Antimatter Launches to Tackle AI's Energy Crisis with New 'Neocloud'

CANNES, France – April 21, 2026 – A new company, Antimatter, has launched with a bold plan to rewire the physical infrastructure of artificial intelligence. Led by serial tech entrepreneur David Gurlé, the venture emerged today through the strategic merger of three specialized firms, creating what it calls the world's first vertically integrated "neocloud" for AI inference. Antimatter claims its distributed network of micro data centers can deliver AI computing power five times faster and 50% cheaper than the hyperscale giants that currently dominate the cloud market.

Antimatter is the culmination of Datafactory, a US-based energy infrastructure firm; Policloud, a modular micro data center network; and Hivenet, a distributed cloud provider. This combination creates a single entity that controls the entire stack, from securing power at the source to managing the software that runs AI models. The company is entering a market grappling with the voracious energy appetite of AI, a problem that threatens to stall the technology's rapid expansion.

"In the age of AI, intelligence is not the bottleneck — energy is," said David Gurlé, who takes the helm as Cofounder, Executive Chairman, and CEO of Antimatter. "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. That is the infrastructure Antimatter is building."

The Hyperscale Bottleneck

The explosive growth of artificial intelligence has created an unprecedented demand for computing power, but its Achilles' heel is quickly becoming apparent: access to electrical power. The first wave of AI focused on training massive models inside colossal, centralized data centers owned by hyperscalers like Amazon, Google, and Microsoft. The next phase, however, is dominated by inference—the process of running those trained models billions of times a day to power applications like chatbots, copilots, and real-time analytics.

This shift is straining the traditional data center model to its breaking point. Industry analysts confirm that the single biggest constraint on building new AI facilities is no longer land or capital, but the ability to connect to the power grid. In the United States alone, nearly 2,300 gigawatts of power generation and storage projects are stuck in interconnection queues, with wait times stretching for years. In Europe, over 12 terawatt-hours of renewable electricity were wasted in 2023 because the grid couldn't transport it to where it was needed, representing billions in lost value.

Hyperscale data centers, which can take over two years to build, are struggling to keep pace. Antimatter’s core strategy is to invert this paradigm: instead of bringing massive amounts of energy to a data center, it brings the data center directly to the energy.

A Distributed, Energy-First Model

Antimatter’s solution is a global network of modular, containerized micro data centers called Policlouds. Each unit, housing up to 400 GPUs, can be deployed in as little as five months—a fraction of the time required for traditional builds. This rapid deployment is made possible by a vertically integrated model that begins with energy.

The company has already secured over 1 gigawatt of power capacity through grid agreements and site reservations across the US, Europe, and the GCC. It plans to co-locate its Policloud units at existing power generation sites, including wind, solar, and hydro plants, effectively turning stranded or curtailed energy into high-value AI compute capacity.

This decentralized physical infrastructure is unified by a proprietary software layer that orchestrates the distributed hardware into a single cloud fabric. The company claims this architecture can deliver sub-10ms latency for edge workloads, a critical requirement for real-time AI applications, while also providing full data sovereignty by keeping data within local jurisdictions—a key selling point for governments and regulated industries.

The cost advantages are significant. Antimatter projects its capital expenditure per megawatt is around $7 million, compared to an estimated $35 million for a traditional hyperscale facility. These savings, combined with operational efficiencies, are the foundation for its promise of 50% lower customer pricing.

A Serial Entrepreneur's Next Act

Leading this ambitious charge is David Gurlé, a French entrepreneur with a formidable track record of building disruptive technology companies. Gurlé is best known for founding and leading Microsoft's Real-Time Communications business, the precursor to Microsoft Teams, and later founding Symphony, a secure communication platform that grew to a $1.4 billion valuation by serving the world's top financial institutions. His experience building secure, distributed systems at massive scale provides a credible foundation for Antimatter's vision.

His career, which also includes leadership roles at Skype and Thomson Reuters, demonstrates a pattern of identifying and capitalizing on major technological shifts. With Antimatter, he is betting that the future of AI infrastructure is not bigger, but smarter and more distributed. The merger of his more recent ventures, Hivenet and Policloud, with Datafactory's energy expertise forms the strategic core of this new company.

Ambitious Goals and Investor Confidence

Antimatter is launching with significant momentum and aggressive growth targets. The company is in the process of securing €300 million in funding to deploy its first 100 Policloud units by 2027, which will represent 40,000 GPUs and 3.6 exaFLOPS of computing power. By the end of 2030, it plans to operate a network of 1,000 units with over 400,000 GPUs, projecting revenues to exceed $3 billion.

Early investors are backing the vision, seeing the model as a solution to critical industry-wide problems. "AI infrastructure is now a strategic asset class, and the winners will be those who can combine hard assets with software at scale," said Alex Manson, CEO of SC Ventures at Standard Chartered Bank. "Antimatter's vertically integrated model... is exactly the kind of infrastructure we believe can define the next decade of digital growth."

Other backers highlighted the model's alignment with regional needs for sovereign and sustainable AI. Stéphanie Hospital, Founder and CEO of OneRagtime, noted the company's ability to "deploy micro data centers in months, on existing power assets, while meeting the most demanding regulatory constraints." This sentiment was echoed by investors focused on emerging markets and European deep tech, who see the distributed, capital-efficient approach as a way to leapfrog legacy infrastructure.

Beyond its speed and cost, Antimatter is also promoting a greener approach to the AI boom. The company claims its model results in a 70% lower carbon footprint and uses zero water for cooling—a significant claim as data center water consumption becomes a major environmental concern. Its use of closed-loop liquid cooling and placement near renewable energy sources are key components of this sustainable design, addressing a critical challenge for an industry whose energy consumption is projected to skyrocket.