Regulation

AI-generated text

Amazon will drop NDAs for data center deals as AI agent startups seek sensitive access

Amazon announced it will stop using nondisclosure agreements (NDAs) when negotiating data center agreements with local governments, following a similar step by Microsoft earlier this year.

Amazon will drop NDAs for data center deals as AI agent startups seek sensitive access

Amazon has announced it will cease using nondisclosure agreements (NDAs) when negotiating data center deals with local governments. The move follows a similar decision by Microsoft earlier this year. The change responds to rising demands for transparency: secrecy around data center plans has provoked community backlash against AI infrastructure deployments.

Secrecy and behind-the-scenes negotiations have led to opposition across multiple U.S. jurisdictions, prompting proposed or enacted moratoriums on data center construction from New York to San Francisco. The shift away from NDAs is intended to reduce local mistrust and improve the public visibility of infrastructure plans.

Personal AI agents raise trust and permission issues

At the same time, a wave of startups is betting consumers will grant personal AI agents access to their inboxes, files, and credit cards — provided websites and services permit such integrations. On the latest episode of TechCrunch’s Equity podcast, hosts Anthony Ha, Sean O’Kane, and Rebecca Bellan examined whether trust will become the primary product for these personal AI companies.

The discussion highlighted startups such as Tab, Underdog, and Hark pitching privacy as a core offering, while also exploring business models that could turn agents into salespeople acting on users’ behalf. Meanwhile, platform decisions will matter: companies including United, Apple, and Amazon are erecting barriers to outside agents, which will shape how widely these agent models can spread.

Other topics covered on the Equity episode

The podcast also covered several related technology and corporate developments:

  • The White House’s plans and coordination around a so-called “Super Intelligence Force.”
  • Ghost’s $3,499 computer marketed to run AI agents continuously, which the hosts argued is unlikely to achieve mass-market adoption.
  • Reports of Lambda raising about $4 billion in funding and an assertion that a single customer accounts for $35 billion of its backlog.
  • Uber’s $2.3 billion all-cash acquisition of ezCater, signaling an expansion into office catering.

Conclusion

Amazon’s decision to abandon NDAs in local data center negotiations signals that major cloud providers are responding to public pressure for greater transparency in AI infrastructure decisions. At the same time, the prospects for personal AI agents depend heavily on user trust and the access policies of platforms and websites. The TechCrunch Equity episode explored these tensions alongside several significant corporate moves in the AI and cloud space.

The podcast also promoted upcoming coverage and events, including TechCrunch Disrupt 2026, and offered listeners a discount code for tickets.