Sales Usage Based
Exa 🇺🇸 San Francisco, CA	AI Search Infrastructure	Series C May 2026	Funding $357M	Proprietary AI-native web search platform with own crawlers, embeddings, and vector DB for AI agents and enterprise.	Will Bryk Co-Founder & CEO Marcus Holm CRO	Usage Based Subscription Recurring
Redis 🇺🇸 San Francisco, CA	Real-time AI Data Platform	Series G Apr 2021	Funding $357M	In-memory data platform with vector search, agent memory, and context engine (Iris) powering 43% of enterprise AI agent stacks.	Rowan Trollope CEO Yiftach Shoolman President Benjamin Renaud CTO Jon Fritz CPO	Subscription Recurring Usage Based Managed Services
H2O.ai 🇺🇸 Mountain View, CA	Machine Learning Platform	Series E Nov 2021	Funding $251M	Converged predictive/generative AI platform with AutoML, agentic AI, open-weight models, and FedRAMP High sovereign deployment.	SriSatish Ambati CEO Arno Candel CTO Jason Finney President & CRO	Subscription Recurring Usage Based Professional Services
Aerospike 🇺🇸 Mountain View, CA	Real-Time NoSQL Database	Series E Apr 2024	Funding $241M	Real-time distributed NoSQL database with vector search and AI-native developer tooling. 50% of revenue from AI workloads.	Subbu Iyer CEO	Subscription Recurring Professional Services Freemium
Alif Semiconductor 🇺🇸 Pleasanton, CA	Edge AI Microcontrollers	Series B	Funding $235M	AI-enabled Arm-based microcontrollers and fusion processors with on-chip NPUs for battery-powered edge devices.	Syed Ali CEO Reza Kazerounian President	Hardware Sales
Domino Data Lab 🇺🇸 San Francisco, CA	Enterprise AI Platform / MLOps	Series F Jun 2022	Funding $224M	Enterprise AI platform for building, deploying, and governing AI/ML in regulated industries. 20%+ of Fortune 100 as customers.	Nicholas Elprin CEO Christopher Yang CTO Thomas Robinson COO	Subscription Recurring Professional Services
Tensordyne 🇺🇸 Sunnyvale, CA	AI Inference Systems	Series C Jul 2025	Funding $211M	Rack-scale AI inference systems with proprietary logarithmic math (Pareto/LNS), 3nm custom silicon, and sub-microsecond interconnect.	Marc Bolitho CEO RK Anand CPO	Hardware Sales
Goodfire 🇺🇸 San Francisco, CA	AI Interpretability Infrastructure	Series B Feb 2026	Funding $207M	Mechanistic interpretability platform (Silico) for understanding, debugging, and designing foundation models across life sciences, robotics, and LLMs.	Eric Ho CEO Tom McGrath Chief Scientist Dan Balsam CTO	Subscription Recurring Usage Based Professional Services
Labelbox 🇺🇸 San Francisco, CA	AI Training Data Platform	Series D Jan 2022	Funding $189M	AI data factory and RL data engine with 2.6M+ expert network for frontier AI model training.	Manu Sharma CEO Brian Rieger President	Subscription Recurring
CAST AI 🇺🇸 Miami, FL	AI Cloud Cost Optimization	Series C Jan 2026	Funding $181M	Kubernetes automation platform reducing cloud compute costs 50-60% with GPU optimization (OMNI Compute) for 2,100+ customers.	Yuri Frayman CEO	Subscription Recurring Usage Based Marketplace Commission
EnCharge AI 🇺🇸 Santa Clara, CA	AI Semiconductors	Series B Feb 2025	Funding $163M	Analog in-memory computing AI accelerator chips (EN100) delivering ~200 TOPS at 8W for edge-to-cloud inference.	Naveen Verma CEO Kailash Gopalakrishnan CTO Echere Iroaga COO	Hardware Sales
LangChain 🇺🇸 San Francisco, CA	AI Agent Infrastructure	Series B Oct 2025	Funding $160M	Open-source agent frameworks and LangSmith platform for building, evaluating, and monitoring AI agents. 1B+ downloads.	Harrison Chase CEO	Subscription Recurring Usage Based
Deep Infra 🇺🇸 Palo Alto, CA	AI Inference Cloud	Series B May 2026	Funding $154M	Purpose-built AI inference cloud processing 5T tokens weekly with owned GPU infrastructure across 8 U.S. data centers.	Nikola Borisov CEO	Usage Based
Pinecone 🇺🇸 New York, NY	Vector Database	Series B Apr 2023	Funding $138M	Serverless vector database for AI retrieval, RAG, and agent memory at billion-vector scale. 9,000+ customers.	Ash Ashutosh CEO Edo Liberty Chief Scientist	Subscription Recurring Usage Based Freemium
MinIO 🇺🇸 Redwood City, CA	AI Object Storage	Series B Jan 2022	Funding $126M	S3-compatible exabyte-scale object storage for AI with MemKV context memory and AIStor Tables. 77% of Fortune 500.	Garima Kapoor Co-CEO Anand Babu Periasamy Co-CEO Ugur Tigli CTO	Subscription Recurring Licensing Royalties
Lightbits Labs 🇺🇸 San Jose, CA	Software-Defined Block Storage	Series B Jul 2018	Funding $113M	NVMe/TCP disaggregated block storage plus LightInferra KV cache acceleration for AI inference.	Eran Kirzner CEO Avigdor Willenz Chairman Sagi Grimberg CTO	Subscription Recurring Professional Services
Lightning AI 🇺🇸 New York, NY	AI Cloud Infrastructure	Series B Jun 2022	Funding $108M	AI-native cloud platform unifying development, training, deployment with 36,000+ owned GPUs post-Voltage Park merger.	William Falcon CEO	Usage Based Subscription Recurring
Syntiant 🇺🇸 Irvine, CA	Edge AI Semiconductors	Series D Mar 2026	Funding $106M	Ultra-low-power Neural Decision Processors and MEMS microphones for always-on edge AI. 100M+ chips shipped.	Kurt Busch CEO Pieter Vorenkamp COO Stephen Bailey CTO	Hardware Sales Licensing Royalties
Phaidra 🇺🇸 Seattle, WA	AI Data Center Infrastructure	Series B Oct 2025	Funding $92.5M	RL agents for autonomous thermal, power, and cooling control of AI factories and hyperscale data centers.	Jim Gao CEO Katie Hoffman President & COO Veda Panneershelvam CTO	Subscription Recurring
Defined.ai 🇺🇸 Seattle, WA	AI Training Data	Series B May 2020	Funding $78.6M	World's largest ethical AI data marketplace with 817+ datasets and 1.6M+ vetted contributors across 150+ countries.	Daniela Braga CEO	Marketplace Commission Professional Services

What AI infrastructure is

AI infrastructure is the compute, data and tooling layer that AI model builders and application companies run on, rather than the models or apps themselves. The 35 companies on this page sit in seven overlapping segments: data, storage and training infrastructure, GPU cloud and compute, AI chips and accelerators, MLOps and agent platforms, inference and model serving, vector and search infrastructure, and foundation models and training stack.

What separates AI infrastructure from AI application software is where the value is created. An application company sells a finished product, such as a coding assistant or a search engine, to an end user. An infrastructure company instead sells capacity, a platform or a data layer to other AI companies, which is why usage-based and consumption pricing is common here alongside subscription revenue.

Hardware is a bigger share of this sector than most: 10 of 35 companies (29%) generate revenue from hardware sales, mostly AI accelerator and inference chip designers competing on power efficiency and cost per token rather than raw compute alone.

This list covers privately held, US-headquartered companies at Series B or later. Public companies, application-layer AI software and foundation-model labs whose primary product is the model rather than infrastructure to run it are out of scope.

How akta.pro classifies this industry

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

• 518210 Computing Infrastructure Providers, Data Processing, Web Hosting
• 334413 Semiconductor and Related Device Manufacturing

SIC

• 7372 Prepackaged Software
• 3674 Semiconductors and Related Devices

Internal taxonomy

• Data, storage & training infra
• GPU cloud & compute
• AI chips & accelerators
• MLOps, agents & platforms
• Inference & model serving
• Vector & search infra
• Foundation models & training stack

Industry codes are standard federal classification, mapped by revenue model (hardware sales vs. software/cloud/services); they are provided for cross-referencing against public filings and procurement records, not confirmed against akta.pro's own internal assignment. akta.pro classification is independent of them, so a company may carry an internal segment without a matching NAICS or SIC registration.

Methodology

• Universe sourced from akta.pro company list API using 26 AI infrastructure industry codes, filtered to private, operating, US-headquartered companies at Series B+ funding stage with AI maturity scale >= 3.
• 125 candidates retrieved sorted by total funding descending. Each candidate was enriched via akta.pro company_data_concise and semantically qualified against the AI infrastructure sector.
• Qualification criteria: AI infrastructure must be the company's primary product or service, including GPU/compute cloud, AI semiconductors, inference platforms, model serving, AI data infrastructure, vector databases, MLOps platforms, agent frameworks, edge AI chips, AI storage, and data center optimization. Application-layer AI products without core infrastructure components were excluded.
• 35 companies passed full enrichment and semantic qualification. The final count is below the requested 50 due to the depth of per-company enrichment required; all included companies are defensible AI infrastructure matches.
• Ranked by total funding (USD) descending, with latest funding date as tiebreaker.
• All firmographic data — funding, revenue estimates, revenue model, category, management, and offering — sourced from akta.pro enrichment. No fields were populated from web search or model knowledge.

Data via akta.pro Company Database & Management/Product Profiles Compiled 18 Sep 2026