KuppingerCole analyst predicts consolidation across AI security categories

Several AI security categories will be absorbed into larger platforms rather than surviving as standalone products, according to Jonathan Care, lead analyst and practice lead for AI at KuppingerCole Analysts. He compares agentic AI security to the early fraud market, where point solutions eventually gave way to a decisioning engine that orchestrates signals in real time.
Care expects runtime authorisation for agents, identity threat detection for non-human identities and AI-assisted security operations to hold up as categories. He bets against standalone prompt-injection filtering, AI security posture management as a separate product, and AI firewalls that claim to make any model safe through inspection alone.
Consolidation is already under way: Cisco bought Astrix, 1Password bought Apono and Check Point moved on Cyata this year, folding agent and machine identity security into bigger platforms. On funding, Care said capital is concentrating in fewer, larger rounds, with Pinpoint counting 62 seed and Series A deals in Q3 worth $1.16bn, 60% of all transactions.
He also addressed Israel’s share of cyber venture investment, roughly 46% of Pinpoint’s $5.26bn global total, warning that the bigger risk is ecosystem monoculture rather than geography. For agent deployments he listed five controls: a distinct identity per agent, least privilege with short-lived credentials, human approval for irreversible actions, readable logging, and a tested kill switch.