返回新闻
工业AI Understanding 简报

Ault & Company 投资 5500 万美元建设 Hyperscale Data 密歇根人工智能数据中心

Ault & Company 通过可转换优先股向 Hyperscale Data 投资 5500 万美元,获得 62% 的实益所有权股份,该公司准备将前采矿场转变为以人工智能为中心的计算能力。

4 min readRead the linked source
Source-provided image accompanying Ault & Company invests $55 million in Hyperscale Data’s Michigan AI data center
来源参考来源记录
出版商
remio.ai
来源链接
remio.aihttps://www.remio.ai/post/ault-hyperscale-data-investment-deepens-its-ai-bet-but-the-contract-still-has-to
来源类型
链接来源——主要来源状态尚未确定。
背景60 秒内了解这一点

从这里开始

关键术语

计算
训练和运行模型所需的处理资源,通常以 FLOPS 或 GPU 小时来衡量。
延迟
发送请求和接收模型输出之间的时间。
测试一下自己人工智能测验的未来

发生了什么

Ault & Company announced a $55 million investment in Hyperscale Data, delivered through multiple purchases of convertible preferred stock. The transaction gives Ault roughly a 62% beneficial ownership position, combining common shares, preferred securities, warrants and Class B shares. The capital is earmarked for converting a Michigan campus—previously used for cryptocurrency mining—into a high‑performance computing facility for AI workloads. Hyperscale Data has signed a 20‑megawatt master services agreement with an unnamed California‑based neocloud provider, with the deployment slated for the fourth quarter of 2026 and a ten‑year term that can be extended to 20 years. The agreement includes an optional 32‑megawatt expansion, which could raise total contracted capacity to 52 megawatts and generate up to $3 billion in revenue over the maximum term, according to management estimates.

On September 25 2026, Ault & Company disclosed that it had invested approximately $55 million directly into Hyperscale Data, using a series of convertible preferred stock purchases (Series C, G and H). The filing valued the Series C purchase at $50 million, with smaller amounts allocated to Series G ($960 k) and Series H ($4 million).

The investment brings Ault’s beneficial ownership to roughly 62 % of Hyperscale Data, though its voting power remains around 14 %. The ownership calculation includes common shares, preferred securities, warrants and Class B shares, reflecting a broader economic exposure than the outright share count.

Hyperscale Data’s Michigan campus, previously a cryptocurrency mining and colocation site, is being re‑engineered for AI . Alliance Cloud Services, a wholly‑owned subsidiary, has already invested more than $70 million in the facility’s power and cooling upgrades.

A master services agreement with an unnamed neocloud provider covers 20 megawatts of AI‑focused GPU capacity, with a ten‑year term and two five‑year extension options. The contract also grants the customer a right to add another 32 megawatts, potentially raising total contracted capacity to 52 megawatts.

来源详情: remio.ai ↗

为什么这很重要

The investment underscores the growing appetite for dedicated AI capacity and highlights how legacy, energy‑intensive sites are being repurposed for AI workloads. By securing a large, insider‑controlled stake, Ault provides Hyperscale Data with patient capital that could accelerate construction, power upgrades, and cooling infrastructure—critical hurdles for AI‑grade data centers. If the 20‑megawatt deployment proceeds on schedule, it will serve as a tangible proof point that the company can deliver AI hosting services, potentially validating management’s valuation range of $750 million to $1.25 billion for the campus. Conversely, delays or failure to attract the customer’s hardware could expose common shareholders to dilution from the convertible preferred securities and erode confidence in the broader AI infrastructure boom. The deal also illustrates how private capital is being funneled into AI‑specific projects, complementing larger, institutional funding streams and influencing competitive dynamics among data‑center operators.

The capital infusion directly addresses the high upfront costs of converting a mining‑grade site into an AI‑ready data center, a process that typically requires substantial power delivery, advanced cooling, and low‑ networking.

A successful launch would provide a concrete data point for the broader AI infrastructure market, demonstrating that repurposed sites can meet the stringent performance and reliability standards demanded by AI workloads.

Management’s revenue projections—up to $1.2 billion for the initial 20 megawatt deployment and $3 billion if the expansion option is exercised—are contingent on multiple downstream events, including the customer’s hardware deployment, financing of additional infrastructure, and sustained demand over a two‑decade horizon.

Interactive Mechanism

互动机制:它实际上是如何运作的

以交互方式探索这一发展背后的基础技术。

Model Parameter Size:8B Parameters
VRAM Required5.5 GBGPU memory footprint
Target HardwareMacBook / Single GPUDeployment tier
Privacy100% Air-GappedLocal device capability
Core takeaway: Small, quantized models (3B–8B) now run directly inside smartphones and laptops with complete data privacy, while mammoth 400B+ models remain the domain of datacenter clusters.
交互式概念检查+10 Points
Future of AI Quiz

What should a useful AI forecast state?

接下来看什么

Investors and observers should monitor three key milestones: (1) the actual energization and operational start of the 20‑megawatt capacity by Q4 2026; (2) the unnamed neocloud customer’s deployment of hardware and any public disclosure of its identity, which will affect credit assessment; and (3) the exercise of the 32‑megawatt expansion option, which would signal confidence in the initial phase and trigger additional financing needs. Additionally, the pending divestiture of Ault Capital Group in 2027 could reshape the ownership structure and clarify the company’s focus on AI data‑center assets.

Operational readiness: Confirmation that the Michigan campus receives power and completes construction by Q4 2026 will be the first tangible test of the investment’s premise.

Customer execution: Evidence that the unnamed neocloud provider actually installs GPU hardware and begins generating hosting revenue will validate the contract’s economic assumptions.

Expansion decision: Whether the customer exercises its 32‑megawatt option will indicate confidence in the initial phase and trigger a larger capital requirement for Hyperscale Data.

Financing and dilution: Future financing rounds, especially any conversion of the preferred securities into common equity, could dilute existing shareholders and affect the overall return profile.

相关指南和测验

AI 的未来人工智能代理人工智能模型解释测试你所知道的——尝试免费的人工智能测验在我们的词汇表中查找人工智能术语关注 AI 资金追踪器
觉得这有用吗?