Identity for the AI Agent Era
Rico Mallozzi@rico
· 1
−2.7%
since Oct 15
The idea
I believe AI agents are becoming a new class of user inside every company, and someone has to verify, limit and switch them off, which makes identity the layer AI stretches hardest. So the core holdings are the stand-alone identity players plus Microsoft's login layer, surrounded by platforms that bought their way into identity, and rounded out with cloud, network and data security names plus a broad cybersecurity fund for steadiness. I keep a slice in cash and avoid concentrating too heavily in any single name, planning to hold this for five to ten years as the thesis plays out.
What it owns
- As AI agents proliferate, every agent that needs a login and permissions creates direct new demand for identity platforms that can grant and revoke access instantly. · Okta, Inc.It is the largest independent identity company still publicly traded and has built AI agent controls into its platform, making it a direct play on the idea that every agent needs a verifiable login and switchable permissions., SailPoint, Inc.It decides who and what can access which systems across enterprises, and has added AI agent identity governance tools that treat agents like any other identity requiring approval and oversight., Microsoft CorporationIt already manages logins for most large enterprises through its cloud identity platform, so AI agents built on Microsoft tools will likely get their identities there by default.27%
- Large security platforms have acquired identity startups, so these companies win if identity becomes a standard feature bundled into broader security suites rather than sold standalone. · Palo Alto Networks, Inc.It closed a major acquisition of privileged identity security, which covers the most powerful accounts and the exact kind of access that risky AI agents would need., CrowdStrike Holdings, Inc.It acquired identity controls that grant and revoke access in real time, positioning it to win if identity protection becomes a standard feature inside the largest security platforms., Cisco Systems, Inc.It owns login tools for employees and acquired technology to secure non-human identities like AI agents, giving it identity controls and a massive existing customer base to sell them to., ServiceNow, Inc.It runs the workflows where many companies will deploy AI agents that take real actions, and its access control tools can limit what each agent is allowed to touch.24%
- AI agents make constant requests across the internet to reach apps and data, creating ongoing demand for platforms that inspect and control that traffic. · Zscaler, Inc.It inspects every connection before granting access, a practice called zero trust, and AI agents making constant requests throughout the day represent new traffic that needs that scrutiny., Fortinet, Inc.It sells network firewalls to a wide range of companies, adding network-level security depth that complements identity controls across the portfolio., Cloudflare, Inc.It controls a large portion of internet traffic and has demonstrated secure transactions for AI agents, positioning it as a potential gatekeeper that verifies agent identity and activity on the open web., Check Point Software Technologies Ltd.It is a long-running firewall company with steady profits that adds stability to a multi-year hold while adapting its products to new AI agent threats.19%
- Agents with access to sensitive company information raise the stakes on knowing where data lives, who touches it, and how to recover systems when things go wrong. · Rubrik, Inc.It backs up company data and can restore identity systems after an attack, which becomes more critical as AI agents with broad access increase the odds of costly breaches., Varonis Systems, Inc.It tracks what and who touches sensitive data inside companies, and AI agents pulling files is exactly the activity it is built to watch and restrict., SentinelOne, Inc.It protects servers and computers and is expanding into guarding AI systems as they run, covering the attack prevention side of the view that AI raises security demand., Tenable Holdings, Inc.It maps weak spots and misconfigurations across company systems, including the new accounts and connections that AI agents create, before attackers can exploit them.15%
- A diversified cybersecurity fund spreads exposure across the whole industry, so individual company stumbles matter less over a long holding period. · First Trust NASDAQ Cybersecurity ETFIt tracks a wide range of cybersecurity companies, so the take stays broadly exposed to the whole industry even if any single company underperforms.7%
- Cash8%