ai&
ai& is a Japan-headquartered AI infrastructure company offering sovereign, Japan-hosted inference of frontier open-weight models via an OpenAI/Anthropic-compatible API, paired with the Build.io PaaS for application deployment, targeting Japanese enterprises and developers in regulated sectors.
- Company typePrivate
- Founded2026
- HeadquartersKanagawa, Japan
- Headcount11–50
- GTM typeB2B
- OfferingSoftware
What ai& does
ai& is a Japan-headquartered AI infrastructure and model services company that operates a vertically integrated stack spanning owned modular data centers, heterogeneous compute, frontier-model inference, and an application deployment platform. Founded in 2026 by David Bennett (former CCO of Tenstorrent and former CEO & President of NEC Personal Computers / Lenovo Japan) and Shimpei Hara (former technology executive at Tenstorrent), the company launched on March 25, 2026 alongside a $50 million seed round and more than $2 billion in committed data center capital. Its core user-facing product, ai& Inference, exposes a single OpenAI/Anthropic SDK-compatible REST API serving frontier open-weight models (Kimi, DeepSeek, GLM, Qwen, gemma, gpt-oss) from Japan-resident infrastructure, with marketing claims of under-50ms Japan latency, up to 90% cost savings versus incumbents, and APPI-aligned data residency.
The underlying platform is built around ten data center sites in development across Japan, the US, and Asia — including a 100MW+ campus commitment over three years — using heterogeneous compute from AMD, NVIDIA, and Tenstorrent, with two of four established sites already serving design partners. In April 2026, ai& acquired Build.io, adding a full-stack Platform-as-a-Service with nearly 60 real-world deployments, bare-metal infrastructure expertise, and Japan's first SBOM-native deployment platform, extending the stack down to the application layer. The company is also establishing a dedicated AI Research Lab in Japan to train Japanese-language and culturally localized models.
ai& monetizes through a hybrid revenue model: usage-based per-token pay-as-you-go inference for developers and SMBs via the self-serve console, quote-based enterprise contracts with reserved compute, custom SLAs, and on-premises or private-network deployment, and a Build.io PaaS subscription tier. Go-to-market combines self-serve API motion (global, PLG-style), direct enterprise sales (Japan-led, expanding to APAC, Europe, and the Middle East), and a channel partner network including Ryoyo Ryosan (hardware procurement and enterprise reselling), Tomorrow Net (SI and production deployment across manufacturing, financial services, public sector, energy, telecom, and IT services), and Plug and Play Japan (startup and corporate innovation pipeline across 60+ global locations). Strategic vertical co-development is anchored in healthcare via GENSHI AI, with full-scale medical AI rollout targeted for fiscal year 2026.
ai& firmographics
Firmographics- Name
- ai&
- Legal name
- ai& Inc.
- Website
- https://aiand.com
- Company type
- Private
- Founded year
- 2026
- Operating status
- Operating
- Headcount range
- 11–50 employees
- Short description
- ai& is a Japan-headquartered AI infrastructure company offering sovereign, Japan-hosted inference of frontier open-weight models via an OpenAI/Anthropic-compatible API, paired with the Build.io PaaS for application deployment, targeting Japanese enterprises and developers in regulated sectors.
- Ownership category
- akta.pro rank
ai& industry classification
Industry- Product category
- AI Infrastructure
- NAICS
- Computing Infrastructure Providers, Data Processing, Web Hosting, and Related Services (5182)
- SIC
- Services-Prepackaged Software (7372)
- akta.pro primary industry
- Model Serving, Inference & Deployment Platforms (APIs, Edge/On-Prem) (HDAEANAC)
- akta.pro secondary industries
- AI Application Enablement Platforms (Copilot/Agent Frameworks, SDKs) (HDAEANAJ), End-to-End Enterprise AI Platforms (MLOps & Model Lifecycle Management) (HDAEANAA), Model Hosting, Serving & Inference Platforms (HDAAACAB)
Keywords
Where ai& is headquartered
LocationHeadquarters
- HQ city
- Kanagawa
- HQ country
- Japan
- HQ region
- Asia
Offices2 records
Markets served
ai& business model
Business model- GTM type
- B2B
- Offering type
- Software
- Cost components
- Infrastructure, Technology or R&D, Personnel, Marketing or Sales, Operations
Revenue model
- ai& Inference (API): Usage-based consumption revenue from pay-as-you-go API inference; priced per million input/output tokens across multiple frontier models (Kimi, DeepSeek, GLM, Qwen, gemma, gpt-oss). No contracts required; pay only for what is used.
- Enterprise tier: Custom enterprise contracts with reserved compute, custom SLA, advanced security, dedicated infrastructure, priority support, and terms built around customer requirements.
- AI Data Center & cloud services: Revenue from AI compute, data center services, and AI cloud offered to select design partners and enterprise customers using equipment from AMD, NVIDIA, and Tenstorrent; two of four established data centers currently operational.
- Build.io PaaS revenue (inherited): Build.io contribution of multi-million-dollar revenue from nearly 60 real-world deployments across US, Europe, and Japan following April 2026 acquisition.
Pricing tiers
| Model | Billing | Price |
|---|---|---|
| Usage-based | Pay-as-you-go | Inference – Pay as you go (self-serve) |
| Subscription | Multi-year contract | Enterprise (custom) |
| Subscription | Monthly | Build (developer deployment tier) |
Go-to-market motion4 records
Distribution channels6 records
Marketing channels7 records
ai& product offering
Product offeringCore offering
ai& operates a vertically integrated AI platform that unifies ai&-owned modular data centers, heterogeneous compute (AMD, NVIDIA, Tenstorrent), frontier model services, and an application deployment layer into a single stack. Its two named products are ai& Inference, an OpenAI/Anthropic-SDK-compatible REST API serving frontier open-weight LLMs from Japan-resident infrastructure with pay-as-you-go per-token pricing, and Build (formerly Build.io, acquired April 2026), a full-stack PaaS for shipping apps, agents, and products to production. Underpinning the platform are AI Data Center, AI Compute, and AI Cloud services available to enterprise design partners.
Product overview
ai& offers a vertically integrated global AI platform built end-to-end in Japan that unifies owned data centers, heterogeneous compute, and frontier model services into a single stack. The platform centers on two named products: ai& Inference, the core OpenAI-compatible API that serves frontier open-weight models (Kimi K2.6, Kimi K2.7 Code, GLM-5.1, GLM-5.2, DeepSeek V4 Flash/Pro, gemma-4-31B-it, Qwen3.6-27B, gpt-oss-120b) from Japan-resident infrastructure, and Build (the Build.io PaaS acquired in April 2026), which lets teams deploy apps, agents, and products to production on the same infrastructure. Underpinning these are supporting services — AI Data Center (ten sites in development, including a 100MW+ campus commitment), AI Compute (heterogeneous compute across AMD, NVIDIA, and Tenstorrent hardware), AI Cloud, and an AI Research Lab focused on localized Japanese model development. Together, Inference plus Build form the user-facing 'full stack' that combines 1-click AI deployment with sovereign Japan-hosted infrastructure.
Differentiator
Problem solved
Functional benefit
Brands
- ai& Inference: Frontier AI inference platform serving models from Japan via an OpenAI/Anthropic-compatible API.
- ai& Build
- Build.io
Products and services
- ai& Inference Frontier-model inference API hosted entirely in Japan, delivering frontier open-weight LLMs (Kimi K2.6, Kimi K2.7 Code, GLM-5.1, GLM-5.2, DeepSeek V4 Flash, DeepSeek V4 Pro, gemma-4-31B-it, Qwen3.6-27B, gpt-oss-120b) through a single OpenAI-compatible REST endpoint. Offers <50ms Japan latency, 100% Japan-hosted data residency, no infrastructure to manage, and an enterprise tier with reserved compute, custom SLA, and advanced security. Targets developers and enterprises needing sovereign, low-cost AI inference.
- Build (formerly Build.io) Full-stack Platform-as-a-Service that takes product teams from code to production in minutes, powered by bare-metal infrastructure and AI workflows across the US, Europe, and Japan. Provides SBOM-native supply chain integrity, zero-config TLS, global edge, secure-by-default deploys, and keeps IP and database in Japan. Targets product teams and developers deploying apps, agents, APIs, and e-commerce stores.
- AI Data Center ai&'s owned data center infrastructure, with ten sites in development across Japan, the US, and Asia, including a committed 100MW+ campus in Japan over the next three years. Two of four established data centers are currently operational, serving design partners using equipment from AMD, NVIDIA, and Tenstorrent. Targets enterprises and design partners needing sovereign, large-scale AI capacity.
- AI Compute
- AI Cloud
Quantifiable outcome
- Up to 90% potential savings vs. OpenAI or Anthropic
- +4 more outcomes
Companies that use ai&
Customer profileNamed customers4 records
Segments5 records
Ideal customer profiles5 records
ai& technology and API
TechnologyTechnology focussed Yes
API detail
- Has API
- Yes
- API docs
- API detail
Core technology
AI maturity
App detail
Integration9 records
AI capability8 records
Feature7 records
ai& partnerships and signals
Strategic signalPartnerships
Ten partnerships are on record, tiered flagship ecosystem partnership — first of its kind across plug and play's global network of 60+ locations; provides startups and corporate partners direct access to ai&'s compute and localized ai models., core japan channel partner — provides hardware procurement, configuration expertise, and enterprise sales network for ai& inference solutions., major japan si alliance targeting manufacturing, financial services, public sector, energy, telecom, and it services — designed to move enterprise ai from pilot to production., core vertical (healthcare) co-development partner — flagship proof point for sovereign medical ai in japan., core hardware / compute partner — ai& integrates tenstorrent galaxy blackhole infrastructure alongside amd and nvidia., acquired paas — extends ai& stack from infrastructure all the way to the application layer; brings nearly 60 real-world deployments, sbom-native security, and contributors from cisco's cilium, rapid7, and the linux kernel., indirect — background ip and contributor pedigree inherited through the build.io acquisition; relevant to sbom-native deployment and security posture., core heterogeneous compute vendor — amd hardware is one of three pillars of ai&'s heterogeneous inference platform., core heterogeneous compute vendor — nvidia hardware is one of three pillars of ai&'s heterogeneous inference platform. and foundational infrastructure contributor — unsung fields transitioned its data center cloud operations into ai& at launch..
- Plug and Play Japan KKflagship ecosystem partnership — first of its kind across plug and play's global network of 60+ locations; provides startups and corporate partners direct access to ai&'s compute and localized ai models.Collaboration to help enterprises build and deploy AI with greater speed, control, and confidence. Connects Plug and Play's AI Center of Excellence (AI CoE) with ai&'s vertically integrated AI platform. ai&'s AI data centers and cloud services provide direct access to large-scale compute for AI CoE partners, integrating hardware from AMD, NVIDIA, and Tenstorrent. ai&'s continual learning models are combined with Plug and Play's network across 60+ locations. Models optimized for Japanese language, culture, and enterprise environment.
- Ryoyo Ryosan, Inc.core japan channel partner — provides hardware procurement, configuration expertise, and enterprise sales network for ai& inference solutions.Partnership to accelerate enterprise adoption of AI inference in Japan. Ryoyo Ryosan acts as ai&'s key hardware partner (procurement and optimal configuration of GPU-accelerated server infrastructure) and as enterprise sales partner for ai& inference solutions, leading proposals and implementation support for domestic enterprises.
- Tomorrow Net Co., Ltd.major japan si alliance targeting manufacturing, financial services, public sector, energy, telecom, and it services — designed to move enterprise ai from pilot to production.Business alliance to provide generative AI and AI/IT infrastructure solutions for enterprise customers. Establishes an integrated support structure spanning design and construction of AI inference environments, generative AI implementation, production operations, and continuous improvement. Joint solution combines ai&'s inference platform with Tomorrow Net's CAT.AI conversational AI service and supports on-premises, closed/private network, hybrid, and public cloud deployments with enterprise-grade 24/7 operations.
- GENSHI AI Inc.core vertical (healthcare) co-development partner — flagship proof point for sovereign medical ai in japan.Collaboration to build a secure Japan-based LLM platform for healthcare institutions. PoC underway, with full-scale rollout targeted in fiscal year 2026. Validates GENSHI AI's 'MedLocal' solution on ai&'s domestic inference solution; develops AI agents for clinical research and EMR integration; aligns with the Act on the Protection of Personal Information and the Three Ministries / Two Guidelines.
- Tenstorrentcore hardware / compute partner — ai& integrates tenstorrent galaxy blackhole infrastructure alongside amd and nvidia.ai& is named among deployment partners integrating Tenstorrent Galaxy Blackhole AI compute systems into their AI platforms and services. ai& also operates Tenstorrent hardware in its established data centers serving design partners.
- Build.ioacquired paas — extends ai& stack from infrastructure all the way to the application layer; brings nearly 60 real-world deployments, sbom-native security, and contributors from cisco's cilium, rapid7, and the linux kernel.Acquisition of Build.io, a full-stack PaaS provider. Build.io takes product teams from code to production in minutes, powered by bare-metal infrastructure and AI workflows across US, Europe, and Japan. Combined with ai&, creates a complete vertically integrated AI stack spanning hardware to application layer.
- Cisco (Cilium) and Rapid7 (open-source contributions via Build.io team)indirect — background ip and contributor pedigree inherited through the build.io acquisition; relevant to sbom-native deployment and security posture.Build.io team includes contributors to Cisco's Cilium Software Defined Networking for Datacenters, the Rapid7 open-source security suite, and the Linux Kernel. Not a direct partnership with ai&, but a contributor-pedigree signal relevant to platform capabilities.
- AMDcore heterogeneous compute vendor — amd hardware is one of three pillars of ai&'s heterogeneous inference platform.ai& integrates AMD hardware alongside NVIDIA and Tenstorrent across architectures optimized for performance, efficiency, and cost. Two of four established data centers are currently operational providing hardware and services to a select group of design partners using AMD equipment. AMD Japan (Jon Robottom, CVP and GM) has appeared publicly with ai& CEO David Bennett.
- NVIDIAcore heterogeneous compute vendor — nvidia hardware is one of three pillars of ai&'s heterogeneous inference platform.ai& integrates NVIDIA hardware alongside AMD and Tenstorrent across architectures optimized for performance, efficiency, and cost. NVIDIA hardware is used in ai&'s established data centers serving select design partners.
- Unsung Fields, Inc.foundational infrastructure contributor — unsung fields transitioned its data center cloud operations into ai& at launch.With the launch of ai&, Unsung Fields, Inc., a Japan-based data center cloud company, transitioned its infrastructure and operations into the ai& platform and ceased operating as an independent entity.
Scale indicators6 records
Recent moves7 records
Expansion highlights7 records
ai& competitors and assessment
Company assessmentRegional players
- Preferred Networks: Preferred Networks is a Japan-based deep learning and AI infrastructure company with its own MN-3a supercomputer; comparable as a Japanese provider of AI infrastructure targeting domestic enterprise and research workloads.
- Sakana AI: Sakana AI is a Tokyo-based AI research company building Japan-focused foundation models; comparable as a Japan-headquartered AI company pursuing sovereign AI for Japanese enterprise and government customers.
- Sakura Internet: Sakura Internet is a Japanese cloud and data center provider that has explicitly positioned around sovereign Japanese AI and operates domestic GPU clusters, comparable in providing Japan-resident AI compute for regulated workloads.
Direct peers
- OpenRouter: OpenRouter is a unified API and routing layer for multiple LLM providers, directly comparable in offering OpenAI-compatible endpoints, multi-model selection, and developer-first inference with cost and latency optimization.
- Modal Labs: Modal Labs offers serverless AI compute and inference infrastructure with a developer-first API, comparable to ai& in providing frictionless GPU-backed model serving for AI-native applications.
- Replicate: Replicate provides cloud-based model serving for open-weight ML models via API, comparable to ai& in serving multiple third-party models through a standardized endpoint targeting developers.
- Anyscale: Anyscale (creators of Ray) runs Anyscale Endpoints and an enterprise AI compute platform, directly comparable in serving open-weight LLMs at scale with infrastructure-level optimization and enterprise deployment tooling.
- DeepInfra: DeepInfra operates a low-cost serverless inference platform for open-weight LLMs and diffusion models, directly comparable in API-first open-weight model serving with cost leadership positioning.
- Together AI: Together AI operates an AI cloud and inference platform serving open-weight models (Llama, DeepSeek, Qwen) via an OpenAI-compatible API, directly comparable to ai& Inference in product, business model, and target customer base.
- Fireworks AI: Fireworks AI is a model inference and fine-tuning platform serving frontier open-weight models via API, with similar developer-first positioning, OpenAI-compatible endpoints, and a focus on cost-per-token economics versus frontier proprietary APIs.
Market position
Strengths5 records
Weaknesses5 records
Competitive moat7 records
Key risks6 records
Key highlights7 records
Customer concentration
ai& social profiles
Digital presenceai& compliance and trust
Trust signalCompliance2 records
ai& financial estimates
Financial estimateRevenue estimate
Valuation estimate
ai& leadership team
Management profileNumber of profiles
Profiles2 records
ai& subsidiaries and ownership
Company hierarchySubsidiaries2 records
ai& funding detail
Funding detailFunding overview
Funding rounds1 record
Investors
Funding detail is available on the Subscription and Enterprise plan.Contact sales →
ai& M&A and investment
M&A and investmentM&A1 record
Investments
M&A and investment is available on the Subscription and Enterprise plan.Contact sales →
Frequently asked questions about ai&
What does ai& do?
ai& operates a vertically integrated AI platform that unifies ai&-owned modular data centers, heterogeneous compute (AMD, NVIDIA, Tenstorrent), frontier model services, and an application deployment layer into a single stack. Its two named products are ai& Inference, an OpenAI/Anthropic-SDK-compatible REST API serving frontier open-weight LLMs from Japan-resident infrastructure with pay-as-you-go per-token pricing, and Build (formerly Build.io, acquired April 2026), a full-stack PaaS for shipping apps, agents, and products to production. Underpinning the platform are AI Data Center, AI Compute, and AI Cloud services available to enterprise design partners.
Is ai& a public or private company?
ai& is a private company. It is classified as venture growth investor backed and is currently operating.
When was ai& founded?
ai& was founded in 2026. It employs 11 to 50 people.
Where is ai& based?
ai& is headquartered in Kanagawa, Japan, in the Asia region.
How does ai& make money?
Four revenue lines are on record. ai& Inference (API) is the primary driver. The others are enterprise tier, AI Data Center & cloud services and build.io PaaS revenue (inherited).
Who are ai&'s main competitors?
Regional players on record are Preferred Networks, Sakana AI and Sakura Internet. Direct peers are OpenRouter, Modal Labs, Replicate, Anyscale, DeepInfra, Together AI and Fireworks AI.
Does ai& have an API?
Yes. ai& provides a public OpenAI-compatible REST API for running frontier AI models served from Japan. The single API endpoint (BASE_URL=https://api.aiand.com/v1) supports text and code generation across frontier open-weight models including Kimi K2.6, Kimi K2.7 Code, GLM-5.1, GLM-5.2, DeepSeek V4 Flash, DeepSeek V4 Pro, gemma-4-31B-it, Qwen3.6-27B, and gpt-oss-120b. The API is drop-in compatible with the OpenAI and Anthropic SDKs (users can switch by changing one line of code), supports streaming, tool calling, and structured outputs, and is offered as a hosted inference product (console.aiand.com) with pay-as-you-go pricing and an enterprise tier offering reserved compute, custom SLA, and advanced security. Developer documentation is at docs.aiand.com.
What industry is ai& in?
ai&'s product category is AI Infrastructure. Its primary akta.pro industry code is HDAEANAC, Model Serving, Inference & Deployment Platforms (APIs, Edge/On-Prem), with a secondary code of HDAEANAJ, AI Application Enablement Platforms (Copilot/Agent Frameworks, SDKs). Its NAICS code is 5182 and its SIC code is 7372.