1. **Sovereign AI Emerges as a National Imperative:** Shalev Hulio, controversial founder of NSO Group, is now spearheading Dream, a venture advocating for national governments to host their own AI models to ensure data sovereignty, mitigate geopolitical risks, and safeguard critical infrastructure from foreign influence and cyber threats.
2. **Market Shift Towards Bespoke, Secure AI Solutions:** Dream’s success in raising $400 million and achieving a $3 billion valuation signals a significant market demand for specialized, on-premise AI platforms designed for sensitive government applications, challenging the general-purpose offerings of hyperscale cloud providers.
3. **Cybersecurity and AI Converge:** The escalating sophistication of autonomous cyber attacks is driving the integration of AI into national defense strategies, creating a lucrative niche for firms like Dream that can offer advanced, AI-powered cyber defense and vulnerability management tools to states.
In an increasingly fractured global landscape, where technological supremacy and data sovereignty are paramount, a provocative argument is gaining traction: nations must seize control of their artificial intelligence infrastructure. This call comes from an unlikely, yet uniquely experienced, figure: Shalev Hulio, the entrepreneur who once co-founded NSO Group, the Israeli firm notorious for its Pegasus spyware. Now, through his new venture, Dream, Hulio asserts that governments should host their own AI models, rather than entrusting sensitive national data to the dominant US technology behemoths.
“Every government should have their own model,” Hulio told the Financial Times, underscoring a deepening concern among nation-states about digital independence. “They should not take the risk that somebody will wake up one day and shut their model because AI is a critical infrastructure. You want to control your own AI.” This sentiment resonates strongly in an era marked by escalating geopolitical tensions, data localization mandates, and a global race for AI leadership, pushing governments to reconsider their reliance on foreign-owned cloud and AI services.
Hulio’s proposition, while timely, arrives steeped in controversy. His legacy from NSO Group, the company behind the Pegasus tool used to hack phones globally, casts a long shadow. Yet, Hulio positions his past not as a liability, but as an invaluable asset. He argues that his deep understanding of state-level cyber capabilities, honed by years of providing surveillance tools to governments, uniquely qualifies him to build robust cyber defense and sovereign AI solutions. “Despite the public opinion, governments managed to prevent terror attacks and managed to save hundreds of thousands of lives along the years using NSO products,” he stated, demonstrating an unwavering conviction in his past contributions and a clear strategy to leverage his controversial “black book” of government contacts.
Dream, co-founded in 2022 with former Austrian Chancellor Sebastian Kurz, initially focused on cyber security, recognizing early the transformative potential of AI in defense. Hulio recalls facing skepticism from American VCs in late 2022 when he predicted autonomous cyber attacks would necessitate autonomous defenders. “I told them five years from now… the cyber attack is going to be completely autonomous and the only way to be ready is to build autonomous defenders,” he recounted. “They told me, ‘listen you’ve lost your mind’.” This initial market reticence, however, quickly evaporated as the pace of AI development and the sophistication of cyber threats accelerated beyond expectations. The market now increasingly recognizes the imperative for AI-driven defense mechanisms, a shift validated by Dream’s rapid growth.
Dream has since demonstrated remarkable market traction, raising over $400 million and achieving a staggering $3 billion valuation by June. This significant investment round includes prominent names like Bain Capital Ventures and Tau Capital, an Abu Dhabi-headquartered investment firm backed by Sheikh Tahnoon bin Zayed al-Nahyan, the UAE’s powerful national security adviser. Such backing from both Western venture capital and strategically significant Middle Eastern funds underscores the perceived value and geopolitical importance of Dream’s offerings in protecting national digital assets.
The company’s product suite is tailored for state-level defense and intelligence. Sphere, its flagship, is marketed as a national-scale cyber defense system, leveraging AI to protect governments and critical infrastructure from state-sponsored attacks. Hero, a second product, autonomously identifies and patches software vulnerabilities, addressing a persistent and costly challenge for large organizations. The urgency of these solutions was starkly highlighted in August when Dream’s researchers observed what they described as the first “end-to-end autonomous” attack on a government target, compressing Hulio’s five-year prediction into a mere three and a half.
Beyond defense, Hulio and Kurz identified a parallel need for governments to better harness their own data. This led to the launch of Atlas in June, an AI platform designed to empower governments to develop and operate customized AI models on their *own* infrastructure. This direct challenge to the reigning cloud hyperscalers like Amazon, Microsoft, and Google is rooted in the principles of data sovereignty, security, and domain specificity. Unlike the broad, resource-intensive large language models (LLMs) such as OpenAI’s ChatGPT or Anthropic’s Claude, Dream’s Atlas focuses on bespoke, “narrow but very smart” models. These are trained exclusively on a government’s own classified data for specific tasks, like intelligence analysis or optimizing procurement, promising greater accuracy, lower operational costs, and complete control over sensitive information. “We are saying to entities within the government, you should have a very narrow but very smart model that can run very fast on one or two or three GPUs,” Hulio explained, emphasizing efficiency and security over expansive, general capabilities.
Hulio’s decade at NSO Group, which ended in 2022, was fraught with controversy, particularly after the FT revealed its exploitation of a WhatsApp vulnerability. NSO was eventually blacklisted by the US government and faced lawsuits. Yet, Hulio’s past engagements have seemingly gifted him with unparalleled access to top-tier government officials globally. “Governments really appreciate, I think, the relationship that we had before,” he noted, highlighting the unique trust dynamic he believes he brings to the table for highly sensitive national security projects. Dream, headquartered in Tel Aviv with offices in Vienna and Abu Dhabi, has rapidly expanded to 350 employees, having sold its cyber security tools to over a dozen countries across Europe, Asia, and the Middle East, with its sovereign AI product already deployed in two nations.
An early Dream investor aptly summarized the company’s unique blend of leadership: “Shalev has a fantastic black book,” adding that his “chutzpah” combined with Kurz’s government experience creates a “good combination.” This potent mix allows Dream to navigate the intricate geopolitical landscape of national security tech, offering solutions that promise not just technological superiority, but also the crucial elements of “trust and confidence” that Hulio believes are paramount when working with governments on their most critical digital assets.
Market Impact:
Dream’s ascent and its focus on sovereign AI and autonomous cyber defense represent a significant inflection point in the global technology market, particularly for government and critical infrastructure sectors. This model directly challenges the dominance of hyperscale cloud providers by demonstrating a viable, secure alternative for highly sensitive data and AI workloads. The market is likely to see increased fragmentation, with specialized vendors like Dream capturing niches that Big Tech may struggle to address due to geopolitical sensitivities, trust issues, or the sheer breadth of their general-purpose offerings. This trend will spur further investment in localized AI infrastructure and national data sovereignty initiatives, potentially leading to varied AI ecosystems across different geopolitical blocs. Furthermore, the imperative for AI-powered autonomous cyber defense will drive substantial R&D and procurement spending in the defense technology sector, pushing traditional cybersecurity firms to integrate advanced AI capabilities or risk obsolescence. Ultimately, Dream’s success could accelerate a global shift towards a more decentralized and nationally controlled digital future, impacting supply chains, regulatory frameworks, and international tech collaborations for decades to come.