Energy & Sustainability
How San Jose's Grid Is Being Rewired for the AI Data Center Boom
Published by IOT San Jose Research & Editorial Team

- The Problem: Physical Grid Capacity Cannot Scale as Fast as Demand
- Why This Matters Beyond One Substation
- How the Interconnection Queue Actually Works
- Implementation Lessons for Utilities and Large Energy Users
- Common Challenges in Modernizing Grid Capacity
- What This Means for Renewable Energy Integration
- The Bigger Picture: Sensors and Software as Grid Infrastructure
California's grid operator has forecast that data center growth in the San Jose area alone could add up to 500 megawatts of new electricity demand, a load increase big enough to strain a substation built for a very different era of energy consumption. Rather than wait years for a new transmission line, currently scheduled for 2032, PG&E chose a faster fix: deploying connected, software-controlled power flow devices at its Los Esteros substation in the Alviso community that can dynamically redirect electricity across the existing network. It is a case study in a broader pattern reshaping utility infrastructure nationwide, one where sensors and control software, not new copper and steel, are becoming the fastest way to add grid capacity.
The Problem: Physical Grid Capacity Cannot Scale as Fast as Demand
Traditional grid capacity expansion means building new transmission lines and substations, a process that routinely takes the better part of a decade once permitting, environmental review, and construction are factored in. That timeline was tolerable when electricity demand grew slowly and predictably. It is not tolerable when a single wave of AI data center development can add hundreds of megawatts of new demand to a specific substation within a two or three year window, which is roughly the situation PG&E faced in the Alviso area as multiple data center projects moved forward simultaneously near the Los Esteros substation.
The technology PG&E deployed to bridge that gap, an advanced power flow control system from a company called Smart Wires, works by actively managing how electricity flows across existing transmission lines rather than simply carrying whatever load is pushed onto them. Conventional grid infrastructure is largely passive: power flows according to the physical properties of the lines and the laws of electricity, with limited ability to redirect it toward underutilized capacity elsewhere in the network. The SmartValve devices PG&E installed introduce active, software-directed control, sensing thermal loads in real time and redirecting power away from overloaded lines toward lines with spare capacity.
Why This Matters Beyond One Substation
The Los Esteros project increased available capacity at that single substation by more than 100 megawatts, enough to unlock power for data center projects that would otherwise have waited years for a new transmission line to be built. PG&E has stated that similar approaches across its broader service territory have unlocked nearly 4 gigawatts of capacity, described as enough to power more than 2.5 million homes, supporting faster integration of both new demand and clean energy generation. That scale matters because it demonstrates this is not a one-off engineering fix for a single congested substation. It is becoming a standard tool utilities reach for whenever new load growth outpaces the physical construction timeline for permanent infrastructure.
For San Jose specifically, this has direct business implications. Companies planning data center, manufacturing, or large commercial development in the South Bay have historically faced multi-year interconnection queues tied to physical grid capacity constraints. Dynamic power flow control technology shortens that timeline in specific, congested locations, though it does not eliminate the underlying need for new transmission capacity over the longer term. Businesses evaluating a Silicon Valley expansion should treat this as a meaningful but partial solution: it buys years of breathing room, not a permanent fix.
How the Interconnection Queue Actually Works
It helps to understand why a single substation upgrade generates so much attention from the business community. Before any large facility, whether a data center, a manufacturing plant, or a commercial campus, can draw significant new power from the grid, the project has to move through an interconnection study process managed by the grid operator. That process evaluates whether existing infrastructure can safely absorb the new load without overloading transmission lines or destabilizing the surrounding network. When too many large projects request interconnection in the same area within a short window, as happened near the Los Esteros substation with multiple data center developments moving forward simultaneously, the queue backs up and projects can wait years for approval even when the underlying business case is otherwise ready to proceed. Dynamic power flow control technology directly addresses this bottleneck by increasing the amount of load a given substation can safely absorb without new construction, which is why its deployment translated so quickly into unlocked capacity for the Alviso area data center projects waiting behind it.
Implementation Lessons for Utilities and Large Energy Users
- Deploy dynamic control at known bottleneck points first. PG&E targeted a specific substation facing an identified, near-term capacity crunch rather than attempting a blanket rollout across its entire territory.
- Treat this as a bridge, not a replacement, for permanent infrastructure. The Los Esteros deployment explicitly coexists with a planned new transmission line arriving in 2032, with the SmartValve system continuing to provide dynamic balancing value even after that line is complete.
- Involve the grid operator early in technology selection. The California Independent System Operator evaluated multiple options, including reconductoring existing lines and energy storage, before identifying advanced power flow control as the most effective, timely, and cost-efficient path for this specific location.
- Plan for modular, relocatable deployments. A key advantage of the SmartValve technology is that it can be redeployed or reconfigured as grid needs shift, which matters given how quickly data center demand projections have changed in the current AI infrastructure buildout.
Common Challenges in Modernizing Grid Capacity
| Challenge | Why It Occurs |
|---|---|
| Demand growth outpacing construction timelines | New transmission lines take years to permit and build, while data center demand can materialize within one or two years |
| Passive grid infrastructure | Traditional lines cannot actively redirect load toward underutilized capacity without added control technology |
| Interconnection queue backlogs | Multiple large projects competing for capacity at the same substation simultaneously |
| Balancing renewable integration with new demand | Dynamic load management must account for variable renewable generation alongside growing consumption |
These challenges connect directly to broader trends tracked across the energy and utilities sector, where connected sensor and control technology is increasingly the fastest lever available to utilities facing demand growth that physical construction alone cannot match on the required timeline.
What This Means for Renewable Energy Integration
Grid capacity constraints do not only affect new demand like data centers. They also limit how much renewable generation a region can bring online, since new solar and wind projects need available transmission capacity to actually deliver power to where it is consumed. A dynamically managed grid that can redirect load in real time creates room for both sides of that equation simultaneously, absorbing new demand while also making space for intermittent renewable generation that does not always align neatly with existing transmission capacity. This is part of why utilities describe technology like SmartValve as supporting clean energy integration and not just data center growth. A substation that can only handle static, predictable load has little flexibility to accommodate solar generation that peaks at midday or wind generation that varies by the hour. A substation with active, software-directed power flow control can absorb that variability far more gracefully, which matters increasingly as California continues adding utility-scale and distributed renewable capacity across the region.
For San Jose, which has set its own municipal targets around clean electricity and carbon reduction, this connection between grid modernization and renewable integration is not incidental. The same infrastructure investment that unlocks data center capacity in Alviso also improves the grid's ability to absorb solar and battery storage projects elsewhere in the county, since both problems ultimately come down to managing variable load and generation on infrastructure that was originally designed for neither.
The Bigger Picture: Sensors and Software as Grid Infrastructure
What makes this shift genuinely significant is the underlying philosophy it represents. For most of the grid's history, capacity was a physical property determined by the size and number of wires in the ground. Increasingly, capacity is becoming a function of how intelligently existing physical infrastructure is managed in real time, using the same category of connected sensor and control technology that governs adaptive traffic signals and predictive maintenance systems on a factory floor. In each case, the pattern is the same: existing infrastructure gets a software and sensor layer that makes it dramatically more responsive without requiring a full physical rebuild.
For San Jose, a city simultaneously pursuing an AI data center boom and long-standing climate commitments around renewable energy integration, that pattern is likely to define grid investment for the rest of this decade. Expect more substations across the region to receive similar dynamic control upgrades as utilities work to keep pace with demand growth that has already outrun the traditional planning timeline for new transmission infrastructure.
