Public Opposition Could Doom Data Center Expansion
Across the U.S., communities are mobilized against data center expansion—and with good reason. For most residents, data centers remain invisible until major changes to utility bills arrive. But in Q2 2025 alone, opposition to data centers in the U.S. rose 125%, with an estimated $98B across 20 projects blocked or delayed since 2023 (Data Center Watch).
In the U.S. alone, data centers consume enormous amounts of resources with little regulation or oversight. Together, they consume 163 billion gallons of water annually (Project Censored)—equivalent to Denmark’s annual consumption. What’s more, data centers account for 4% of U.S. electricity demand (170 TWh per year), equivalent to Poland’s annual consumption. This is estimated to at least triple by 2029, driving consumption to mirror that of Germany.
This surge reminds me of the mining industry’s struggle to expand mineral extraction in the face of public opposition, especially from vulnerable and indigenous communities. Just as a mine pit depletes local resources, so does the commissioning of massive data centers. But an important distinction exists: the mining sector has clear market signals for expansion tied to real products—for example, lithium for batteries. Meanwhile, the data center boom lacks the same market signals; rather, expansion is based on forecasted and somewhat abstract AI processing needs.
Big Tech is scrambling. Billions in capital are deployed, and the promise of AI’s investment returns hangs over every boardroom meeting. So where is the industry turning when communities, regulators, and even the power grid itself begin to say no to data center expansion dreams?
They’re Going Up
A well-funded race is underway to bring data center infrastructure to space. No competition with local resources, infinite access to solar power, and space’s naturally cooling environment make it sound like the holy grail for data centers.
Money is flowing:
- Star Catcher raised $65M to build a power grid in space to power orbital data centers.
- Cowboy Space Corp (formerly Aetherflux) raised $275M in Series B funding from prominent investors including Index Ventures and Breakthrough Energy Ventures to launch its first satellite into orbit in 2026 to demonstrate space-to-Earth power beaming.
- Earlier this year, Starcloud raised $170M in Series A funding from EQT Group, Macquarie Asset Management, and others for its low-Earth orbit data centers.
- Orbital Chenguang received up to $8.4B in potential credit lines from 12 major Chinese financial institutions to build orbital data centers by 2035 to compete with SpaceX.
The Story Big Tech Is Selling Isn’t Adding Up
Most orbital data center proposals rest on SpaceX achieving launch cost reductions for its Starship rocket. Current economics sit around $3,600 per kilogram for Falcon 9’s launch costs—its primary rocket—before accounting for build costs. Google researchers identified $200 per kilogram launch costs as necessary to make orbital data centers cost-competitive with terrestrial ones. This requires approximately 180 Starship launches per year over the next decade—an ambitious target since Starship has yet to reach operational scale, and no other provider comes close.
Mass efficiency is critical to closing cost gaps. Google researchers showed early progress on radiation tolerance, demonstrating its chips can survive a five-year mission. But radiator panels large enough to support power-dense AI chips would add significant mass and weaken economics. On thermal management, the paper offers little more than an acknowledgement that it remains a future milestone with no economic modeling to forecast resolution.
Google’s paper treats technology constraints in isolation as problems to be solved sequentially. But Caltech researchers recently stress-tested Google’s claims by building physics-based economic models. These models require simultaneous closure of every constraint—mass efficiency, thermal management, space-to-ground communications, utilization, and lifetime maintenance.
Even under the most favorable assumptions, combined launch and build costs would need to be near $250 to $1,000 per kilogram just to match terrestrial economics. That requires at least a 3X reduction from Falcon 9’s current costs—ambitious even under the most optimistic forecasts—and up to 14X at the lower bound, which is unfeasible near-term. Caltech’s reporting is a stark warning that incremental progress alone will not close the gap.
What to Make of It
You’ll remember that in 2025 we flagged the space sector as emerging in our Global Cleantech 100. It received some pushback from our internal team on grounds that space wasn’t cleantech, but ultimately, we couldn’t ignore the signals. Now, little more than a year later, space is commanding headlines and nine-figure funding rounds. We weren’t wrong about the direction. But the question now is the timeline and the costs.
Right now, tremendous capital is being deployed behind infrastructure that faces extraordinary engineering challenges that would stop any Earth-based project dead in its tracks. Beyond the economics, space is home to violent and unpredictable operating conditions—meteorite impacts, orbital debris, and solar radiation events—that could result in the unrecoverable failure of multi-billion-dollar assets with no technician able to intervene.
Space gets a pass on that scrutiny in part because the narrative tugs on childhood dreams of space exploration, and the incumbents backing it are credible. That’s precisely why investors should be cautious.
We’re monitoring this space closely—but don’t look up just yet.
