The Israeli AI Infrastructure Layer: WEKA, NextSilicon, Pinecone, Run:ai, Gambit
The Israeli AI infrastructure layer anchored by five operators: WEKA (storage), NextSilicon (compute), Pinecone (vector databases), Run:ai (GPU orchestration, acquired by Nvidia), and Gambit (inference). The category where Israeli operators hold disproportionate global share.
The Israeli AI infrastructure layer entered Q1 2026 as one of the most institutionally important segments of the broader Israeli AI economy — anchored by five identifiable category operators: WEKA, NextSilicon, Pinecone, Run:ai, and Gambit. The layer sits structurally below the application AI ecosystem and above the hyperscaler footprint covered in The Israeli AI Hyperscaler Footprint, providing the storage, compute, vector-database, GPU-orchestration, and inference-infrastructure capability that underpins enterprise AI deployment globally.
The five operators
WEKA operates a high-performance data platform designed for AI and high-performance-computing workloads, with documented customer base across hyperscaler, enterprise AI, and research-computing segments. The company has executed multi-stage institutional financing through 2024–2025 at meaningful unicorn-tier valuation, per Globes and Calcalist coverage, anchoring the Israeli AI-storage-infrastructure category.
NextSilicon operates an AI-optimized compute platform designed to deliver performance gains across high-performance-computing and AI workloads through dynamic software-defined hardware acceleration. The company has executed substantial institutional financing rounds through 2024–2025, positioning it as one of the most institutionally capitalized Israeli AI-compute operators outside the hyperscaler footprint.
Pinecone (Pinecone Systems Inc., Israeli-founded, US-headquartered) operates the leading vector-database platform for retrieval-augmented generation (RAG) and other AI-inference workloads. Founded by Edo Liberty (former AWS Research lead, former Amazon Search lead, former Yahoo Research head) in 2019, Pinecone has executed multiple institutional financing rounds and operates as one of the category-defining AI infrastructure operators globally.
Run:ai operates a GPU orchestration platform for AI workloads, with documented customer base across enterprise AI, research-computing, and hyperscaler segments. The company was acquired by Nvidia in 2024 at a reported $700 million transaction, integrating Israeli AI-compute orchestration capability into Nvidia's broader enterprise AI infrastructure platform.
Gambit operates AI-infrastructure capability at the inference and model-serving layer, with documented institutional financing through 2024–2025 and positioning inside the broader Israeli AI-inference-infrastructure category alongside NextSilicon and adjacent operators.
The category architecture
The Israeli AI infrastructure layer sits structurally between the application-AI ecosystem (consumer and enterprise AI applications built on top of foundation models) and the hyperscaler footprint (Google, Microsoft, AWS, Meta, Nvidia Israeli R&D operations). The five identified operators address discrete categories: storage (WEKA), compute (NextSilicon), vector databases (Pinecone), GPU orchestration (Run:ai), and inference infrastructure (Gambit).
Per IVC-LeumiTech, Startup Nation Central, and trade-press coverage, the Israeli AI infrastructure layer carries materially higher institutional capital concentration than the broader Israeli AI application layer. The structural reason: AI infrastructure operators serve a global enterprise customer base with established procurement processes, durable customer-retention dynamics, and meaningfully higher unit economics than typical application-AI operators.
The hyperscaler relationship
The Israeli AI infrastructure layer operates in a structurally important relationship with the major US hyperscalers — Google, Microsoft, AWS, Meta, and Nvidia — covered in The Israeli AI Hyperscaler Footprint. The Israeli AI infrastructure operators either partner with the hyperscalers (WEKA, Pinecone, NextSilicon, Gambit) or operate inside them (Run:ai inside Nvidia).
The institutional read: the Israeli AI infrastructure layer extends the institutional reach of Israeli technology operators into the most economically important segment of global AI infrastructure spending. Where the broader Israeli AI application ecosystem competes against thousands of global operators, the infrastructure layer competes against a smaller, structurally more concentrated category — and Israeli operators hold disproportionate category share.
The structural read
The Israeli AI infrastructure layer Q1 2026 reflects a category at institutional inflection. The five identified operators — WEKA, NextSilicon, Pinecone, Run:ai (now inside Nvidia), and Gambit — collectively anchor an Israeli position inside global AI infrastructure that materially exceeds the country's share of the broader global AI application ecosystem.
The next institutional questions: whether any of the standalone operators (WEKA, NextSilicon, Pinecone, Gambit) executes a public listing inside 2026; whether the hyperscaler acquisition pattern that absorbed Run:ai into Nvidia continues across the remaining cohort; and whether additional Israeli AI infrastructure operators emerge as category-defining positions at scale alongside the existing five.
Source data: IVC-LeumiTech Q1 2026 capital data; Startup Nation Central AI infrastructure tracker; SEC filings of Nvidia regarding the Run:ai acquisition; company disclosures of WEKA, NextSilicon, Pinecone, Run:ai, and Gambit; coverage in Globes, Calcalist, Bloomberg, Reuters, TechCrunch, The Information. Related coverage: The Israeli AI Hyperscaler Footprint: Google, Microsoft, Meta, AWS, Nvidia. Data current as of Q1 2026.



