Guest: Arpit Joshipura
Company: The Linux Foundation
Show: 2026 Predictions
Topic: Cloud Native

The networking industry is at a pivotal moment. After years of incremental progress, 2026 marks the year when several transformative technologies reach critical mass. Arpit Joshipura, General Manager of Networking, Edge & IoT at the Linux Foundation, has been making annual predictions for a decade—and unlike many forecasters, he grades his own accuracy. This year’s predictions signal fundamental shifts in how networks are architected, deployed, and monetized.

AI-Native Networks Cross the Chasm

The first major prediction is bold: AI-native networks will cross the 50% adoption threshold across core, edge, and radio access networks (RAN) in 2026. This isn’t just about adding AI features to existing infrastructure. “We had networks that had autonomy. We had networks that were intent-based. But with the addition of AI and code that actually allows you to go fully AI native, we believe that we are on track to cross the 50% mark,” Joshipura explains.

This transition encompasses AI-driven automation, fully mature AI-driven RAN innovation, and comprehensive shifts to AI-enabled planning and optimization. The convergence of these capabilities represents a fundamental architectural change, not merely an incremental improvement.

Agents and APIs Unlock Network Value

For five years, the telecommunications industry has struggled to expose network value through APIs. That’s about to change. Joshipura’s second prediction centers on AI agents exploiting APIs to drive revenue—not just for telecommunications providers, but across the enterprise ecosystem.

“Agents and APIs will become a big revenue engine,” he states. Projects like Kumara, SEDM, and initiatives within the recently announced Agentic AI Foundation are converging to create new business models. The foundation already has over 100 members, demonstrating strong industry appetite for collaboration in this space.

This isn’t theoretical. Use cases are emerging across finance, automotive, and transportation sectors where AI agents interact with network APIs to deliver tangible business outcomes.

Edge AI Keeps Data Local

The third prediction addresses a critical constraint in AI deployment: latency and cost. “Every time you want to run a model, you can’t go to a public cloud. It’s latency sensitive and expensive,” Joshipura notes. The solution? Keep AI processing and data at the edge.

Linux Foundation projects like EdgeLake, Infinite Edge, and AI Eve are pioneering this approach. Healthcare, manufacturing, and automotive use cases are demonstrating that the cloud isn’t always the optimal location for AI workloads. When milliseconds matter and data sovereignty is non-negotiable, edge AI architecture becomes essential.

AI Data Centers Create Networking Opportunities

The AI data center boom is well-documented, but Joshipura highlights an often-overlooked implication: the massive networking opportunity this creates. These data centers require sophisticated network operating systems to stitch infrastructure together.

Enter Sonic, the open source network operating system that has become “the Linux kernel of a switch in a data center,” as Joshipura describes it. Originally deployed at hyperscale, Sonic is now expanding into enterprise environments and has been significantly modified to handle AI workloads. As sovereignty concerns shape data center architecture globally, open source solutions like Sonic are becoming strategic infrastructure.

Open Source RAN Enters the Efficiency Era

The fifth and perhaps most disruptive prediction: 2026 is the year open source software steals the spotlight in the radio access network (RAN) domain. After massive hardware investments in 5G, the industry is entering an efficiency era where doing more with less becomes imperative.

“We’re past the point where I can predict that open source software will steal the spotlight,” Joshipura says. Projects like Aukur Open, Codu, Randa Open, and the O-RAN Alliance Software Community are driving innovation in what he calls “the biggest network element that’s going to be shaken up significantly.”

This disruption parallels the cloud computing transformation. As 6G standards are being developed and 5G reaches maturity, the RAN is experiencing its own reinvention through open source innovation. Major announcements are expected at Mobile World Congress and beyond.

Navigating Challenges and Opportunities

Geopolitical tensions pose challenges, but the open source networking community remains inherently global. “You cannot separate that workforce, because phones and connectivity need to work across countries, across politics, across regions,” Joshipura observes. Projects like Silva are addressing regional requirements while maintaining global collaboration.

The opportunity lies in the convergence of technologies during this three-year window as 6G standards are finalized, 5G matures, and AI capabilities expand. The RAN is experiencing maximum innovation focus, with AI in RAN, AI on RAN, and projects like Duranta and Uphulu (funded by the U.S. government to open source CU and DU components) leading the charge.

Joshipura’s advice for enterprise leaders is straightforward: participate. “Open source is all about democracy. Join, become a member, attend events. If you don’t do that, you will be sidelined, because the community is going to move 10 to 100 times faster than your own individual innovation inside a company.”

The Linux Foundation’s focus for 2026 centers on AI infrastructure and harmonization. “Every building block that a vendor thinks they’re better at but isn’t differentiated, we will be hosting, harmonizing, and creating framework blueprints,” Joshipura explains. The rapid growth of the Agentic AI Foundation exemplifies this approach—creating collaborative spaces where industry can move faster than any single company could alone.

As networking enters this transformative phase, participation in open source communities isn’t just beneficial—it’s becoming essential for competitive survival.