I've been thinking about this a lot lately, and I wanted to throw it out to the community because I know a lot of you work on industrial and infrastructure siting.
Semiconductor fabs are one of the most unforgiving building types out there when it comes to site selection. We're not just looking for flat land near a highway anymore. A modern fab needs massive, stable power (we're talking hundreds of megawatts), millions of gallons of ultrapure water a day, seismic stability down to fractions of a millimeter of vibration tolerance, and proximity to a labor pool with very specific technical skills. Get any one of those wrong and you're looking at billions in stranded investment.
Here's where I think GIS stops being a "nice to have" and becomes the only real way to do this seriously. You can't evaluate that many overlapping constraints in a spreadsheet. You need to be able to layer utility infrastructure maps against floodplain data, seismic hazard zones, water rights and aquifer recharge data, transmission line capacity, rail and highway access, and workforce demographics, all at once, across multiple candidate regions — and actually see where the viable envelope is.
What I find most interesting is the water piece. Everyone talks about power availability, but ultrapure water sourcing and wastewater discharge permitting can quietly kill a site faster than power ever will. Overlaying watershed boundaries, existing water rights allocations, and drought risk models onto candidate sites tends to surface problems way earlier than a traditional feasibility study would.
The other thing I keep coming back to is incentive-zone overlap. A lot of states are stacking opportunity zones, foreign trade zones, and CHIPS Act-adjacent incentive districts on top of each other right now. If your GIS analysis isn't tracking those boundaries alongside the physical constraints, you might rule out a site that's actually financially the strongest option once incentives are factored in, or vice versa.
Where I'd push back on myself a little: GIS is incredible at narrowing the field, but it can create a false sense of precision. A weighted suitability model is only as good as the weights you assign to power vs. water vs. seismic vs. workforce, and those weights are ultimately a judgment call, not a purely data-driven output. I've seen teams treat a suitability score like it's gospel when it's really just a structured way of organizing an opinion.
Curious how others here are approaching the multi-criteria weighting side of this. Are you leaning on AHP (analytic hierarchy process) to defend your weighting choices to stakeholders, or something looser? And is anyone incorporating grid interconnection queue data into their models yet, given how long some regions are taking to approve new industrial-scale power draws?
Would love to hear how other GIS folks are tackling fab siting, especially anyone who's worked a project through to groundbreaking. What did the data get right, and where did it steer you wrong?
