戻る
Enhanced AI-specific peering and intelligent route optimizations from the Netskope NewEdge Network secure AI without performance trade-offs
SANTA CLARA, Calif. – July 21, 2026 – Netskope (NASDAQ: NTSK), a leader in modern security and networking for the cloud and AI era, today announced performance results derived from continued innovations of its NewEdge AI Fast Path technology. Part of the Netskope NewEdge private cloud, AI Fast Path optimizes the network path between users, sites, and agents to AI destinations for faster inference results and minimizing time to first token (TTFT), accelerating complex multi-prompt agentic AI workflows, as well as optimizing LLM performance when accessing large volumes of data. In testing performed across the NewEdge global network with real-world traffic, AI Fast Path was shown to reduce latency by as much as 90% to popular AI destinations.
More than 80%[1] of infrastructure and operations (I&O) leaders today aren’t confident that their current team and budget can meet the intensifying network performance, resilience, and security demands of the AI era. Through the combination of extensive peering and continuous, automated route optimization, AI Fast Path is shown to significantly reduce latency for AI traffic, giving enterprises fast, secure access to AI without the performance trade-offs common to approaches that rely on the public internet or third-party infrastructure.
“Netskope’s optimizations strip away network overhead associated with AI, which compounds as traffic shifts from simple inference to complex agentic workflows,” said Frank Dickson, Group Vice President, Security & Trust, IDC. “By eliminating latency, NewEdge AI Fast Path also protects against costly GPUs sitting idle, wasting money, and creating compute bottlenecks.”
Turning network telemetry into AI-ready performance
Netskope security, network, and AI services are delivered by Netskope NewEdge, a private cloud infrastructure consisting of more than 120 data centers in over 80 regions worldwide. Every NewEdge data center is designed to offer full compute and all Netskope services, including Netskope One SASE and Netskope One AI Security. As of June 2026, NewEdge maintains more than 12,000 network adjacencies to over 750 unique autonomous system numbers (ASNs), including direct peering relationships with leading AI destinations. This peering density ranks NewEdge among the largest networks in the world by global Internet Exchange (IX) participation, alongside hyperscale networks including AWS, Google, and Microsoft[2].
Available to Netskope customers as part of the Netskope One platform, AI Fast Path brings together the NewEdge network’s extensive peering with its unique Route Control technology that continuously analyzes network conditions. Every day, NewEdge analyzes tens of millions of routes, evaluating latency, jitter, and packet loss to identify the fastest, most reliable path, and automatically shifting traffic when a better path becomes available.
In addition, AI Fast Path takes into account measurements specific to AI traffic, including TTFT, sustained token generation rate, time per output token (TPOT), the percentage of total AI latency attributable to network roundtrip time, and other factors related to AI provider response times or geographic variations. AI Fast Path improves AI performance for customers, and bolsters network resilience and availability even with unexpected events that could trigger route changes, which might include:
- Internet “weather” events, such as traffic spikes
- Congestion ISP connectivity failures, such as fiber cuts
- AI provider infrastructure or cloud outages
In each case, Route Control automatically routes around the disruption, helping maintain a consistent digital experience for users, sites, and agents interacting with AI applications even when the underlying internet is not cooperating. In some of the most demanding AI use cases (such as a high-speed financial trading application, medical scan detecting a life-threatening anomaly or smart sensor triggering a security alert) such routing can also prevent catastrophic impacts
“Every AI workflow, whether it’s a person prompting a chatbot or an autonomous agent chaining together a dozen calls to an LLM, is only as fast as the network performance underneath it,” said Joe DePalo, Chief Platform Officer, Netskope. “Netskope is recognized as a market leader not only for AI security, but also for the network performance that underpins successful AI workflows. Thanks to NewEdge, our customers meet their need for AI speed without adding unnecessary risks or affecting the user experience.”
About Netskope
Netskope (NASDAQ: NTSK), a leader in modern security and networking for the cloud and AI era, addresses the needs of both security and networking teams by providing optimized access and real-time, context-based security for the AI ecosystem inclusive of agents, applications, tools, LLMs, people, devices, and data. Thousands of customers, including more than 30% of the Fortune 100, trust the Netskope One platform, its Zero Trust Engine, and its powerful NewEdge network to reduce risk and gain full visibility and control over cloud, AI, SaaS, web, and private applications — providing security and accelerating performance without trade-offs.
Learn more at netskope.com, on LinkedIn, and on Instagram.
Media Relations Contact:
Investor Relations Contact:
Methodology note: Route Control and NewEdge Traffic Management figures reflect NewEdge telemetry collected June 5–12, 2026. The up-to-90% latency reduction claim reflects testing to a representative AI destination from a single NewEdge data center during the same period; exact results may vary by location, destination, and network conditions. AI-specific performance measurements (TTFT, TPOT, throughput, network fraction, P95/P90 ratio, geographic COV) are currently derived from synthetic, ongoing monitoring of AI provider endpoints.
[1] Netskope research, “Crucial Conversations: Line of Sight” (January 2026) https://www.netskope.com/resources/reports-guides/crucial-conversations-line-of-sight
[2] Source: Hurricane Electric BGP Toolkit, bgp.he.net