HIVE Secures $129 Million Financing for 2,303 NVIDIA AI GPUs

Under Cohere Contract

Introduction

HIVE Digital Technologies has secured approximately $129 million in fixed-rate financing through its BUZZ High Performance Computing (HPC) business to fund an advanced artificial intelligence computing cluster intended for Canadian AI company Cohere.

Announced on October 27, the financing carries a fixed annual interest rate of 9% over three years. The funds will support the purchase of 2,303 NVIDIA GB200 GPUs and associated infrastructure, forming part of a project valued at approximately $165 million.

The hardware is intended to support a three-year contract with Cohere that HIVE values at $225 million in total contract value, including GPU computing, connectivity and storage services. HIVE expects delivery of the GPUs toward the end of the fourth quarter of 2026.

The transaction highlights how cryptocurrency mining companies are increasingly using their infrastructure, energy resources and financing capabilities to enter the AI computing market. It also illustrates the financial commitments involved in converting contracted demand for AI services into an operating business.

Key Highlights

• $129 million financing: HIVE's BUZZ HPC subsidiary secured a fixed-rate credit facility through its indirect subsidiary BUZZ Alpha Cloud.

• 9% fixed annual interest: The facility has a three-year term, with principal and interest amortized over the loan period.

• 2,303 NVIDIA GB200 GPUs: The financing supports an AI computing cluster valued at approximately $165 million.

• $225 million Cohere contract: HIVE expects the cluster to support a three-year agreement covering GPU computing, connectivity and storage.

• Approximately 80% loan-to-cost: The financing covers around 80% of the stated cluster cost.

• No new common shares issued for the loan: HIVE says the facility is structured as debt rather than an equity issuance.

• Delivery expected in late 2026: The GPUs have not yet been delivered, so deployment and revenue realization remain future milestones.

1. How the $129 Million Financing Is Structured

The credit facility was entered into by BUZZ Alpha Cloud Inc., an indirect subsidiary of HIVE Digital Technologies through BUZZ HPC. The lender is described as an institutional lender, but HIVE has not publicly identified it in the announcement.

The main terms are:

Financing detail	Reported terms
Credit facility	Approximately $129 million
Annual interest rate	9% fixed
Loan term	Three years
Project value	Approximately $165 million
Loan-to-cost ratio	Approximately 80%
Intended repayment source	Contract revenue
Security	GPUs and related assets, plus specified subsidiary-level support
Common-share issuance for this facility	None
Expected GPU delivery	Toward the end of Q4 2026

The loan is secured by the GPUs and related assets of BUZZ Alpha Cloud. It also includes a limited-recourse guarantee from BUZZ HPC and a pledge of BUZZ HPC's equity in BUZZ Alpha Cloud.

Principal and interest are scheduled to amortize over the life of the loan. HIVE says it expects to own the GPUs outright at the end of the term without a residual payment.

The financing is structured around a defined infrastructure project and a contracted customer relationship. That gives the lender identifiable collateral and a planned source of repayment, although it does not eliminate credit, operational or asset-value risks.

HIVE also says the facility contains customary covenants, including required debt-service coverage ratios. The company has not publicly detailed every aspect of the repayment schedule or the lender's identity in the announcement.

2. The NVIDIA GPU Cluster and Cohere Agreement

The financing supports a cluster of 2,303 NVIDIA GB200 GPUs and associated infrastructure. HIVE has placed the project cost at approximately $165 million.

The cluster is intended to serve Cohere under a three-year fixed-term agreement announced in June 2026. HIVE now reports a total contract value of $225 million, including GPU compute and additional revenue attributed to connectivity and storage.

The distinction between the overall contract value and the computing component matters. The $225 million figure represents the reported value of the complete services agreement, not pure GPU rental revenue or guaranteed profit.

HIVE has said that the hardware is expected to be delivered toward the end of the fourth quarter of 2026. The deployment schedule therefore remains an important factor in when the company can begin delivering services and recognizing revenue under the agreement.

Cohere develops enterprise AI models and services. Demand from AI companies for advanced computing infrastructure has created opportunities for data-centre operators that can provide high-performance GPUs, reliable power, networking and storage.

However, a signed contract is only one part of the business model. The project must also be delivered, installed, commissioned and operated at a cost that supports debt repayment and an acceptable commercial return.

3. How the Company Plans to Fund the Remaining Project Cost

The $129 million facility represents approximately 80% of the stated $165 million cluster cost. The arithmetic implies a difference of roughly $36 million before considering any additional hardware, software or other deployment expenses.

HIVE says its equity contribution toward the cluster and associated costs will be funded from proceeds of zero-interest convertible bonds issued earlier in 2026.

The company reported gross proceeds of $245 million from those bonds, with net proceeds of approximately $200 million after capped calls.

This structure combines secured debt with previously raised capital rather than issuing new common shares specifically for the $129 million facility.

The approach may help HIVE limit immediate common-share dilution from this financing. However, convertible bonds carry their own terms and potential dilution implications, depending on how they are settled and whether conversion occurs.

The project also remains exposed to costs that may not be fully represented by the cluster's headline purchase price. These can include deployment, power, networking, cooling, maintenance, staffing and other operating expenses.

4. Why Bitcoin Miners Are Moving Into AI Infrastructure

HIVE's announcement reflects a broader strategic shift among digital-asset mining companies.

Bitcoin mining relies on specialized computing hardware, access to electricity, data-centre operations and the ability to manage large-scale infrastructure. AI computing requires a different class of hardware and software, but it also depends on reliable power, cooling, networking and operational expertise.

Companies that already own data-centre facilities or have established energy infrastructure may be able to use some of those capabilities to support AI and high-performance computing workloads.

The revenue models are different, however.

Bitcoin mining revenue depends on factors such as the Bitcoin price, network difficulty, mining efficiency, transaction fees and electricity costs. AI cloud revenue is more directly linked to customer contracts, GPU availability, utilization, service delivery, pricing and customer creditworthiness.

A multiyear AI contract can provide greater revenue visibility than relying entirely on cryptocurrency market conditions. At the same time, it creates obligations to deliver capacity and maintain service over the contracted period.

HIVE's financing therefore illustrates a move toward contract-backed infrastructure investment rather than a complete departure from the risks associated with digital assets.

5. What the $225 Million Contract Means for Revenue Visibility

HIVE reports that the Cohere agreement has a total contract value of $225 million over three years, including compute, connectivity and storage.

Dividing the total contract value by three gives an arithmetic average of approximately $75 million per year. This is a simple average of the announced contract value, not a confirmed annual payment schedule or a measure of profit.

The facility's 9% fixed rate gives HIVE greater certainty about the stated interest rate than it would have under a variable-rate loan. Nevertheless, the company must generate enough cash to cover principal repayments, interest, operating costs and contractual obligations.

A complete assessment of the project's economics would require further details, including:

• The precise timing of Cohere's payments.

• The cluster's power and operating costs.

• The full repayment schedule and applicable covenants.

• Installation and commissioning expenses.

• Expected GPU utilization and service pricing.

• The useful economic life and potential resale value of the GPUs.

These details are important because contract value alone does not determine cash flow or profitability. The project must generate revenue on time and at sufficient margins to cover its expenses and financing commitments.

6. Sovereign AI Infrastructure and Canadian Data Control

HIVE has framed the project as part of its strategy to expand sovereign AI infrastructure in Canada.

The company says the computing capacity funded through the facility will remain within Canadian borders. This positioning is relevant to organizations that consider data location, jurisdictional control and infrastructure independence important when selecting AI services.

Sovereign AI infrastructure can be attractive to enterprises, researchers and public institutions that want more control over where their data is processed and where their models operate.

The practical meaning of sovereignty depends on the complete arrangement, including where the servers are located, who operates them, which laws apply, how data is protected and who has access to the systems.

HIVE's announcement supports its stated intention to keep the computing infrastructure in Canada. It does not, by itself, establish that every customer workload will have the same legal, operational or data-governance characteristics.

7. Key Risks: Debt, Delivery and Hardware Depreciation

Despite the contract-backed financing structure, the project faces several material risks.

GPU delivery and commissioning: HIVE expects the GPUs toward the end of Q4 2026. Delays could push back deployment and the start of revenue generation.

Customer concentration: The financing depends partly on performance under the Cohere contract. A customer failing to meet its obligations, changing the contract or not renewing it could affect projected cash flow.

Debt-service requirements: Revenue must be sufficient to cover principal, interest, operating expenses and required financial covenants. The fixed interest rate does not guarantee that the loan will be repaid without difficulty.

Hardware depreciation: AI hardware can lose value as newer and more efficient GPU generations enter the market. Faster-than-expected depreciation could reduce the equipment's resale value.

Operating expenses: Electricity, cooling, networking, maintenance, staffing and facility costs can affect margins. HIVE has not disclosed a complete project-level operating-cost breakdown in the financing announcement.

AI cloud competition: Other providers may compete through pricing, service quality, hardware availability and infrastructure scale. Lower prices or weaker utilization could reduce the project's returns.

Financing and refinancing risk: HIVE has additional GPU cloud agreements and may need further capital to fulfil them. The availability and cost of future financing remain uncertain.

These risks do not establish that the project will underperform. They identify the factors that will determine whether the financing and customer contract deliver the expected results.

8. HIVE's Wider AI Expansion Strategy

HIVE says it has signed approximately $600 million in total contract value across three GPU cloud agreements during 2026.

That figure suggests the Cohere project is part of a broader effort to expand BUZZ HPC's AI infrastructure business. It should not be interpreted as $600 million in revenue already earned or cash already collected.

The distinction between signed contract value, recognized revenue, operating cash flow and profit is particularly important for infrastructure businesses that require substantial upfront capital expenditure.

The company's strategy combines digital-asset mining with AI infrastructure and high-performance computing. Over time, the relative contribution of each business will depend on deployment progress, customer demand, hardware utilization, energy costs and the economics of Bitcoin mining.

The Cohere cluster is therefore an important execution milestone: it can demonstrate whether HIVE can convert a large contract into operating infrastructure and cash flow while meeting its debt obligations.

9. What Happens Next?

Several developments will help investors and industry observers assess the project:

1. GPU delivery: Whether the NVIDIA GB200 systems arrive within the expected late-2026 window.

2. Deployment and commissioning: When the cluster becomes operational and begins serving Cohere.

3. Revenue realization: Whether services are delivered and contract revenue begins according to the expected schedule.

4. Debt servicing: Whether the project generates sufficient cash flow to meet scheduled principal, interest and covenant requirements.

5. Operating performance: The cluster's utilization, power costs, service reliability and margins.

6. Further AI contracts: Whether HIVE can finance and deploy additional GPU cloud agreements on commercially sustainable terms.

The results of these milestones will provide a clearer view of whether the company can scale AI infrastructure alongside its established digital-asset mining operations.

Final Takeaway

HIVE Digital Technologies' $129 million financing marks another step in the shift from cryptocurrency mining toward contract-backed AI computing infrastructure. The facility provides approximately 80% of the stated cost of a 2,303-GPU NVIDIA GB200 cluster intended for Cohere under a reported $225 million, three-year services contract.

The fixed 9% interest rate and hardware-backed structure provide a defined financing framework, but they do not guarantee profitability. Delivery timing, customer performance, operating costs, debt servicing and GPU depreciation will determine the project's ultimate financial outcome.

For the broader industry, the deal demonstrates how digital-infrastructure companies are trying to turn power, data-centre expertise and access to capital into AI computing capacity. The next test will be execution: bringing the cluster online and converting the contracted demand into sustainable operating cash flow.

Disclaimer: This article is for informational and educational purposes only and does not constitute financial, investment or trading advice. Coinccino does not provide buy or sell signals. Company announcements and forward-looking statements may not reflect actual future results. Readers should review official disclosures and conduct independent research before making investment decisions.