Its Time to Get Serious about Data Centers

I have been thinking a lot about data centers recently. They might not seem sexy or interesting from a planning perspective, but they keep coming up as a topic of concern in my work. At the APA conference in Riverside, I sat in on a discussion on the signing of assembly bill 98 banning large format logistics uses near residents, schools, parks and nursing homes in Riverside County. The intent of the new law is to mitigate supposed negative health impacts that warehouse and logistics facilities have upon nearby communities—especially in Southern California’s Inland Empire region. What's going on here?

The New York times has been writing extensively on the impacts of data centers on our country; there are already 5000 of them in the United States, McKinsey predicts that in a few years, data centers will be an $8 trillion dollar industry, requiring access to massive tracts of cheap land, energy and water to service. These centers can be incredibly loud, reports of adjacent residents say they are as loud as a leaf blower operating 24/7. They are often ringed with guard towers and razor wire fences for security purposes.

In my town of Los Angeles, there has been local buzz around conversion of high rise office space to data centers; One Wilshire Tower, a 30 story office building was recently converted completely, alarm was raised mostly of in terms of the power it uses to function. U. S. Power demand for Data Centers is currently about 3% of the energy usage in the US, and could account for as much as 11% of power demand by 2030 according to analysts at Goldman Sachs. It’s not a surprise that the Three Mile Island nuclear plant will be reopened to service Microsoft’s nearby data centers in Harrisburg, PA.

Many newly planned data centers are moving to rural areas, where the promise of jobs and economic boost to the local economy are often carrying the day, if only  because  non-disclosure agreements put in place by data users before negotiations could take place with city officials often limit public debate of the projects' merits before approvals are granted.

Athough Governor Newsom has made a positive step toward mitigating the impacts of Data Centers on communities with the passage of proposition 98, a number of factors also need to be addressed:

1) understand the full impacts of building a data center on the local economy: the post COVID world has pushed many people to want to stay at home and not interact the way they used to in person within city centers. This has put a glut on office space and a deficit on the energy and bandwidth needed for home computing. Data centers have proliferated to meet this demand. Although the benefit of a proposed data center being built in a local community often focusses on job creation, the power demands for these centers has left the local communities that accommodate them often scrambling to pay for the additional energy demand, with local residents often footing the additional cost through upcharges on their monthly electrical bill. This impact is rarely discussed until after the fact; and needs to be considered before approvals are granted by being coupled with a long term infrastructure investment strategy.

2) Office conversion should require some level of  human habitation; This shift has not been anticipated in current land use polices or urban framework plans, and in the short term we have generally just scrambled to fill up empty office spaces with 'whatever' comes along. Data Centers in a high rise office building are a placeholder use which should not be encouraged to stay in the longer term. Civic leaders need to recognize this trend and pivot quickly to accommodate what the market will bear while making it more attractive to office developers find another use that will contribute to what makes urban districts ultimately successful; people, street activity, a diversity of uses.

3) Promote human connection in the 'real world': there is all kinds of evidence that increased stress levels from on line activities are impacting human health negatively. We have to encourage land uses that promote community interaction in the 'real world' in addition to accommodating the needs of the 'online world'. For example, data companies should not be allowed to require non-disclosure agreements to negotiate with community leaders—it completely short circuits the democratic process of planning … part of a healthy community is having committed citizens who participate in visualizing the city of the future. If they did, they would perhaps see the linkage between their online behaviors, and the quality of the land uses around them. Thinking long term would perhaps prioritize different land uses over time.

4) promote localism (in a good way): although tech produces  some of the highest job creation numbers compared with other sectors,  we need to address the leakage of jobs away from the locality in which the data center is found. Creating local jobs should be a performance metric within development agreements. We need to think holistically about the value of a diverse local economy where tech, and consulting jobs are balanced with manufacturing, agriculture, services, and other sector jobs.

Steve Jobs once famously said “its not the customers job to know what they want”. AI is inevitable and we already have made it a part of our daily use. But the days where we can be ignorant to the ramifications of our online behaviors on the built environment are over. It is important for us all to get intentional about the impact of land planning on our quality of life. We need to urge the powers that be to better steward the land to produce healthier communities.  To simply let the market drive the land use discussion around computing power is to be thoughtless about how we live away from our screens.