Why Palo Alto Networks Is Betting Big on AI Gateway Startup Portkey

By ✦ min read

<h2 id="rise">The Rise of AI Gateways</h2><p>In the early days of large language models (LLMs), developers faced a chaotic landscape: dozens of model providers, each with their own SDKs, authentication methods, and endpoints. The solution emerged in the form of AI gateways—tools that unified these disparate systems under a single, OpenAI-compatible interface. Portkey, LiteLLM, Kong AI Gateway, and Cloudflare AI Gateway each tackled this fragmentation, but they were primarily developer tools, invisible to security teams.</p><figure style="margin:20px 0"><img src="https://cdn.thenewstack.io/media/2026/05/2346e9a0-getty-images-cemfwj6t27u-unsplash-1024x690.jpg" alt="Why Palo Alto Networks Is Betting Big on AI Gateway Startup Portkey" style="width:100%;height:auto;border-radius:8px" loading="lazy"><figcaption style="font-size:12px;color:#666;margin-top:5px">Source: thenewstack.io</figcaption></figure><p>Portkey stood out by processing trillions of tokens monthly for Fortune 500 clients, requiring just three lines of code to integrate and supporting over 3,000 LLMs, MCP servers, and agents. Its simplicity made it a favorite among developers, but its position—sitting between every agent and every model call—gave it a unique vantage point that the security industry soon recognized.</p><h2 id="gateway-security">The Gateway as a Security Checkpoint</h2><p>An AI gateway sees everything: every prompt, every model response, every tool call, memory read, and MCP server interaction. For regulated industries like finance, healthcare, and government, this creates an indispensable audit trail. As agentic workflows proliferate—each task potentially triggering dozens of LLM calls—the gateway transforms from mere plumbing into a critical log of autonomous decisions.</p><p>Palo Alto Networks understood this shift. By acquiring Portkey and integrating it into Prisma AIRS, the company aims to turn the gateway into a unified control plane for securing AI transactions across the enterprise. The deal, reportedly valued around $700 million, signals that the layer between agents and models is no longer optional infrastructure—it is a security checkpoint.</p><h3 id="additions">What Palo Alto Networks Brings</h3><p>Palo Alto is layering enterprise security capabilities onto Portkey’s foundation: identity and authentication, artifact scanning, automated red teaming, and runtime security. All of these are enforced at the gateway, where every agent call passes through. This transforms the gateway from a convenience tool into a place where organizations discover what their agents are <em>actually</em> doing, not what they hoped they would do.</p><figure style="margin:20px 0"><img src="https://cdn.thenewstack.io/media/2026/05/2346e9a0-getty-images-cemfwj6t27u-unsplash-scaled.jpg" alt="Why Palo Alto Networks Is Betting Big on AI Gateway Startup Portkey" style="width:100%;height:auto;border-radius:8px" loading="lazy"><figcaption style="font-size:12px;color:#666;margin-top:5px">Source: thenewstack.io</figcaption></figure><p>The strategy mirrors past moves in developer-owned infrastructure. Consider web application firewalls (WAFs): originally a network team’s concern, they evolved through developer adoption until Cloudflare turned them into a platform. The pattern is consistent: developer convenience leads to visibility, which leads to control, and ultimately to acquisition by security majors.</p><h2 id="implications">Implications for Enterprises</h2><p>For enterprises, the acquisition means that AI security is no longer an afterthought. With the gateway as a central enforcement point, companies can:</p><ul><li><strong>Audit all agent activity</strong> for compliance and governance.</li><li><strong>Enforce access controls</strong> based on identity and authentication.</li><li><strong>Scan artifacts</strong> for vulnerabilities before they reach models.</li><li><strong>Automatically test</strong> for adversarial inputs and prompt injections.</li></ul><p>This shift is especially relevant as competitors like Kong push agent gateway capabilities and A2A traffic management. The space is rapidly consolidating around security as a core feature, not just a bolt-on.</p><h2 id="future">The Future of Agentic Workflows</h2><p>As autonomous agents become more common, the gateway’s role will only grow. It will evolve from a simple proxy into a policy engine that governs every interaction between an enterprise’s AI systems and the outside world. Palo Alto’s bet on Portkey is a bet that this layer will become as essential as the firewall itself.</p><p>For developers, the trade-off is clear: the convenience of a unified endpoint now comes with enterprise-grade security built in. The security team is finally in the room.</p>