Startup brief
AI Context Engines and Shadow-AI Defense Drive Israeli Early-Stage Deals
Tricentis acquires AI pioneer Tabnine as stealth launches from Modus and Bloom Security target enterprise model context and endpoint security.
Startup
Tricentis acquired Tabnine to solve the context gap in automated testing—combining enterprise QA workflows with deep codebase intelligence.
BackgroundFounded in 2013 as Codota, Tabnine pioneered early AI-assisted software development before shifting toward enterprise codebase context engines. The startup previously raised $55M from Khosla Ventures, Atlassian, OurCrowd, and Hetz Ventures to build code prediction models.
- Tabnine's technology maps complex software codebases to provide autonomous AI agents with full architectural context during automated development cycles.
- Tricentis plans to deploy Tabnine's context engine across its quality assurance suite, allowing AI test agents to auto-generate edge-case software tests.
- The acquisition marks an exit for early backers including Headline and TPY Capital, demonstrating persistent enterprise demand for deep AI codebase intelligence.
Startup
Bloom Security secured $20M to plug endpoint AI blind spots—giving security teams real-time control over shadow AI tools running on corporate laptops.
BackgroundThe rapid adoption of browser-based AI extensions and localized model servers has introduced unmonitored security vectors on corporate laptops. Enterprise IT teams lack critical visibility into confidential data shared with third-party endpoint AI tools.
- Seed investors include Okta Ventures, Runtime Ventures, and individual founders of Snyk, Demisto, Dig Security, and Talon, signalling strong cyber ecosystem backing.
- Bloom's security platform monitors endpoint AI extensions, unvetted scripts, and localized model servers to prevent real-time corporate data leakage.
- Former Palo Alto Networks engineering leaders Itay Keren, Ofir Balassiano, and Itay Frishman launched Bloom in 2025 to target shadow AI threats.
Startup
Modus raised $10M to solve enterprise AI token waste—building a Context Warehouse that slashes LLM compute expenses for autonomous agents.
BackgroundEnterprise AI implementations frequently fail or generate massive API bills due to redundant data ingestion and missing access permissions. Filtering raw database context before submitting queries to LLMs is essential for cost-effective enterprise deployment.
- Insight Partners led the round alongside Soma Capital and prominent technology founders including Wix co-founder Nadav Abrahami, anchoring early commercial momentum.
- The platform functions as an intermediary context layer, enforcing precise business permissions and filtering query data sent to external AI models.
- Modus claims its dedicated context architecture slashes API token consumption and underlying database query volumes by up to 90%.
Startup
Maia Fertility secured strategic funding to commercialize its enzyme platform—using livestock markets to validate technology meant for human reproductive care.
BackgroundAgricultural biotechnology firms apply advanced biomolecule delivery platforms to optimize commercial livestock breeding yields and biological efficiency. Improving fertility efficiencies in livestock establishes crucial preclinical data and manufacturing scale for future human therapeutics.
- Maia Fertility developed Ferizyme, a specialized enzyme delivery matrix designed to optimize reproductive efficiency and biological yield in commercial mammals.
- Initial commercial deployment targets dairy and livestock production facilities across European and American agricultural hubs to establish recurring revenue.
- The livestock clinical trials will provide essential safety and efficacy data required to support subsequent human reproductive health regulatory filings.
Unlock the full brief
Sign in to read every signal, takeaway, and source. Free account — Apple, Google, or email.