It’s an interesting question because, at a first glance, the answer seems obvious. Of course they do.
Networks are built from hardware - every switch, access point and gateway is designed, manufactured and supported by someone. Reliable hardware has always been the foundation of enterprise networking, and I don’t see that changing anytime soon.
But perhaps that’s the wrong question. Maybe what we should be asking is whether modern networks still need proprietary dependency.
For years, those two ideas have been treated as if they were the same thing. Buying networking hardware usually meant buying into the entire ecosystem. The controller came from the same vendor and the management platform came from the same vendor. Security, AI, licensing, support, hardware refreshes and even future architectural decisions gradually became tied to that original purchase.
That approach made perfect sense for a long time.
For decades, hardware was where most innovation happened, where vendors differentiated themselves through ASIC design, switching performance, chassis architecture and proprietary protocols. Hardware and software were developed together, and the tighter that integration was, the better the overall experience often became. Buying a complete platform wasn’t just a commercial decision, it was usually the technically correct one.
Today's conversations sound very different. When we speak with customers, very few of them ask about switching silicon or forwarding tables. They’re trying to automate deployments across hundreds or thousands of sites, and they’re integrating networking with identity providers, security platforms and cloud services. They want better visibility, APIs and operational consistency. Increasingly, they want AI to help them understand what’s happening across the network before users notice a problem.
In other words, the conversation has shifted. The network is no longer only defined by the hardware sitting in the wiring closet, it is increasingly defined by the software operating around it.
That’s why I think it's important to distinguish between proprietary hardware and proprietary dependency. There’s nothing inherently wrong with the proprietary hardware. Every physical device has a manufacturer and every platform has engineering decisions that make it unique. Competition between hardware vendors has pushed the industry forward for decades, and it should continue to do so.
Dependency is something else entirely. Dependency is what happens when selecting a switch also determines your management platform, your licensing model, your AI roadmap, your security integrations and the pace at which you can evolve your infrastructure over the next seven years. At that point, you’re no longer choosing hardware, you’re choosing an ecosystem. And leaving that ecosystem becomes far more difficult than joining it.
The interesting part is that we have already solved this problem elsewhere.
We’ve seen a different pattern before in other parts of enterprise IT. Servers became virtual machines, applications moved to SaaS, infrastructure moved into the cloud, and containers separated applications from operating systems. None of these changes eliminated hardware. Data centers didn’t disappear because virtualization arrived. Cloud providers didn’t stop buying servers.
What changed was the relationship between hardware and software. The hardware continued to matter, but it stopped dictating every decision above it. Networking feels like it's approaching a similar point.
Imagine expanding into a new region where one hardware vendor has long supply chain delays while another is readily available. Should that delay your project? Imagine refreshing wireless infrastructure while your existing switches continue to perform perfectly. Should replacing access points require replacing your operational model? Or imagine adopting a different SSE platform because it better aligns with your security strategy. Should a networking decision made years earlier prevent that choice?
These aren’t hypothetical situations, they happen every day. The more I think about it, the more I believe enterprise networking isn’t facing a hardware problem at all - it’s facing an architectural one.
Customers don’t necessarily want fewer hardware vendors. They want the freedom to choose between them. Hardware companies should continue building better hardware while software companies should continue building better operational platforms. And security vendors should continue pushing innovation in their own space.
Those are all healthy forms of competition. What no longer seems necessary is forcing customers to consume all of them as a single inseparable package. Perhaps that’s where enterprise networking is heading.
Not towards a world where hardware becomes irrelevant, but towards one where hardware becomes one decision among many instead of the decision that defines everything else.
At Join, that’s the direction we’ve chosen to follow. We work with certified hardware platforms because we believe customers should be able to select infrastructure that fits their business while maintaining consistent operational experience. Whether that network extends into SD-WAN, campus networking, SASE or cloud environments shouldn’t depend on the logo printed on the front of a switch. Ultimately though, this isn’t really about Join, it’s about how enterprise networking continues to evolve.
For years the industry has rewarded the tightest integration between hardware and software. That approach helped build the networks we rely on today but the next chapter may require something different. Not less integration, but more freedom. The freedom to adopt new technologies without replacing everything underneath them. The freedom to choose the right tool for the problem instead of the tool that happens to belong to the same ecosystem.
And perhaps most importantly, the freedom to change your mind.
To learn more about Join's approach to implementing network solutions that are hardware independent, you can read about our approach to Open Enterprise Networking here.