NED

**Product and the concrete problem it solves.** NED is addressing a physical bottleneck behind Israel's software and AI ambitions: locally available data-center capacity has not kept pace with cloud adoption, GPU-intensive workloads, sovereign-data requirements, and the power and cooling densities of modern accelerators. The company is not an AI model vendor and does not claim to own a proprietary processor. It is building the sites, power strategy, cooling systems, connectivity, operations, and commercial structure that allow hyperscalers, cloud providers, enterprises, and public-sector users to place demanding workloads inside Israel. That distinction is strategically important. An Israeli startup can produce excellent AI research while still depending on compute located abroad, creating latency, jurisdiction, availability, and supply-chain constraints. NED's proposition is to turn local physical infrastructure into a dependable platform: customers can lease secured pods, data halls, or build-to-suit capacity rather than finance and operate a complete campus themselves. The initial Alpha Campus is planned near Netanya, approximately 20 kilometers from the Tel Aviv metropolitan area, and the company also lists a second site in Dimona and additional Netanya phases. Public reporting describes Alpha as a NIS 1.3 billion, roughly $350 million project, which places NED in the execution-heavy infrastructure category rather than the low-capital software-startup category.

**Core technology and how it actually works.** NED's technical differentiation is systems engineering across the data-center stack. Its solutions page specifies direct liquid cooling for dense AI deployments, support for NVIDIA DGX-class configurations, ring-based chilled-water distribution, fully redundant power and cooling, carrier-neutral facilities, and 24/7 on-site operations and monitoring. The Alpha Campus brochure describes direct liquid cooling of up to 100 kW per rack, while independent data-center coverage reports multiple cooling modes, including in-row cooling and rear-door heat exchangers. These choices matter because AI racks can exceed the thermal envelope of legacy enterprise colocation halls, and cooling failure can be as destructive as a power outage. NED also positions its facilities around hardened construction, protected customer disaster-recovery offices and storage, and Tier III-oriented design. Its sustainability claims include LEED Gold design, renewable-energy integration, market-leading power-usage effectiveness targets, and desalinated-water reuse. The architecture is therefore not a single invention but a coordinated control of power availability, thermal management, physical protection, network neutrality, phased construction, and operational continuity. The public record does not establish achieved PUE, uptime history, live GPU utilization, independent certification of the completed site, or a production operating track record in Israel, so these remain diligence items rather than proven performance.

**Market, customers, and go-to-market.** NED sells infrastructure capacity to customers that need large, reliable, and locally governed compute without building a facility from scratch. The target market includes hyperscalers, neo-cloud operators, enterprise AI teams, high-performance-computing users, financial and regulated organizations, and public-sector or government workloads that benefit from Israeli data residency and lower local latency. The commercial model is wholesale colocation and build-to-suit development. NED secures land and power, designs the campus, works with construction and infrastructure partners, and then provides halls or capacity to multiple large customers. This is a more concentrated buyer motion than ordinary colocation, but it can support large contracts and long-lived infrastructure relationships when power and permitting are scarce. NED's public materials say it is in advanced negotiations with major international clients, while a Calcalist report says the Netanya facility is intended to serve local activity from global cloud companies such as Amazon, Microsoft, and Google. Those references describe target or potential activity, not publicly confirmed signed customer contracts, and NED has not disclosed customer names, contracted megawatts, pricing, revenue, or occupancy. The go-to-market advantage is location and execution: a global cloud or AI operator can use NED's local development platform while relying on international data-center expertise from its ecosystem. The commercial risk is equally clear: a site can be technically excellent yet financially exposed if anchor tenants delay, power delivery slips, or AI demand shifts before commissioning.

**Traction, funding, and third-party validation.** NED has materially stronger evidence than a concept-stage infrastructure proposal, but its validation is project-based rather than revenue-based. The company says it was founded in 2021, while its Alpha Campus brochure and several partner materials describe a 2020 founding, so the founding year should be treated as a 2020-2021 range pending corporate-record confirmation. The company secured land and power for the Netanya campus in 2022, established a joint venture with the Levinstein Group, and broke ground in 2024 according to its own timeline. Data Center Dynamics independently reported the April 2025 groundbreaking for Alpha's first phase and described a $350 million, 42MW campus with direct liquid cooling from 100 kW per rack and expected 2027 operations. Globes separately reported an investment of approximately NIS 1.3 billion and said the first facility was expected to become operational in early 2027. The official brochure identifies strategic support from Goldacre and Noé Group and collaboration with Kao Data, while the corporate site identifies Electra as part of the local development and construction context through public reporting. This combination of a real site, secured power, construction activity, experienced infrastructure partners, and repeated independent coverage is meaningful third-party validation. It is not proof of commercial success: NED has not publicly disclosed a venture financing round, valuation, project debt structure, signed anchor tenant, revenue, profitability, or completed facility acceptance test.

**Founders and team background.** NED's founding group combines local operating access with international technology and infrastructure experience. The official company history names Daniel Efrati as founder and CEO, Shelly Landsmann as advisory-board chair, and David Bloom as a founding partner and chairman figure associated with Kao Data. The Alpha Campus brochure describes Efrati as a former head of investment banking at IDB New York, Landsmann as the former CEO of Microsoft Israel, and Bloom as the founder and chairman of Kao Data. Efrati's public interview with Goldacre explains that he became interested in data centers through his work in U.S. telecom and technology investment banking and selected Israel as the starting point for a Middle Eastern platform. Landsmann contributes senior Israeli technology-market credibility and enterprise relationships, while Bloom and the Kao Data connection provide direct relevance to operating large European data centers. NED's own materials say the broader management team covers site selection, design, construction, marketing, operations, and financing, which is the right functional shape for a capital-intensive infrastructure developer. The evidence does not support a precise employee count, a complete executive biography, or a claim that NED itself has operated a live Israeli campus. Diligence should therefore distinguish founder and partner experience from NED's own delivery record, including who owns each engineering, operations, security, and financing responsibility.

**Competitive dynamics.** NED competes in a market where capital, electricity, permits, connectivity, and customer trust are more decisive than a novel software feature. Israeli alternatives include MedOne, an established data-center and cloud-services operator with multiple underground facilities and domestic connectivity assets; Anan Data Centers, an AI-ready platform developing large liquid-cooled sites in Afula and Tzora; and Serverz, which markets an underground, blast-protected central-Israel facility with Tier III design and AI infrastructure services. Larger global providers such as Equinix, Digital Realty, and hyperscaler-owned facilities compete through international scale, existing customer relationships, and standardized operating processes. NED's distinct position is the combination of a new Israeli campus pipeline, protected underground construction, wholesale and build-to-suit flexibility, and the Kao Data, Goldacre, Noé, Levinstein, and Electra ecosystem. That can appeal to buyers that need local execution while retaining global standards. The edge is not yet a durable moat. Competitors can secure power, add liquid cooling, partner with real-estate owners, or use their own balance sheets, and a hyperscaler may prefer to own its strategic capacity. NED must prove that its site selection, resilience design, commissioning discipline, energy economics, and customer service produce lower total risk than existing operators. Its strongest near-term defensibility is execution speed and partner alignment; its weakest point is concentration in a small number of very large construction and occupancy milestones.

**Defense, security, and resilience dual-use relevance.** NED's core technology is credibly dual-use because resilient compute infrastructure supports both commercial and national-security functions, although the public evidence does not show a dedicated defense contract or classified deployment. The strategic case has several layers. First, government AI, cyber defense, intelligence analysis, emergency communications, and defense-industry workloads require high availability, controlled jurisdiction, and the ability to continue operating during regional disruption. Second, a hardened facility with redundant power and cooling, local connectivity, protected customer recovery space, and optional segregation can reduce the blast radius of physical attacks, grid disturbances, construction accidents, or network outages. Third, local AI capacity can reduce dependence on foreign cloud regions for sensitive workloads and can make it easier for Israeli institutions and allied partners to place data and inference within an agreed legal and operational boundary. NED's underground and fortified design was explicitly reported as a response to Israel's missile-threat environment, and Data Center Dynamics described the project as intended to serve both government clients and the private sector. That is a stronger resilience link than a generic claim that any cloud company is dual-use. The calibration is essential: NED is an infrastructure developer, not a cyber-defense vendor or weapons company. Its public materials do not confirm classified accreditation, defense customers, military service-level commitments, sovereign-cloud authorization, or operation through an actual emergency. The dual-use score reflects strategic infrastructure relevance and physical protection, not fielded defense capability.

**Growth stage, trajectory, and key diligence risks.** NED is best classified as mid-stage because it has moved from formation and land assembly into construction and multi-site development, but its flagship Israeli capacity is still transitioning toward operations. The official site lists 60 MVA under development and more than 600 MVA in pipeline, while the Projects page lists Netanya facilities and a 240 MVA Dimona campus. These figures are company-reported development targets and should not be treated as commissioned IT load. The near-term trajectory depends on bringing Alpha online, converting negotiations into signed customer commitments, and demonstrating that the protected design can meet AI-density, cooling, availability, and cost requirements. The central diligence risks are: (1) construction and commissioning risk, including schedule, contractor coordination, and grid interconnection; (2) power scarcity and price risk, because AI economics are governed by available megawatts and electricity cost as much as by land; (3) occupancy and financing risk, because a large campus needs anchor demand and substantial capital before recurring cash flow; (4) cooling and water risk, including actual performance under sustained GPU load and the environmental cost of desalinated-water reuse; (5) security-certification risk, because designed-for standards are not the same as independently certified operations; (6) concentration risk from reliance on a small set of infrastructure, real-estate, and strategic partners; and (7) geopolitical and supply-chain risk affecting equipment delivery, international customers, and the physical sites themselves. NED is strategically important to monitor because Israel's AI and digital-sovereignty ambitions cannot be separated from local power, cooling, and protected compute, but its ultimate value will be determined by operational evidence rather than development announcements.