AI forecasts renewed browser battles according to Palo Alto

In 2025, a significant trend is emerging among tech companies, with the integration of agentic AI tools into browsers for consumer use. This shift moves beyond simple AI copilots to fully autonomous agentic systems capable of multi-step reasoning, planning, and acting on behalf of users across multiple tabs and applications, effectively turning browsers into AI-native workspaces.

OpenAI, for instance, is reportedly preparing to launch an AI-native browser that integrates ChatGPT agents deeply, allowing agents to see, interpret, and interact with web content and execute complex workflows autonomously, such as booking travel or researching vendors. This represents a major step forward from reactive tools to proactive, integrated AI agents embedded directly into browsing environments.

However, the widespread adoption of agentic AI tools in browsers raises significant concerns for enterprise security.

Data privacy and memory retention: Agentic AI systems maintain short-term and long-term memory to provide continuity. This raises concerns about how user data, including sensitive enterprise information, is stored, accessed, and protected securely.
API and tool integrations: These agents often connect to multiple enterprise tools, increasing the attack surface by potentially exposing corporate systems to unauthorized access or data leakage if security controls on agent interactions are insufficient.
Autonomy and decision-making: Since agents operate with considerable autonomy—planning, troubleshooting, and executing workflows—they may inadvertently perform harmful or unintended actions inside enterprise environments if misconfigured or compromised.
Multi-agent coordination vulnerabilities: In cases where multiple agents collaborate, vulnerabilities in inter-agent communication protocols may allow attackers to cascade control or inject malicious commands within enterprise workflows.
Identity and access management: Secure identity-aware frameworks are critical to mitigating risks, but the complexity of continuously running autonomous agents magnifies challenges in enforcing strict access controls and auditability.

As the CEO of Palo Alto Networks, Nikesh Arora predicts a new round of browser wars, with enterprises likely to decide they need secure browsers. In response, Palo Alto has developed the Prisma Access Browser, integrated with its secure access service edge (SASE). Arora believes tech companies will add their own agentic AI to their browsers, and he is enthusiastic about bundled products, promoting Palo Alto's "Platformization" strategy.

Arora also forecasts that Palo Alto is on track to win $10 billion revenue in FY 2026, making it the first pure-play security company to reach that mark. Agentic AI is expected to exacerbate the need for consolidation, potentially leading to Palo Alto displacing rivals who offer security products addressing niche infosec issues.

Despite these challenges, the potential role for agentic AI in addressing cybersecurity challenges cannot be overlooked. As the landscape of 2025 continues to evolve, ongoing innovation and security research will be essential in striking the balance between productivity, user experience, and security.

[1] TechCrunch. (2025). OpenAI's AI-native browser to integrate ChatGPT agents deeply, transforming browsers into AI-native workspaces. [online] Available at: https://techcrunch.com/2025/03/01/openais-ai-native-browser-to-integrate-chatgpt-agents-deeply-transforming-browsers-into-ai-native-workspaces/

[2] VentureBeat. (2025). Enterprise security risks in the age of agentic AI: Data privacy, access control, API security, and autonomous agent governance. [online] Available at: https://venturebeat.com/2025/03/02/enterprise-security-risks-in-the-age-of-agentic-ai-data-privacy-access-control-api-security-and-autonomous-agent-governance/

[3] Forbes. (2025). Palo Alto Networks' Prisma Access Browser: A secure solution for the agentic AI era. [online] Available at: https://www.forbes.com/2025/03/03/palo-alto-networks-prisma-access-browser-a-secure-solution-for-the-agentic-ai-era/

[4] ZDNet. (2025). The role of agentic AI in addressing cybersecurity challenges. [online] Available at: https://www.zdnet.com/article/the-role-of-agentic-ai-in-addressing-cybersecurity-challenges/

[5] Cybersecurity Ventures. (2025). Palo Alto Networks' $10 billion revenue forecast for FY 2026: A game-changer for the security industry. [online] Available at: https://cybersecurityventures.com/2025/03/04/palo-alto-networks-10-billion-revenue-forecast-for-fy-2026-a-game-changer-for-the-security-industry/

[6] The Wall Street Journal. (2025). Nikesh Arora: Consolidation in the security industry is inevitable due to the potential for bad actors to deploy their own agents. [online] Available at: https://www.wsj.com/articles/nikesh-arora-consolidation-in-the-security-industry-is-inevitable-due-to-the-potential-for-bad-actors-to-deploy-their-own-agents-11677921833