Antimatter, a newly formed infrastructure company created through the merger of three existing firms, has announced its launch as what it describes as the world's first vertically integrated neocloud built specifically for AI inference workloads.

The company plans to establish its global headquarters in Hong Kong and is moving rapidly to secure financing and deploy hardware across the United States, Europe, and the Gulf Cooperation Council region.

Three Companies Become One

Antimatter was formed through the strategic combination of Datafactory, a US-based energy and power infrastructure company; Policloud, which operates a modular micro data center network; and Hivenet, a distributed cloud provider.

The company announced its launch on May 4, 2026, stating that the merged entity creates what it calls the industry's first fully integrated AI infrastructure platform spanning energy sourcing, physical hardware, and cloud software.

The company is led by David Gurlé, who previously founded Microsoft's Real-Time Communications business, which became Microsoft Teams, led Skype's enterprise division through its sale to Microsoft, and founded Symphony Communication Services. Gurlé serves as cofounder, executive chairman, and chief executive officer of the new entity.

"In the age of AI, intelligence is not the bottleneck — energy is," Gurlé said in the company's announcement. "The infrastructure built for the first era of cloud and AI was designed around centralized scale. But the inference era requires a different model: more distributed, faster to deploy, and sovereign by design."

Moving AI from the Cloud to the Real World

Antimatter's founding rationale rests on the argument that the shift from AI model training to AI inference fundamentally changes the infrastructure requirements of the industry.

Training, in Antimatter's framing, was concentrated in large centralized data centers. Inference — running trained models billions of times per day to power applications such as copilots, agents, and real-time decision systems — demands something different: infrastructure closer to users, faster to deploy, and distributed across geographies.

The company contends that traditional hyperscalers, built around massive centralized campuses that can take years to construct and require enormous upfront capital, are poorly suited to meet this demand.

Antimatter's stated approach inverts the conventional model by bringing data centers to existing energy sources rather than routing energy to centralized locations.

The company cites market data indicating the global data center capacity market is projected to grow from 55 gigawatts in 2023 to 220 gigawatts by 2030, a compound annual growth rate of 22 percent.

It also notes that more than 12 terawatt-hours of renewable electricity were curtailed in Europe alone in 2023, representing over USD 4.92 billion in lost value, and that more than 1,000 gigawatts of additional renewable capacity remain stuck in permitting and grid-connection queues across Europe and the GCC (Gulf Cooperation Council).

Infrastructure Architecture and Deployment Plans

The core physical unit of Antimatter's network is the Policloud, a modular containerized micro data center housing up to 400 GPUs. The company says each unit can be deployed in as little as five months, compared with 24 or more months for a traditional hyperscale build.

Antimatter currently operates 17 units across 8 sites and says it has a commercial pipeline of more than 500 additional units.

Antimatter is securing USD 351 million to fund the deployment of its first 100 Policloud units in 2026, which it says will represent 40,000 GPUs and over 3.6 exaFLOPS of active compute capacity.

By the end of 2030, the company plans to have a network of 1,000 Policlouds providing more than 400,000 GPUs and over 36 exaFLOPS of distributed AI inference capacity, which it describes as the equivalent of five traditional hyperscale data centers deployed across dozens of countries.

The company has secured more than one gigawatt of power capacity across distributed micro-power sites, including over 160 megawatts already operational in Texas and Oregon.

Policloud units are deployed directly at or near existing power assets, including wind, solar, hydro, and biogas sites, converting what Antimatter describes as stranded generation into productive AI infrastructure.

Connecting the distributed hardware is a proprietary software layer that the company describes as a distributed computing and storage platform. Antimatter says this platform provides orchestration intelligence that links individual units into a single cloud fabric with global default Tier 3 capability, supporting sub-10ms latency for edge workloads and full data sovereignty for regulated industries.

Cost and Performance Claims Against Hyperscalers

Antimatter makes a series of direct cost comparisons against traditional hyperscale providers. The company states its capital expenditure per fully loaded megawatt runs approximately USD 7 million, compared with roughly USD 35 million for a traditional hyperscale facility. It says customer pricing runs approximately 50 percent below hyperscaler market rates.

Additional claims include approximately 70 percent lower carbon emissions, zero water cooling requirements, and a sovereign-by-design data architecture that keeps data under local jurisdiction rather than relying on what the company characterizes as bolt-on compliance solutions from existing providers.

Revenue and Commercial Position at Launch

The company enters the market, describing itself as cash-flow positive. At launch, it reports USD 20 million in current annual revenue and 4 million dollars in earnings before interest and taxes. It has 4,500 GPUs deployed, with demand, it states, exceeding 10,000 units.

Antimatter's existing customer base is described as diversified across sectors, with the energy sector accounting for 35 percent of business, the public sector at 30 percent, corporates at 20 percent, and agriculture at 15 percent.

The company has set a revenue target of more than 250 million dollars within the next 18 months and states it is targeting more than 2.5 billion dollars in revenue by the end of 2030.

Asia-Pacific Data Center Pipelines Are Moving Fast — Is Your Intelligence Keeping Up?

The Asia-Pacific data center sector is expanding at a pace that leaves little room for guesswork. From hyperscale campuses breaking ground in Singapore and Malaysia to edge deployments pushing into emerging markets across Southeast Asia and beyond, the window to act on the right opportunities is narrowing by the quarter.

Staying competitive in this environment demands more than periodic market reports. You need a continuous, structured view of every project moving through the development lifecycle — from early planning signals to completed facilities — across every key Asia-Pacific market simultaneously.

The Global Project Tracking (GPT) platform by Blackridge Research gives data center developers, contractors, equipment suppliers, and investors exactly that visibility. Built to surface actionable intelligence at every stage of a project's journey, the platform ensures you are never caught off guard by a contract awarded or a tender closed before you knew it existed.

• Upcoming Projects

• Tender Notices

• Contract Awards

• Projects Under Construction

• Completed Projects

To see how the Global Project Tracking (GPT) platform by Blackridge Research can sharpen your Asia-Pacific data center strategy, contact our team today and Book a Free Demo.