Analysis of CXOToday’s October 15 coverage.
AI Startups Security Automation 2026-10-15
What NetApp announced
NetApp launched the fifteenth cohort of NetApp Excellerator with five startups from India, the United States and Israel. CXOToday says the six-month program gives founders access to NetApp technology leaders, mentorship and data-infrastructure expertise, plus opportunities to explore proof-of-concept engagements. A startup may receive a US$15,000 grant for a potential PoC; “potential” matters, because admission does not mean that a PoC, product integration or customer deployment already exists.
The announcement reports 90 startups supported across the first 14 cohorts, more than US$600 million raised collectively, 30 PoC engagements and six commercial contracts inside NetApp. Those are program-level totals, not outcomes attributed to Cohort 15.
Cohort 15, mapped by enterprise problem
- LightSprint AI — collaborative AI agents intended to automate software development.
- Wald.ai — semantic data-loss prevention for AI communications, using contextual detection rather than pattern matching alone.
- Lasso Security — security for generative-AI applications, agents and large language models.
- Cogover — no-code and low-code workflow automation intended to connect data, applications, people and AI.
- QUIPU AI — enterprise knowledge intelligence that aims to turn source data into contextual, AI-ready information.
Cogover and QUIPU AI also participate in ExcellerateHER, NetApp’s program for women-founded and women-led startups. CXOToday reports that initiative has supported 23 startups.
The pattern: context, control and orchestration
This is not primarily a storage-product cohort. Four recurring control-plane problems connect it to NetApp’s “intelligent data infrastructure” positioning: deciding which enterprise context an AI system may use, preventing sensitive information from leaving its boundary, coordinating work across applications, and validating software produced or modified by agents. The relevant NetApp surface is therefore broader than capacity and throughput.
For operators, the useful adjacent reading is the site’s AI operations hub, ONTAP automation guide, security-hardening guide and backup & restore reference. Those pages describe the platform controls a PoC would have to respect; they do not imply that any Cohort 15 product integrates with ONTAP today.
What a credible PoC would have to prove
The five categories need different success measures, but each should begin with a declared data boundary and a baseline. A useful evaluation would record:
- Data access: which repositories, APIs and identities the system can reach; whether access is read-only; and how secrets and service accounts are governed.
- Quality: task completion or detection accuracy against a representative, versioned test set, including false positives and failure cases—not a curated demo alone.
- Control: audit events, human approval points, isolation between tenants and the behavior when a model, connector or data source is unavailable.
- Operations: latency, cost per completed task, recovery behavior, rollback and the owner who responds when automated output is wrong.
- Business conversion: whether the trial advances from discovery to a funded PoC, production use and a commercial contract.
For an ONTAP-connected experiment, API access should use a least-privilege role, a non-production scope and explicit request logging. The ONTAP REST API guide provides the platform-side starting point. No ONTAP integration was announced for these five startups, so any such design remains a hypothetical evaluation pattern.
The signal—and its limit
Cohort 15 shows where NetApp is seeking external innovation: agentic development, AI-era DLP, model and agent security, workflow orchestration, and knowledge context. The program’s historical PoCs and internal contracts indicate that Excellerator can produce more than mentoring, but the published totals do not establish a conversion rate for this cohort or customer adoption of its products.
The next evidence checkpoints are concrete: named PoC scopes, documented integrations, measured before-and-after results, production references and commercial agreements. Until those appear, Cohort 15 is best read as a structured portfolio of enterprise AI experiments—not as five finished additions to the NetApp product stack.
Source: CXOToday — “NetApp Excellerator Accelerates Deep-Tech Innovation for the AI Era with Cohort 15”. NetApp Black Box analysis supplies the control-plane framing and PoC evidence checklist.