The news
The Equity podcast examined whether the AI industry could reduce resistance to new data centers by abandoning non-disclosure agreements. Hosts and guests focused on recent moratoriums and the sector’s broader communication failures with communities. The discussion centered on one practical step: ending routine NDAs that keep project details hidden from the public until deals are already signed.
Context
Data centers form the physical backbone for training and running large models. Several localities have enacted temporary bans on new facilities while they study power, water, and land use impacts. The podcast treated these moratoriums as a direct response to rapid expansion plans that often arrive with little public detail. Participants noted that the speed of AI infrastructure growth has outpaced the industry’s ability to explain itself to the towns and counties asked to host the facilities.
The core tension is straightforward. Local governments want concrete information before they approve permits or allocate grid capacity. Companies have treated site selection and power contracts as competitive matters best kept under wraps. When those two positions collide, the result is often a moratorium that halts new construction while officials gather data on their own.
Details
Participants argued that NDAs, routinely used in site selection and power negotiations, prevent local officials and residents from understanding project scale or environmental trade-offs. Without those agreements, companies might share more concrete plans earlier, potentially reducing suspicion. The discussion did not produce agreement on whether transparency alone would satisfy opponents who cite grid strain and noise.
One thread explored the possibility that visible data on energy consumption and water use could shift the conversation from secrecy to measurable mitigation steps. Guests pointed out that once a project is announced after land options are secured, communities feel they have no real input. Earlier disclosure might change that dynamic, though some participants doubted it would be enough to overcome organized opposition.
The podcast also touched on the communication problem itself. Several speakers described how announcements often come through press releases or closed-door meetings rather than open forums. That pattern leaves residents learning about massive power demands or cooling systems only after the local utility has already begun planning upgrades. Removing NDAs would not automatically fix the tone of those conversations, but it would remove one barrier to having them at all.
Reactions / counterpoints
Not everyone on the podcast believed dropping NDAs would move the needle. Some argued that the core objections—electricity costs passed to ratepayers, water consumption in drought-prone areas, and noise from cooling equipment—are substantive and would persist even with full disclosure. Others countered that secrecy amplifies distrust and makes any later mitigation efforts look like after-the-fact damage control.
Why it matters
Local resistance now directly affects timelines for new capacity that AI developers need. If NDAs remain standard, communities will continue to encounter projects only after land deals close, leaving little room for negotiation. Removing them could force earlier public scrutiny and force companies to address concerns about electricity demand and infrastructure costs before construction begins.
The alternative is more moratoriums that slow hardware deployment and raise the effective cost of scaling models. The podcast left open whether the industry will test the approach or continue treating site details as competitive secrets. Either path carries clear consequences for how quickly new data centers can come online and how much those delays will add to the overall expense of training and serving larger models.
The debate also highlights a broader pattern. AI companies have moved quickly to secure power and land, assuming technical momentum would carry the day. When that assumption meets local permitting processes, the result has been friction that no amount of later explanation has fully resolved. Changing the NDA practice would not solve every objection, but it would alter the starting point of those conversations from one of surprise to one of measurable trade-offs.
---
Sources:
No comments yet