The volume of traffic moving across the internet has never been higher – and it's accelerating. Streaming, gaming, remote work, IoT, and now the explosion of AI, are driving bandwidth demands to new heights. According to industry forecasts, global network traffic is projected to grow 5x to 9x through 2033. For network operators, this ongoing surge brings both pressure and promise: the pressure to scale infrastructure fast enough to keep up, and the promise of capturing that growing demand to drive increased revenue and deliver more value to customers than ever before.

But scaling capacity is only one part of the equation. The other (and arguably more difficult) challenge is economic. As demand grows, are the economics of delivering all that traffic still working in your favor? Can you confidently answer how traffic translates into spend, and whether those costs are justified by the revenue your business is receiving in return?

For many service providers, the answer is murky at best.

Even with powerful observability tools, most providers still lack clear visibility into which portions of their traffic are driving global connectivity spending. Bridging the gap between traffic data and financial insight remains one of the most difficult and consequential challenges in operating a network.

The cost of connectivity – high stakes, low visibility

Network operators have to manage a wide range of costs: data center space, hardware refresh cycles, operations and maintenance (O&M), OSS/BSS systems, just to name a few. For large-scale providers and content platforms, one of the key cost centers is interconnection – the exchange of traffic at the network edge. These costs typically include a blend of:

• IP transit, billed by 95th percentile usage or commit-based tiers
• Backbone and metro transport circuits
• Peering, both settlement-free and paid
• Private network interconnects (PNIs) with CDNs, OTT apps, and cloud platforms
• Internet Exchange (IX) ports
• Cross-connect fees
• Cloud interconnects like AWS Direct Connect and Azure ExpressRoute

These traffic exchanges form the connective tissues of the internet. But for network teams, understanding and tracking costs across all of these connectivity types is a big challenge. Here's why:

Billing complexity: Every provider and connectivity type uses different models – flat fees, 95th percentile, tiered rates, commit-based pricing. Bursts, overages, and usage ceilings add layers of nuance – and don’t get us started on peak of sums vs. sum of peaks. With so much variability, ISP billing errors can easily slip through unnoticed unless you have a reliable secondary data source to validate and verify monthly charges.

Data fragmentation: Traffic flows and network telemetry live in one platform. On the other hand, contracts, pricing, and billing are usually buried in finance and procurement systems, inaccessible in any programmatic way for teams to gain valuable insight necessary for critical business decision-making. 

Global variability: As interconnection grows at the edge, the footprint becomes increasingly global with an increased number of currencies and exchange rates that must be calculated into the mix.

No system of ownership: Network engineers manage operations, planning, and architecture. Legal manages contracts. Finance owns pricing agreements. No single team has the tools, or mandate, to bring it all together. Legacy monitoring and observability platforms were not built for merging elements of all these disciplines.  

Recognizing the critical need for better visibility, Kentik developed its Connectivity Costs workflow to solve this problem with clarity and automation. By combining interface classification and volumetric data with contractual pricing, Kentik gives operators a real-time, unified view of what they’re spending – across all providers, interconnection types, and geographic markets. It provides the foundational step towards making connectivity cost data accessible and actionable.

Why traffic-based cost intelligence matters

​​Understanding total connectivity spend is a useful and necessary first step; however, it only provides a surface-level look at cost drivers. The deeper business value comes from being able to analyze costs through the lens of traffic behavior. In other words: Which traffic flows are driving the most spend? Which customers, regions, or applications are generating the highest costs? How does that align with revenue?

These are questions that every corner of the service provider and digital infrastructure ecosystem needs to be able to answer.

• Access providers, such as ISPs and MSOs, are under mounting pressure to deliver ever-increasing volumes of internet content at high performance, all while competition drives down prices and subscriber revenue remains stagnant or in decline. At the same time, they’re expected to shoulder the bandwidth demands of OTT services like Netflix, YouTube, and Amazon Prime Video – without receiving any direct revenue share in return. In a historic shift, streaming viewership in the U.S. has now surpassed the combined viewership of cable and broadcast TV – a trend that shows no signs of slowing. Without the ability to isolate and evaluate the cost of receiving traffic from each OTT or CDN, access providers are left guessing when it comes to optimizing ingress content delivery paths and protecting per-subscriber profitability.
   
• For wholesale providers, who sit between upstream and customers, the cost-versus-quality dilemma is ever-present: where does it make sense to cut costs at the expense of performance, or vice versa – invest more to deliver higher quality of service? With market prices eroding and margins compressing, wholesale providers must be able to evaluate whether the price they’re charging downstream customers is sufficient to cover the upstream costs incurred from carrying that traffic. Not having visibility into the composition of the underlying costs of delivering that service makes the cost vs. performance balancing act nearly impossible. When decisions are made in the dark instead of being data-driven, providers risk underpricing services, overextending resources, and ultimately losing both competitiveness and customers.
   

• Content providers and CDNs have their own calculus. Their business depends on delivering all of that streaming content and data across a complex mesh of ISPs and geographies. Each market carries a different price tag. Knowing which regions are the most expensive to serve – and which eyeball destination networks in those markets are driving costs – can make or break their ability to scale sustainably and profitably. 

Despite different business models, these groups are all grappling with the same underlying question: What is the cost of delivering any given slice of traffic across our infrastructure?

Why this has been so hard to solve

The data needed to answer that question exists – but it’s not in one place, and it’s incredibly difficult to reconcile.

Traffic telemetry lives in network observability platforms. Cost and contract data live in business systems, across spreadsheets, procurement platforms, or vendor-specific invoices. Network topology and interface mapping documentation exists elsewhere. Bringing these together, accurately and at scale, is time-consuming, and error-prone – consuming precious engineering cycles and becoming an operational burden on already-lean infrastructure teams. 

More importantly, the dynamic and constantly evolving structure of traffic itself complicates cost computing. A single OTT application’s traffic will traverse dozens of interfaces across different providers, each with different rates and billing terms. Likewise, one interface might carry traffic for hundreds of different customers, services, or applications. Parsing out the exact cost of one specific flow-based traffic slice for that particular month, especially in real time, is like trying to split gas costs on a long carpool road trip when riders are getting on and off at different stops, across multiple cars, with fluctuating gas prices and shared detours.

This complexity forces most providers to rely on manually munging data points (i.e. V-lookups galore) from siloed spreadsheets, one-off analyses, or blunt averages that hide the real story. Some interconnection teams spend hours each month cobbling together estimates for internal reporting or contract negotiation support for commercial teams. Others skip the exercise altogether, focusing on high-level cost trends while leaving traffic-level cost attribution out of the equation.

Either way, the gap between traffic and cost remains, and that gap has ripple effects far beyond the engineering team.

The organizational disconnect

One of the biggest barriers to cost visibility is structural and cultural in nature – the teams responsible for building and operating networks are not the same teams responsible for pricing and selling services.

Engineering teams are focused and incentivized on performance, reliability, and team efficiency. Their success is measured in uptime, quality of service, and ability to scale and turn-up services quickly. Meanwhile, commercial teams – sales, marketing, finance – are focused, incentivized, and held accountable on revenue growth, margins, and customer management. 

When these teams don’t share a common view of traffic impact and cost, it creates real business risk as teams optimize in silos.

• Engineers may optimize for the most performant path, without realizing it's also the most expensive.
• Sales teams may close customer renewals that look strong on paper, but carry hidden delivery costs that chip away at profitability.
• Peering and infrastructure teams may over-provision capacity or stick with legacy interconnects that no longer make economic sense, all because the financial signals are missing or delayed. 

What comes next – a call for convergence

To thrive in today’s market, service providers need intelligent and innovative solutions to bring cost modeling out of the spreadsheet and into the operational fold. That means treating cost visibility not as a quarterly reporting exercise, but as a real-time, always-on capability. It requires connecting the dots between network telemetry and contractual data – and giving both commercial and technical teams a shared understanding of which traffic is driving spend.

The business of networking demands better answers. The network has become the backbone of our digital lives, yet the connectivity costs of running it is still treated like a black box. When cost becomes a first-class metric, alongside uptime, latency, and throughput, providers can operate smarter, leaner, and more strategically.

At Kentik, we’ve built exactly this. With the launch of Kentik Traffic Costs, we’re delivering the industry’s first real-time, automated workflow that bridges the gap between traffic and cost. Network operators can now instantly see how traffic translates into spend, gain insights to optimize delivery, and take action to protect margins and drive revenue.

Because in the end, success isn’t just about delivering traffic. It’s about delivering it in a way that makes business sense.