When Hybrid Work Became Permanent: The Infrastructure Gap Most Enterprises Still Haven't Closed

The Assumption That Broke Enterprise Video Infrastructure

For most of the 2010s, enterprise video conferencing was treated as a supplemental capability—a nice-to-have for executive meetings and cross-site collaboration that happened a few times a week. Network architects provisioned bandwidth accordingly. IT teams deployed hardware endpoints in a handful of conference rooms and called it done. The underlying assumption was simple: most employees would be in the office most of the time.

That assumption is gone. According to data from Stanford's Work From Home Research project, over 25 percent of US workdays are now performed remotely, and hybrid schedules have become a standard employment expectation rather than a temporary accommodation. The problem is that the infrastructure most enterprises rely on was never redesigned to reflect this shift. It was patched, extended, and jury-rigged—and the cracks are showing.

What Legacy Architectures Were Actually Built For

Traditional enterprise video conferencing architectures were designed around a hub-and-spoke model. A central on-premises MCU (Multipoint Control Unit) or a regionally deployed bridging system would handle call routing, with the assumption that the majority of participants would be on-site, connected to the corporate LAN, and using certified hardware endpoints.

This model performed reasonably well when those conditions held. Internal network traffic stayed internal. Bandwidth was predictable. QoS policies could be applied cleanly at the network edge.

But hybrid work obliterated those conditions. Today, a single conference call might simultaneously include:

• A user on a corporate laptop over a home broadband connection in suburban Ohio
• A mobile participant joining via a smartphone on LTE in downtown Chicago
• A conference room endpoint at a leased co-working space in Austin
• A remote contractor using a third-party client from a different platform entirely

Legacy MCU-based systems were not designed to gracefully handle this kind of endpoint diversity. The result is predictable: degraded call quality, inconsistent participant experiences, and IT teams fielding complaints they have limited ability to resolve.

The Latency Problem Is Structural, Not Incidental

One of the most persistent pain points in legacy hybrid deployments is network latency—and it is often structural rather than incidental. When a remote user's media stream must traverse the public internet, hit an on-premises media node, traverse back out to another remote participant, and then return, the round-trip delay compounds at every hop.

This is not a problem that can be solved by upgrading home internet connections or asking employees to use wired Ethernet. The issue is architectural: centralized media processing creates unavoidable geographic latency for distributed participants.

Modern video conferencing platforms address this by distributing media processing closer to where participants actually are. Pexip's architecture, for example, supports distributed deployment across on-premises nodes, private cloud infrastructure, and public cloud regions. This means that media can be processed at a node geographically proximate to the participant cluster, dramatically reducing round-trip times and the jitter that makes video calls feel unreliable.

For enterprises with a significant remote workforce concentrated in particular metro areas, this distributed model is not a luxury—it is a latency requirement.

Integration Debt: When Your Video Platform Doesn't Talk to Your Collaboration Stack

Beyond raw call quality, hybrid work has surfaced a second category of infrastructure debt: integration fragmentation. Enterprise collaboration stacks have grown substantially more complex over the past five years. Most large US organizations now operate some combination of Microsoft Teams, Slack, Zoom, and legacy SIP/H.323 endpoints—often simultaneously.

Legacy video conferencing systems typically offer limited, proprietary integration pathways. A Cisco hardware endpoint may join a Cisco Webex call natively, but routing that same endpoint into a Microsoft Teams meeting requires a gateway, additional licensing, and operational overhead that IT teams often struggle to justify or maintain.

This fragmentation has real costs. Users who cannot join a meeting from their preferred client or device simply drop off. IT teams spend disproportionate time managing one-off integration exceptions. And organizations end up paying for multiple platforms because no single system can serve all their interoperability requirements.

Interoperable platforms like Pexip are specifically architected to bridge this divide. By supporting both standards-based protocols (SIP, H.323, WebRTC) and native integrations with major collaboration platforms, Pexip allows enterprises to consolidate gateway functions and reduce the number of discrete systems IT must maintain.

Bandwidth Constraints in the Age of Always-On Video

Another structural challenge is aggregate bandwidth consumption. When video conferencing was episodic, bandwidth planning was straightforward. When it is continuous—when employees are expected to be on camera throughout their workday, joining meetings from home networks that also support streaming, gaming, and smart home devices—the calculus changes entirely.

IT teams have limited direct control over residential broadband quality. What they can control is how efficiently their conferencing platform uses available bandwidth. Modern codecs like H.264 High Profile, H.265, and VP9 offer substantially better compression efficiency than older standards. Platforms that support adaptive bitrate streaming can dynamically reduce video quality during periods of network congestion rather than dropping the call entirely.

Enterprise IT teams evaluating conferencing infrastructure should specifically audit codec support and adaptive bitrate capabilities as part of any platform assessment. These are not marketing features—they are operational requirements for hybrid workforces.

Building an Architecture That Reflects How Work Actually Happens

Closing the hybrid work infrastructure gap requires treating video conferencing not as a peripheral service but as a core communications layer—one that deserves the same architectural rigor applied to identity management, network security, or application delivery.

That means:

• Distributed media processing to reduce latency for geographically dispersed participants
• Protocol interoperability to support the full range of endpoints and platforms in your environment
• Scalable call capacity that can expand during peak demand without requiring manual intervention
• Centralized management and observability so IT teams can diagnose quality issues across all participant types from a single plane

Pexip's self-hosted and cloud-connected deployment models are designed to support exactly this kind of architectural flexibility. Enterprises can begin with an on-premises deployment and extend into cloud regions as remote workforce density grows—without replacing the core platform.

The organizations that are managing hybrid work most effectively are not the ones that found the cheapest video conferencing solution. They are the ones that recognized, early, that the architecture underneath their conferencing platform would determine whether hybrid work was a productivity enabler or a persistent source of IT friction. The gap is closable. But it requires treating it as an infrastructure problem, not a user training problem.