Dream, a cybersecurity and sovereign AI company, just raised $260 million at a $3 billion valuation, signaling a major shift in how nations approach artificial intelligence. The funding reflects growing urgency among governments worldwide to build AI capabilities they control entirely, rather than depending on models and infrastructure owned by foreign companies.

Founded in 2023 by Shalev Hulio, former Austrian Chancellor Sebastian Kurz, and Gil Dolev, Dream has already secured nearly $300 million in total contract value from national governments across Europe, the Middle East, and Asia in just over three years. The company now employs approximately 350 people and has achieved zero customer churn, a remarkable retention rate for government technology contracts.

Why Are Governments Suddenly Treating AI as Critical Infrastructure?

The shift reflects a fundamental realization: artificial intelligence is becoming as strategically important as roads, power grids, and defense systems. Yet most governments rely on AI models built by foreign companies, hosted on infrastructure they don't control, and subject to restrictions or shutdowns beyond their authority.

Recent geopolitical events have made this vulnerability tangible. When the U.S. government ordered Anthropic to suspend access to its Claude Fable 5 and Claude Mythos 5 models for foreign nationals in June 2026, the company disabled both models globally because it couldn't reliably separate permissible from impermissible users across all access channels. For one weekend, the world saw that access to advanced AI systems can be interrupted not only at the chip or data-center level, but at the model-access level itself.

"When a government depends on foreign models it doesn't host, it inherits a dependency. Models can be restricted, throttled, or switched off. A nation's most sensitive data ends up on infrastructure it doesn't control. That's a national security risk whether the use case is intelligence, healthcare, finance, or critical infrastructure," said Shalev Hulio, Co-Founder and CEO of Dream.

Shalev Hulio, Co-Founder and CEO of Dream

Dream's founders recognized this gap early. While much of the AI industry focused on building and selling frontier models, Dream saw a different problem: governments needed to adopt AI safely without becoming dependent on systems they couldn't control.

What Does Dream Actually Build for Governments?

Dream operates three interconnected platforms designed to give nations sovereign control over their AI infrastructure:

• Sphere: A unified national cyber defense system that combines cyber intelligence, exposure management, attack path analysis, digital twin technology, and AI-powered detection and response to protect governments against nation-state threats.
• Hero: An autonomous AI security researcher that discovers vulnerabilities, identifies attack paths, and reasons like an adversary at machine speed to prevent sophisticated cyber threats before attackers exploit them.
• Atlas: Dream's sovereign AI platform that enables governments to connect fragmented national data, transform information into structured knowledge, deploy mission-specific AI agents and models, and generate actionable insights entirely within secure government-controlled environments.

What makes Dream's approach distinctive is its use of specialized small language models (SLMs) trained on vast security datasets and deployed fully on-premise, including in air-gapped environments where systems are completely isolated from the internet. These purpose-built models have outperformed leading frontier models on cyber-specific evaluations and achieve superior token efficiency and latency, which is critical for real-time threat detection at nation-scale.

Dream's methodology is straightforward: think like an attacker. Rather than relying on compliance checklists and alert floods, the platform simulates how a nation-state adversary would breach an environment and neutralizes those paths before they're exploited. Customers have already reported discovering risks they had never previously been able to map.

How to Evaluate Sovereign AI Capabilities for National Security

• Data Sovereignty: Verify that all sensitive government data remains on infrastructure the nation controls, not stored on foreign servers or cloud platforms subject to external restrictions or access requests.
• Model Ownership: Ensure the government owns or has permanent, irrevocable access to the AI models themselves, not just a license that can be revoked or modified by a foreign company.
• Operational Independence: Confirm that critical AI systems can operate continuously even if external services are disrupted, restricted, or withdrawn due to geopolitical events or policy changes.
• Threat Detection Speed: Assess how quickly the AI system can identify and respond to cyber threats; Dream's platform identifies vulnerabilities in minutes rather than months, a critical advantage against nation-state adversaries.

Dream's traction tells the story of a category being created in real time. What began as expansion from a handful of local sovereign customers has now expanded to disparate use cases across multiple governments, with the company on track to triple its footprint across sovereign clients in the last 18 months. The company has secured nearly $300 million in total contract value since beginning commercial operations in late 2024.

The pace of adoption has challenged common assumptions about how governments buy technology. Decision-making that typically takes months or years in government procurement has compressed dramatically when national security and sovereignty are at stake.

"The common assumption is that governments are slow and bureaucratic. What we've seen is the opposite. When the issue touches national security, resilience, or sovereignty, decisions happen fast. Some of the gaps we discovered surprised us. The speed with which leaders moved to close them surprised us even more," said Shalev Hulio.

Shalev Hulio, Co-Founder and CEO of Dream

What Does This Mean for the Global AI Landscape?

Dream's expansion signals a broader geopolitical realignment in artificial intelligence. For years, the AI competition focused primarily on chips and computing power. Export controls targeted semiconductor manufacturing and cloud infrastructure. The Anthropic shutdown and Dream's rapid growth suggest the next frontier of AI governance will operate at the model-access and platform level.

Sebastian Kurz, President and Co-Founder of Dream and former Chancellor of Austria, emphasized the stakes: "The defining question for governments is no longer whether they will use AI, but whether they will own it. Nations that want to control their future need the ability to operate advanced AI under their own authority, on infrastructure they govern, and in alignment with their own interests".

Sebastian Kurz, President and Co-Founder of Dream and former Chancellor of Austria

The new capital will accelerate Dream's expansion beyond Europe, the Middle East, and Asia into the Americas, and fund strategic acquisitions that accelerate the broader sovereign AI stack. Dream has raised $412 million total to build the sovereign AI infrastructure governments fully own, control, and operate.

For nations watching geopolitical tensions rise and AI capabilities become more strategically important, Dream's growth reflects a hard truth: dependence on foreign AI systems is increasingly seen as a national security vulnerability, not a business convenience. The race to build sovereign AI infrastructure has begun, and governments are moving with unprecedented speed to close the gap.