SAIHEAT Limited has entered into a definitive merger agreement with Canopy Wave, Inc., a Santa Clara-based AI inference and GPU cloud platform company, to build a global AI inference platform.
The deal, announced on August 10, 2026, is designed to reposition SAIHEAT around AI inference infrastructure for open-weight large language models, while retaining its existing data center infrastructure business.
The acquisition is strategically important because enterprise AI adoption is shifting investment from one-time model training toward ongoing inference workloads – the delivery of AI model outputs or “tokens” at production scale. Demand for cost-efficient, secure inference of open-weight models is expanding across AI coding, agent, and enterprise workloads.
By combining SAIHEAT’s modular data center infrastructure and energy-efficient computing capabilities with Canopy Wave’s full-stack inference platform, the companies aim to provide a U.S.-based global AI inference platform.
The transaction is expected to close by end of 2026, subject to SAIHEAT shareholder approval, Nasdaq approval of initial listing, and concurrent private placement. Combined company will be renamed Canopy Wave Holdings Inc. and trade under ticker CWAV.
Key Details of the Acquisition
| Detail | Information |
| Acquirer | SAIHEAT Limited (Nasdaq: SAIH) |
| Target | Canopy Wave, Inc. |
| Acquirer Country | Singapore (HQ), Nasdaq-listed |
| Target Country | United States |
| Acquirer Industry | Distributed Computing, Liquid-Cooling Data Centers, Energy-Efficient Computing |
| Target Industry | AI Infrastructure, GPU Cloud, AI Inference Platform, GPU-as-a-Service |
| Deal Type | Merger – All-Stock |
| Deal Value | ~$100 Million combined pre-money valuation |
| Canopy Wave Pre-Money Valuation | $60 Million |
| SAIHEAT Pre-Money Valuation | $40 Million |
| Payment | All-Stock – 3,306,269 ordinary shares issuance |
| Share Breakdown | 2,624,152 Class A shares, 496,442 Class B shares, 185,675 Class A for option pool |
| Implied Price | $18.15 per share |
| Concurrent Private Placement | $4.5 Million Class A shares at $18.15 per share |
| Ownership Post-Close | Canopy Wave shareholders ~54.19% economic, 78.44% voting power |
| Founding Team Control | CEO Tao Zhang and CTO James Liao retain majority economic and voting |
| Announcement Date | August 10, 2026 |
| Acquirer Founded | 2019, Singapore (formerly SAI.TECH Global, renamed Aug 2024) |
| Target Founded | 2024, Santa Clara, California |
| Target HQ | Santa Clara, California, U.S. |
| Combined Name | Canopy Wave Holdings Inc. |
| Combined Ticker | CWAV (Nasdaq), subject to approvals |
| Combined HQ | Santa Clara, California |
| Reporting Transition | From foreign private issuer to domestic issuer next fiscal year |
| Expected Closing | End of 2026 |
| Approvals Required | SAIHEAT shareholders, Nasdaq initial listing, private placement closing |
| Structure | Canopy Wave becomes wholly-owned subsidiary of SAIHEAT via Saiheat Merger Sub, Inc. |
SAIHEAT Company Introduction
SAIHEAT Limited is a Nasdaq-listed distributed computing power operator, headquartered in Singapore, listed under SAIH. Founded in 2019, formerly known as SAI.TECH Global Corporation and renamed SAIHEAT Limited in August 2024.
Its major technology areas include:
- Liquid-cooling data centers
- Modular computing power system
- Advanced computing center ecosystem
- High-performance servers, liquid cooling, systems for capturing and recycling computing heat
- Helping energy owners address local energy consumption and efficient resource utilization
- Carbon-negative data center operations by repurposing heat
- Formerly SAIHUB CAB indoor cabinet turning air-cooled bitcoin mining machines to liquid-cooled, ULTIWIT outdoor infrastructure
Canopy Wave Company Introduction
Canopy Wave, Inc. is a Santa Clara, California-based AI inference and GPU cloud platform company, founded in 2024, headquartered in Santa Clara.
Its key capabilities include:
- Full-stack AI inference platform for open-weight generative AI models
- GPU-as-a-Service (GPUaaS)
- Optimized inference services tailored for open-weight LLMs
- OpenAI-compatible API interfaces
- Intelligent GPU resource scheduling
- Enterprise-grade security: SOC 2 Type II certification, data isolation, zero-data-retention policy
- Supports broad catalog of leading open-weight models
- Serves developers and enterprises across AI coding, AI agent, production workloads
- Access to third-party infrastructure through leasing arrangements
Why Is SAIHEAT Merging with Canopy Wave?
| SAIHEAT Already Has | Canopy Wave Brings |
| Modular data center infrastructure, energy-efficient computing | Full-stack inference platform: GPU cloud, orchestration software, API endpoints |
| Global distributed computing power operator | GPUaaS and optimized inference for open-weight models |
| Liquid-cooling technologies | OpenAI-compatible API, intelligent scheduling |
| Public company platform (Nasdaq SAIH) | Exceptional engineering team (CEO Tao Zhang, CTO James Liao), SOC 2 Type II, zero-data-retention |
Flow: Modular Data Centers + Liquid Cooling → GPU Cloud Infrastructure → Inference Platform → API Endpoints → Enterprise Developers → AI Tokens at Scale
What Does Canopy Wave’s AI Inference Technology Do?
1. Enterprise wants to deploy open-weight LLM (e.g., Llama, Mistral)
2. Canopy Wave provides OpenAI-compatible API – no infrastructure management
3. Intelligent GPU scheduling allocates resources
4. Platform delivers tokens at production scale with security (SOC 2, data isolation, zero retention)
5. Developers build AI coding, AI agent workloads
In simple terms: Open-Weight Model → GPUaaS → Inference Platform → Secure API → Enterprise AI Application. Eliminates infrastructure bottlenecks.
Why Is AI Inference Important?
| Model Training (Past Focus) | AI Inference (Future Focus – This Deal) |
| One-time expensive training of frontier models | Ongoing delivery of model outputs/tokens at production scale |
| Closed proprietary models dominant | Open-weight models closing capability gap with proprietary |
| High upfront CapEx | Growing share of AI infrastructure spending |
| Limited enterprise use | Expanding across coding, agents, enterprise AI workloads |
SAIHEAT management believes inference represents growing share as enterprises evaluate open-weight models for performance, control, cost efficiency.
$100 Million All-Stock Merger
- Canopy Wave valued $60M pre-money, SAIHEAT $40M pre-money – result of arm’s length negotiation
- Total consideration 3,306,269 ordinary shares at $18.15 per share
- Breakdown: 2,624,152 Class A shares, 496,442 Class B shares, 185,675 Class A for option pool
- Concurrent private placement $4.5M Class A shares at $18.15 per share
- Former Canopy Wave stockholders own ~54.19% economic, 78.44% voting power
- Tao Zhang and James Liao collectively hold majority economic and voting power post-close
- Combined company to be renamed Canopy Wave Holdings Inc., ticker CWAV, HQ Santa Clara
- Unanimously approved by boards of both companies, CEO Jianwei Li said: “This combination will position the company where the AI market is going: inference at scale”
Canopy Wave’s Revenue Base and Financial Position
| Metric | Figure |
| Founded | 2024, Santa Clara, California |
| HQ | Santa Clara, California |
| Business | AI inference and GPU cloud platform for open-weight LLMs |
| Pre-Money Valuation | $60 Million |
| Platform Features | GPUaaS, orchestration software, API endpoints, SOC 2 Type II, zero-data-retention |
| Customers | Developers and enterprises across AI coding, AI agent, production workloads |
| Operating History | Limited since 2024 (per risk factors) |
| Access | Third-party infrastructure via leasing (does not own) |
| Market | AI inference infrastructure for open-weight models |
SAIHEAT’s Work Across Computing Programs
| Area | Examples |
| Liquid-Cooling Data Centers | SAIHUB CAB indoor cabinet, ULTIWIT outdoor infrastructure |
| Advanced Computing Center | High-performance servers, liquid cooling, heat recycling |
| Distributed Computing | Modular computing power system, energy-efficient computing |
| Energy | Helping energy owners address local consumption, carbon-negative operations |
| Former Business | Cryptocurrency mining liquid-cooling (per older profile), Bitcoin purchase $1.5M in 2024 |
What Could SAIHEAT Gain?
- 1. Pivot into AI Inference Infrastructure – Moves from crypto/distributed computing to high-growth AI inference
- 2. Full-Stack Platform – Combines modular data centers with GPU cloud, orchestration, API endpoints
- 3. U.S.-Based Leadership – Post-close HQ Santa Clara, led by Canopy Wave founding team CEO Tao Zhang and CTO James Liao, transition to domestic issuer reporting
- 4. Public Platform for Scale – Canopy Wave gets public company platform and infrastructure depth to scale faster globally
SAIHEAT Financial Position
| Metric | Figure |
| Stock | Nasdaq: SAIH, formerly SAI.TECH Global |
| Founded | 2019, Singapore, renamed Aug 2024 |
| Business | Distributed computing, liquid-cooling data centers |
| Pre-Money Valuation in Deal | $40 Million |
| CEO | Jianwei Li |
| Post-Deal | Retains existing data center infrastructure business while adding AI inference |
| Private Placement | $4.5M concurrent financing at $18.15 per share |
Potential Challenges
| Risk Area | Details |
| Closing Conditions | Shareholder approval, Nasdaq approval, private placement closing, regulatory clearances |
| Dependence | On third-party open-weight AI models (including outside U.S.) – export controls, trade restrictions |
| Infrastructure Reliance | Reliance on third-party computing infrastructure via leasing – subject to termination |
| Pricing Pressure | Declines in per-token pricing or GPU rental rates |
| Limited History | Canopy Wave founded 2024 – limited operating history |
| Customer Concentration | Per risk factors |
| Voting Concentration | 78.44% voting power with founding team – concentration risk |
| Listing Risk | Continued Nasdaq listing until closing and combined company listing after |
| Transition Costs | Transition from foreign private issuer to domestic issuer |
Outcome
The SAIHEAT-Canopy Wave $100M all-stock merger pivots SAIHEAT from liquid-cooling data centers to AI inference. SAIHEAT, founded 2019 in Singapore as SAI.TECH and valued at $40M, merges with Canopy Wave, founded 2024 in Santa Clara and valued at $60M, via 3.3M shares at $18.15 plus $4.5M private placement. Canopy Wave holders get 54.19% economic and 78.44% voting, led by CEO Tao Zhang and CTO James Liao, with combined company renamed Canopy Wave Holdings and ticker CWAV in Santa Clara. For SAIHEAT, it adds full-stack GPUaaS platform with OpenAI-compatible API, SOC 2 Type II and zero-retention for open-weight LLMs, targeting inference-at-scale market. Risks include closing conditions, third-party infrastructure dependence, token pricing declines, and limited history. If closed end of 2026, it creates U.S.-based global AI inference platform for coding and agent workloads.
Source: SEC, saiheat news
