Many investors assume that Shark Tank personality Kevin O'Leary, widely known as "Mr. Wonderful," is placing his biggest bets on artificial intelligence itself. In reality, his focus is on the more than $5 trillion worth of infrastructure required to power the AI revolution, where much of today's institutional capital is being deployed. One of his key bets is on Bitzero Holdings (AIBZ), a company aiming to address one of AI's most significant constraints: access to electricity.

Companies referenced in this analysis include Bitzero Holdings Inc. (AIBZ), Microsoft Corporation, NVIDIA Corporation, International Business Machines Corporation, Digital Realty Trust, Inc., and Quanta Services, Inc..

While much of the market remained concentrated on AI software and semiconductor companies, Bitzero took a longer-term view. On May 5, the company signed a binding agreement for a 15-year AI power lease, marking its transition from low-carbon Bitcoin mining into supplying electricity and infrastructure to the rapidly expanding AI data center industry.

Major technology companies continue to pour unprecedented sums into AI infrastructure. Amazon is expected to spend approximately $200 billion in capital expenditures during 2026, much of it directed toward data centers. Microsoft's projected spending is close to $190 billion, while Alphabet is expected to invest a similar amount. Meta has outlined plans to commit roughly $600 billion to U.S. infrastructure through 2028. Combined capital expenditures from Amazon, Microsoft, Alphabet, and Meta could reach as much as $725 billion in 2026 alone, largely driven by AI-related investments in data centers, semiconductors, power systems, and supporting infrastructure. Meanwhile, McKinsey estimates that AI infrastructure spending could total $5.2 trillion over the course of this decade.

Growing Pressure on AI Infrastructure Development
A significant number of planned AI data center projects may never become operational because the necessary electrical capacity is unavailable when required. This dynamic creates opportunities for companies such as Bitzero that already possess access to large-scale power resources.

Demand for artificial intelligence computing capacity continues to accelerate, yet securing sufficient electricity is becoming increasingly difficult, time-consuming, and costly. More than 70% of grid interconnection applications are eventually withdrawn, with only a small percentage reaching completion. At the same time, global data center electricity consumption is forecast to approach 945 terawatt-hours by 2030—roughly equivalent to Japan's current annual electricity usage.

Although semiconductor availability was once viewed as the primary bottleneck for AI growth, power availability has increasingly emerged as the industry's most critical constraint. This shift positions energy-intensive businesses such as Bitzero favorably.

According to Mohammed Bakhashwain, founder and CEO of Bitzero Holdings, companies involved in Bitcoin mining secured power resources well before the recent surge in demand from AI and data centers, giving them a competitive advantage as electricity prices rise.

Positioning for the Next Era of Computing
Earlier this month, Bitzero completed engineering due diligence work for a project supporting up to 520 megawatts at its Kokemäki campus in Finland, with long-term expansion potential reaching one gigawatt. The initial 80-megawatt phase is scheduled to launch during the first half of 2027, with subsequent phases expected to add between 400 and 800 megawatts.

The company's operations in Norway are already functioning as a fully operational industrial platform. There, Bitzero mines Bitcoin using electricity that costs less than four cents per kilowatt-hour, allowing the company to generate revenue while simultaneously developing additional infrastructure.

At its Namsskogan site in Norway, a further 70 megawatts of capacity is expected to come online during the fourth quarter of 2026 as part of a broader 325-megawatt expansion strategy. This location also represents Bitzero's official entry into the AI infrastructure market.
On May 5, Bitzero signed a binding letter of intent with OneQode Networks for the full 110-megawatt capacity of its Namsskogan data center campus under a 15-year agreement supporting GPU-based AI workloads. The contract carries an estimated value of approximately $2.6 billion over its duration.

Under its traditional Bitcoin mining model, Bitzero generates revenue through cryptocurrency production using its own power assets. Under the new AI infrastructure arrangement, the company earns revenue by leasing both power capacity and infrastructure to OneQode, while the tenant assumes responsibility for the electricity costs associated with AI operations. This structure enables Bitzero to capture recurring infrastructure revenue without bearing the substantial operating expenses of AI computing workloads.

Company projections indicate that the Namsskogan facility could generate annual revenue between $176 million and $178 million at full utilization. Independent shareholder analysis has estimated annual net operating income of approximately $151 million based on an estimated operating margin of 85%.

The appeal of Bitzero's Nordic assets lies largely in their access to abundant, low-cost, low-carbon energy. Norway's electricity grid is predominantly hydroelectric, while Finland benefits from a diversified mix of hydroelectric, nuclear, solar, and wind power.

Additionally, Bitzero's North Dakota operations provide exposure to the U.S. market, offering access to a different regulatory and pricing environment than its Scandinavian assets.

AI Growth Is Outpacing Power Development
Constructing large-scale electrical infrastructure to support data centers can take as long as seven years. Yet many AI infrastructure forecasts have implicitly assumed that sufficient power capacity will become available when needed. This assumption increasingly clashes with the realities of electricity generation and transmission development.

Securing power involves complex processes, including grid impact studies, transmission access negotiations, permitting, utility agreements, and long-term pricing arrangements. Meanwhile, demand continues to accelerate.

The International Energy Agency projects that electricity consumption from data centers will grow approximately four times faster than electricity demand from all other sectors combined, potentially reaching around 945 terawatt-hours annually by 2030. Similarly, Goldman Sachs forecasts a 175% increase in data center electricity demand by 2030 compared with 2023 levels.

Despite this demand surge, investment in electrical infrastructure remains insufficient. Estimates suggest that approximately $6.7 trillion in capital will be required by 2030, including $5.2 trillion dedicated specifically to AI infrastructure. Current projections indicate that only about $720 billion in grid investments are presently planned.

Why Investors Are Watching Bitzero
Bitzero is developing large-scale campuses supported by secured, low-cost electricity and positioning those assets to serve AI and high-performance computing markets. Rather than choosing between cryptocurrency mining and AI infrastructure, the company intends to participate in both sectors simultaneously.

Bitcoin mining operations provide immediate revenue generation, while the same facilities are being adapted to support higher-value AI and high-performance computing workloads over time.

CEO Mohammed Bakhashwain has highlighted the company's opportunities in high-performance computing, noting that its engineering teams have previously worked on deployments involving Microsoft and Nscale in Norway. By controlling land, power resources, and infrastructure assets, Bitzero aims to attract AI tenants seeking large-scale computing capacity.

This strategy enables the company to generate revenue from existing Bitcoin operations while positioning its infrastructure for potentially more lucrative long-term AI and colocation contracts. By maintaining flexibility, Bitzero seeks to allocate capacity to whichever market offers the strongest economics at a given time.

Other Companies Worth Monitoring
Microsoft Corporation is undertaking the largest infrastructure expansion in its history, committing approximately $80 billion to AI-enabled data centers during fiscal year 2025. More than half of that investment is targeted toward facilities in the United States, including a new AI campus in Wisconsin expected to cost more than $7 billion and support hundreds of thousands of NVIDIA GPUs.

The company's financial performance reflects this investment strategy. In the first quarter of fiscal year 2026, Microsoft generated $77.7 billion in revenue, an 18% year-over-year increase, while Azure and related cloud services grew by 40%. Its commercial backlog expanded to $392 billion, up 51%.

NVIDIA Corporation remains central to the AI infrastructure ecosystem. The company reported record first-quarter fiscal year 2027 revenue of $81.6 billion, representing 85% annual growth, while data center revenue increased 92% to $75.2 billion.

Growth is increasingly being driven by agentic AI applications, which require substantially more computing power than traditional conversational AI systems. Although hyperscale cloud providers account for slightly more than half of NVIDIA's data center revenue, enterprise demand continues to expand rapidly.

International Business Machines Corporation represents a less obvious but increasingly relevant participant in the AI infrastructure market. During the first quarter of 2026, IBM reported revenue of $15.9 billion, up 9% year over year, while software revenue increased 11% to $7.1 billion. Free cash flow reached $2.2 billion, marking the company's strongest first quarter in a decade.

IBM's AI strategy centers around Red Hat OpenShift and its watsonx platform, which enterprises use to deploy and manage AI systems across hybrid environments. CEO Arvind Krishna has emphasized that AI demand is also benefiting IBM's mainframe business.

Digital Realty Trust, Inc. continues to play a major role in the hyperscale data center market. During the first quarter of 2026, the company signed the largest hyperscale lease in its history—a 200-megawatt agreement in Charlotte—and recorded annualized base-rent bookings of $707 million.

The nature of AI-related leases has evolved significantly, with tenants increasingly committing to larger projects, longer contract durations, and earlier pre-leasing arrangements.

Quanta Services, Inc. occupies a critical position in the AI infrastructure ecosystem by constructing and maintaining the electrical networks that connect data centers to the grid. CEO Duke Austin has estimated the company's addressable market opportunity at approximately $2.4 trillion through 2030, driven by data center expansion, grid modernization, and renewable energy integration.

The industry's challenges remain fundamentally physical: high-voltage transformers often require lead times exceeding two years, and skilled electrical labor remains in short supply across many markets.

Hyperscale Data, Inc., an AI-focused data center operator with roots in Bitcoin infrastructure, has announced a new Master Services Agreement (MSA) through its indirect wholly owned subsidiary, Alliance Cloud Services, LLC (ACS). Under the agreement, ACS will deliver colocation and data center support services to a California-based neocloud company at the firm's Michigan data center campus.

The customer specializes in cloud and managed services, offering customized high-performance computing resources and scalable storage solutions in partnership with leading technology providers.

The agreement calls for the deployment of 20 megawatts (MW) of AI computing capacity, which is expected to become operational during the fourth quarter of 2026. The customer also has the option to expand its footprint to as much as 52 MW of AI compute capacity. The contract carries an initial term of 10 years, with two additional five-year renewal options available. If the customer exercises the full 20-year term, the agreement is projected to generate more than $1.2 billion in revenue.

In addition, the MSA grants the customer the right to secure an extra 32 MW of AI computing capacity. Should this expansion be activated within the first two years and remain in effect through both renewal periods, total contract revenue could exceed $3 billion.

To support the initial deployment, ACS has begun acquiring critical electrical and infrastructure components and is retrofitting approximately 60,000 square feet of space at the Michigan campus. The estimated investment for preparing the site for the first 20 MW phase ranges between $100 million and $120 million.

As AI infrastructure is brought online and customer workloads increase, Hyperscale Data plans to gradually redirect power resources currently dedicated to Bitcoin mining at the Michigan campus. The company expects to continue Bitcoin mining operations at its Montana facility and may maintain limited mining activity in Michigan during the transition.

Chief Executive Officer William B. Horne stated that the project marks another step in transforming the Michigan facility from a Bitcoin mining operation into a modern AI and high-performance computing hub. He noted that the site is being positioned to deliver premium AI computing services and could begin generating substantial, high-margin revenue as early as late September 2026.

Executive Chairman Milton "Todd" Ault III described the agreement as a major achievement for the company. He explained that while the Michigan campus currently supports roughly 28 MW of Bitcoin mining capacity, an increasing share of the site's resources will be dedicated to AI and advanced computing workloads as customer deployments ramp up. He added that the strategy is designed to maximize the long-term value of the campus while advancing plans to develop more than 300 MW of total power capacity.

Hyperscale Data believes the Michigan campus offers significant future growth potential through phased expansion initiatives. However, any additional development remains subject to factors such as regulatory approvals, financing availability, infrastructure readiness, engineering assessments, and utility agreements. The company estimates that the site could ultimately support more than 300 MW of power capacity.

Management emphasized that these expansion plans are still in the preliminary stages and face various risks and uncertainties. As a result, there is no guarantee that the proposed capacity increases will be approved, financed, constructed, or prove economically viable.

The company encourages investors and stakeholders to review its public filings, investor updates, and official announcements for additional information regarding Hyperscale Data and its subsidiaries.

Exyte, a global specialist in the design, engineering, and construction of advanced technology facilities, has been awarded three significant data center projects in the Greater Frankfurt area. Secured from leading hyperscale operators and major technology firms, the contracts have a combined value of nearly €750 million, recorded as order intake for 2026. The projects are also expected to create approximately 400 jobs across Germany.

“These new awards highlight the growing importance of the data center sector within Exyte’s portfolio and reinforce its position as a major engine of growth,” said Mark Garvey, Chief Commercial Officer at Exyte. “With demand for artificial intelligence and cloud-based services expanding rapidly worldwide, reliable and scalable digital infrastructure is becoming essential for maintaining our clients’ competitive advantage.”

The latest wins further strengthen Exyte’s reputation as a trusted partner for global technology companies and build upon its extensive experience delivering data center developments in the Frankfurt market.

The three contracts encompass both the construction of new facilities and the expansion of existing sites, with all projects being executed simultaneously. Some of the work will be carried out within operational data centers, requiring meticulous planning and coordination to maintain uninterrupted service while increasing capacity. Together, the facilities will provide approximately 80 megawatts of IT load capacity and are scheduled for completion by 2028. Rapid and dependable delivery remains a critical priority as customers seek to expand infrastructure in response to rising demand.

Workforce Expansion to Support Delivery
To meet project requirements, Exyte plans to add around 400 positions in Germany. Approximately 150 roles will be filled through the redeployment of experienced personnel from other projects within the company, while the remaining 250 positions will be recruited externally.

Hiring efforts will span a variety of disciplines, including engineering, technical services, project controls, construction management, and commissioning. Employees will work within international, multidisciplinary teams on large-scale infrastructure developments, contributing to technologies that support the future growth of AI and digital services.

Advancing Digital Infrastructure for an AI-Driven Future
The Frankfurt projects are part of a wider surge in investment across Europe’s data center sector. Growing AI adoption, increased reliance on cloud computing, and the need for secure and resilient digital ecosystems are driving demand for large, high-density facilities.

“Data centers are now a cornerstone of modern economic infrastructure,” said Jürgen Raschendorfer, Chief Operating Officer for Exyte’s Continental Europe region. “These projects expand our established presence in Frankfurt and help clients increase capacity, improve resilience, and deliver next-generation digital services in one of Europe’s most important data center hubs.”

Focus on Delivery Excellence
Exyte attributes its success to a delivery model that combines advanced engineering expertise with highly standardized and industrialized construction methods.

“What sets Exyte apart is our ability to execute large-scale data center programs with confidence in cost, schedule, and quality outcomes,” said Damian Farr, President of Exyte’s Global Business Unit Data Centers. “This capability is especially important in live operational environments, where careful coordination and integrated execution are essential to avoid disruptions.”

Modern data center designs are also evolving to support higher power densities and sophisticated cooling technologies needed for AI-driven workloads. Exyte leverages offsite manufacturing and prefabrication techniques to accelerate project delivery, improve quality standards, and reduce construction risks.

To meet increasing energy requirements, the company incorporates flexible power infrastructure and resilient backup systems that enhance operational reliability. Digitalized delivery and commissioning processes provide greater predictability across project timelines, costs, and quality metrics. In addition, Exyte’s established supply chain network and category management strategy help secure access to critical equipment and specialist partners despite ongoing market constraints.

These capabilities enable customers to bring facilities online more quickly, achieve greater operational reliability, and expand digital infrastructure with confidence.
 

New research commissioned by Lenovo highlights the urgent need for data center design across EMEA to adapt in order to support long-term business resilience. As organizations seek to run AI workloads more efficiently while responding to tightening sustainability regulations and compliance requirements, legacy data center models are increasingly proving inadequate. Nearly half of IT leaders (45%) acknowledge that their existing infrastructure does not align with their energy efficiency or carbon reduction targets.

Concerns around data governance are equally pressing. Almost all IT and executive decision-makers in the region (99%) say data sovereignty will play a critical role in how data is gathered, stored, and processed in the coming years. Meanwhile, the rapid acceleration of AI-driven data consumption is exposing a disconnect: many organizations remain ill-equipped to deploy AI effectively or in an environmentally responsible way, underscoring a widening gap between digital aspirations and infrastructure readiness.

To explore these challenges, Lenovo partnered with Opinium on the Data Center of the Future study, which outlines the major forces shaping where and how data centers will be designed and operated. This research comes at a time when the data center market continues to expand, while energy consumption, sustainability, and cost pressures grow ever more central to IT decision-making across EMEA.

Sustainability readiness demands a rethink
Although 92% of IT decision-makers say they favor technology partners that actively reduce energy use and carbon emissions, fewer than half (46%) believe their current data center architecture supports sustainability objectives. This shortfall reflects the mounting environmental strain caused by AI, automation, and surging data volumes. Traditional air-cooling systems, in particular, are struggling to deliver the right balance of efficiency, affordability, and emissions reduction.

Data sovereignty remains a top priority
With 88% of IT leaders already treating data sovereignty as a key concern—and nearly all expecting it to remain critical over the next five years—control over data location and compliance is set to strongly influence future data center strategies. In parallel, 94% cite low latency as an essential requirement now and in the future, driven by the rise of real-time workloads and edge computing.

Scaling AI will shape tomorrow’s data centers
Looking ahead, 90% of IT decision-makers anticipate that AI will dramatically increase organizational data volumes over the next decade, while 62% see AI and automation as the most influential factors in shaping IT strategy. Despite this, readiness remains uneven: 41% admit their organizations are not yet prepared to integrate AI in an efficient manner.

“The next generation of data centers will be defined by their ability to scale AI workloads, meet sustainability commitments, and maximize energy efficiency,” said Simone Larsson, Head of Enterprise AI, EMEA at Lenovo. “As compute demand accelerates, customers will increasingly seek infrastructure partners that can deliver high performance while also taking accountability for environmental impact.

“In EMEA, data sovereignty is especially urgent due to complex regulatory landscapes and growing scrutiny from CIOs and executive teams. Organizations must make proactive infrastructure choices now, because the decisions made today will determine future readiness.”

Envisioning the Data Center of 2055
To explore what data centers could look like three decades from now, Lenovo collaborated with engineering firm AKT II and architects Mamou-Mani. Their concepts reimagine the traditional rack-based data center by integrating liquid cooling, natural resources, and unconventional locations to improve sustainability and reduce environmental impact. Proposed designs include:

All concepts rely on liquid cooling to address rising heat densities and the inefficiencies of air-based cooling. Compared to traditional methods, liquid cooling consumes less energy and significantly improves overall sustainability.

“As architects and engineers, our goal is to make data centers smarter and more responsible—not simply larger,” said James Cheung, Partner at Mamou-Mani. “This project brings together business needs and realistic solutions, from repurposed underground spaces to urban data villages and high-altitude cloud modules that return energy to local communities. While no one can predict the future with certainty, these concepts demonstrate ideas that could move from theory to real-world pilots with reduced risk.”

Building future-ready infrastructure today
To address rising compute demands alongside stricter sustainability expectations, organizations must begin modernizing their infrastructure now. Liquid cooling represents a practical and effective pathway forward. Lenovo Neptune liquid cooling technology, for example, can remove up to 98% of system heat directly at the source, significantly lowering energy consumption and reducing dependence on air cooling.

As AI adoption and advanced analytics continue to scale, Neptune solutions provide a foundation that supports both performance growth and environmental responsibility.

“Lenovo is focused on delivering intelligent, sustainable infrastructure at scale,” Larsson concluded. “With Neptune liquid cooling, we are already enabling customers to manage the energy intensity of AI through highly efficient, deployable solutions. Designing future-ready data centers requires a fundamental shift—one where sustainability is embedded from the outset, not added later.”