Loading...

The San Jose Campus Carrying 90 Percent of West Coast Internet Traffic

Smart Utilities

Published by IOT San Jose Research & Editorial Team

The San Jose Campus Carrying 90 Percent of West Coast Internet Traffic

More than 90 percent of internet traffic on the West Coast passes through a single campus in San Jose. That is not a metaphor about Silicon Valley's importance to the technology industry, it is a literal description of Equinix's Great Oaks campus off Great Oaks Boulevard, a cluster of data centers, SV1 through SV19 and counting, that functions as one of the densest interconnection points on the planet, with more than 240,000 individual connections linking network providers, cloud platforms, and enterprise customers in a single low-latency environment. For any conversation about edge computing, infrastructure placed physically close to where data is generated and consumed to minimize latency, San Jose is not just a city that happens to have some of this infrastructure. It is one of the handful of places the entire concept depends on.

Why Physical Location Still Matters for Computing Infrastructure

Cloud computing's marketing has long emphasized location independence, the idea that where a server physically sits should not matter to the person using it. Edge computing exists precisely because that promise breaks down for latency-sensitive applications. A round trip to a distant data center, even a fast one, adds tens or hundreds of milliseconds that autonomous vehicles, real-time industrial control systems, and increasingly, agentic AI applications acting on physical infrastructure cannot tolerate. Equinix's Silicon Valley facilities advertise less than 10 milliseconds of latency to 90 percent of the population across North America and Western Europe, a figure that is only achievable because of the physical density of network interconnections concentrated at this specific location, not because of any particular software optimization.

That density did not happen by accident. San Jose's Great Oaks campus has grown from its original facilities into a sprawling cluster comprising SV1, SV5, SV10, SV11, SV18, and SV19 among others, with the most recent addition, SV12x, energized in January 2026 as the first facility built under a new partnership between the City of San Jose and PG&E specifically designed to guarantee power delivery for data center expansion. That 20 megawatt, 82,000 square foot facility brought the Great Oaks campus to more than 40 megawatts of capacity in that section alone, part of a broader buildout that has made Equinix's San Jose presence one of the largest data center concentrations in the world by interconnection density specifically, as distinct from raw compute capacity.

How This Differs From a Typical Cloud Region

It is worth being precise about what distinguishes an interconnection hub like Equinix's Great Oaks campus from a standard cloud provider's regional data center, since the two get conflated in casual conversation about edge infrastructure. A cloud provider's regional data center is typically operated by a single company for its own customers' workloads, with network access to that facility mediated through the cloud provider's own infrastructure and pricing. An interconnection hub like Equinix's campus is fundamentally different: it is carrier-neutral, meaning any network, cloud provider, or enterprise can establish a direct physical connection inside the facility, and Equinix's business model is specifically built around enabling and monetizing those direct connections between otherwise unrelated parties. That neutrality is precisely why the campus has attracted such extraordinary interconnection density: a company does not need to choose a single cloud provider and accept whatever network path that provider offers, it can establish direct, low-latency connections to multiple providers, content delivery networks, and enterprise partners simultaneously from the same physical location.

Interconnection Density as the Real Edge Computing Advantage

What makes San Jose's role in edge computing distinctive is not simply the number of data centers, plenty of regions have more total square footage of data center space. It is the interconnection density, the sheer number of different networks, cloud providers, and enterprises that can exchange traffic directly with each other inside a single campus without routing through a distant intermediate point. Equinix describes its Silicon Valley facilities as serving the largest concentration of high-tech companies and peering hubs in the world, with campus cross-connects between SV1, SV5, SV10, and SV11 allowing customers in one building to interconnect directly with customers in another without that traffic ever leaving the campus.

For an edge computing application specifically, that density translates into a practical advantage: a company processing IoT sensor data, running real-time inference, or supporting a latency-sensitive application can co-locate compute infrastructure at Equinix's San Jose campus and gain direct, low-latency access to whichever cloud provider, content delivery network, or enterprise partner network it needs, without depending on the public internet's variable performance for that critical connection.

Business Value for South Bay Companies Building Edge Applications

For companies building latency-sensitive IoT or AI applications across the South Bay, understanding this infrastructure changes how a deployment architecture decision actually gets made. Rather than building proprietary edge infrastructure from scratch, a company can colocate compute resources directly at facilities like Equinix's San Jose campus, gaining access to the same interconnection density and low-latency network access that hyperscale cloud providers themselves rely on, without needing to build and operate that physical infrastructure independently. This matters increasingly for the kind of agentic AIoT applications covered elsewhere on this site, where real-time decision-making increasingly needs compute resources genuinely close to where sensor data originates, not simply labeled as edge computing while still routing through a distant regional data center.

Implementation Advice for Companies Evaluating Edge Infrastructure

  1. Distinguish true edge proximity from marketing labels. Not every service marketed as edge computing is physically close to where data originates, and evaluating actual network latency to a specific facility matters more than a vendor's edge computing branding.
  2. Evaluate interconnection options, not just compute pricing, when selecting a colocation or edge facility, since the ability to directly interconnect with specific cloud providers or partners without routing through the public internet is often the actual value driver for latency-sensitive applications.
  3. Plan for power availability as a genuine constraint, not an afterthought, given how directly data center expansion in the region now depends on coordinated utility partnerships like the one that enabled San Jose's SV12x facility.
  4. Consider proximity to San Jose's interconnection density even for applications physically deployed elsewhere in the South Bay, since a facility does not need to be the primary compute location to benefit from a fast, direct connection back to this hub.

Common Challenges in Edge Computing Infrastructure

ChallengeDescription
Power availability constraintsData center expansion increasingly depends on coordinated utility capacity planning, as seen in the San Jose-PG&E partnership behind SV12x
Cost of true edge proximityColocation at high-density interconnection hubs carries a cost premium compared to more distant, less connected facilities
Vendor and facility lock-inOnce an application architecture depends on a specific interconnection hub's network relationships, migrating to a different facility carries real switching costs
Balancing edge and cloud workload placementDetermining which workloads genuinely need edge-level latency versus which can run acceptably in a more distant cloud region requires careful architectural planning

These power and infrastructure constraints connect directly to the grid capacity challenges covered elsewhere on this site, where PG&E's broader work managing data center demand growth across the San Jose area sits upstream of exactly the kind of expansion that made facilities like SV12x possible in the first place.

What's Next for Edge Computing Infrastructure in San Jose

With Equinix continuing to expand its Great Oaks campus and San Jose's partnership model with PG&E establishing a template for guaranteeing power delivery to future data center projects, the region's position as a physical edge computing hub looks likely to strengthen rather than diminish, even as edge computing workloads increasingly distribute toward smaller, more numerous facilities closer to end users elsewhere. For companies building the next generation of latency-sensitive IoT and AI applications, San Jose's interconnection density remains a genuine, physical advantage that a purely cloud-native architecture built somewhere else cannot easily replicate, regardless of how sophisticated that architecture's software layer might be.

Frequently asked questions

More than 90 percent of internet traffic on the West Coast passes through Equinix's Great Oaks campus in San Jose, which functions as one of the densest interconnection points in the world with more than 240,000 individual network connections.

SV12x is a 20 megawatt, 82,000 square foot Equinix data center energized in January 2026, the first facility built under a new partnership between the City of San Jose and PG&E specifically designed to guarantee power delivery for data center expansion.

Interconnection density determines how many different networks, cloud providers, and enterprises can exchange traffic directly within a single campus without routing through the public internet, which is the actual driver of the low latency edge computing applications require, not simply how much total space a data center region has.

Equinix advertises less than 10 milliseconds of latency to 90 percent of the population across North America and Western Europe from its Silicon Valley facilities, a figure made possible by the physical density of network interconnections concentrated at the San Jose campus.

Yes. A facility elsewhere in the South Bay can still benefit from a fast, direct connection back to San Jose's interconnection hub, since a location does not need to be the primary compute site to gain value from proximity to this network density.

Power availability is an increasingly significant constraint, which is why data center expansion now depends on coordinated utility partnerships like the one between San Jose and PG&E that enabled the SV12x facility to guarantee sufficient power delivery.

Related articles

Top