Five Startups, Six Months, One Test: Can Their AI Survive the Enterprise?

NetApp has selected five deep-tech startups for the 15th cohort of its Excellerator programme, with the latest batch centred heavily on a problem enterprises are increasingly confronting: moving artificial intelligence from experiments into secure, scalable production systems.

The six-month cohort brings together startups from India, the United States and Israel. **LightSprint AI** is building collaborative AI agents for software engineering; **Wald.ai** focuses on context-aware data-loss prevention for enterprise AI communications; and **Lasso Security** develops security tools for AI applications, agents and large language models. **Cogover** is working on no-code and low-code workflow automation, while QUIPU AI turns enterprise data into contextual intelligence for AI applications.

Rather than functioning only as a mentorship programme, Excellerator gives the startups access to NetApp technology leaders and data-infrastructure expertise, alongside opportunities to explore proof-of-concept engagements with internal teams. The programme is designed to help founders test whether technologies built in early-stage environments can operate within the demands of larger enterprises.

Two companies in the cohort, Cogover and QUIPU AI, are also part of **ExcellerateHER**, NetApp’s initiative supporting women-founded and women-led startups. NetApp says the initiative has supported 23 women-led companies so far. QUIPU AI, for example, plans to use the programme to validate how its long-term-memory technology can work in enterprise AI deployments.

The accelerator itself has been running since 2017. Across its first 14 cohorts, NetApp says it supported 90 startups that collectively raised more than $600 million, while facilitating 30 proof- of-concept engagements and six commercial contracts within NetApp.

That history makes Cohort 15 less about simply finding interesting AI startups. The real test is whether these five companies can move from technically promising products to systems enterprises are willing to test, integrate and eventually buy.