SAIHEAT has filed a prospectus with the Securities and Exchange Commission (SEC) for a $300M mixed shelf offering. The registration statement, accessible via the SEC's EDGAR database, permits the company to issue various securities including debt and equity. This filing provides SAIHEAT with the flexibility to raise capital over time as market conditions dictate.

Offering Details

The mixed shelf offering allows SAIHEAT to offer different types of securities, either singly or in combination. The total amount offered will not exceed $300M. The prospectus outlines the general terms under which these securities may be sold, though specific details regarding the timing, size, and nature of the securities will be determined in future offerings.

Regulatory Filing

The prospectus was filed with the SEC and is available for public review. The document serves as a preliminary registration, allowing the company to act quickly when it decides to undertake a specific offering. Investors can refer to the official SEC filing for comprehensive details regarding the potential securities and associated risks.

SAIHEAT Limited announced its strategic expansion into the AI inference services business on June 11, 2026, delivering enterprise-level authorized token access to mainstream open-source AI models. The company aims to empower corporate clients to streamline R&D progress and iterative upgrades of intelligent application products by providing high-performance and low-latency capabilities. This expansion addresses challenges businesses face in scaling AI from pilot projects to full production without managing complex GPU clusters or data center operations.

The platform is built on dedicated, high-performance AI infrastructure utilizing proprietary optimization technologies. SAIHEAT enables organizations to scale AI applications rapidly by offering direct access tokens from the latest open-source models. The solution supports both open-source and custom AI models at scale, catering to vertical AI companies and digital enterprises seeking to deploy and optimize AI workloads efficiently.

SAIHEAT's inference platform leverages an advanced cluster architecture to support a wide range of mainstream AI frameworks. It manages highly concurrent workloads across model training and inference, emphasizing high-quality, low-latency, and secure AI inference services. This architecture allows enterprises to integrate cutting-edge open-source AI models into production with speed and simplicity.

"Many businesses are hitting a wall trying to scale AI from pilot projects to full production. Solely using closed models creates bottlenecks. We solve this by combining the best open-source models with robust enterprise infrastructure, offering an adaptable and scalable path to get AI working in your real-world applications," said Jianwei Li, CEO of SAIHEAT. "Scale your AI without worry. Our platform offers the dedicated, high-capacity infrastructure your enterprise needs."

Supported AI Platforms

SAIHEAT's existing and prospective partners cover a diverse range of AI platforms. The following table outlines the key models supported by the new inference service:

As AI adoption accelerates across various industries, SAIHEAT intends to establish itself as a trusted infrastructure partner. The company focuses on delivering secure, high-performance AI inference services that help enterprises unlock the full potential of open AI models and seize opportunities in the AI-driven economy.