AI agents flood public services — and the claims are mostly real
AI made filing a complaint nearly free, and public systems built on friction are now finding out what that does to their workload. Plus: a $1 billion bet on taking data centers off the grid, and a robotics startup stepping out of stealth with $100 million — and robots already on warehouse floors.
Public services are getting swamped — and the flood isn't spam. Complaints to the UK's housing ombudsman more than doubled after ChatGPT arrived, rising from about 2,600 in 2022 to just over 7,000 last year. The US Consumer Financial Protection Bureau saw complaints grow fivefold over the same period, and Brazilian judicial petitions and German parliamentary petitions show similar jumps, according to research by AI researcher Chris Schmitz.
Why it matters: the obvious read is AI slop — machines drowning agencies in noise. Schmitz's finding cuts the other way: most new filings come from real people with legitimate claims who finally cleared the hassle barrier. "The vast majority of cases we find are people who are entitled to claim for something, claiming for that thing," he told TechCrunch. What once took hours of careful prompting is now, in his words, a matter of "taking a photo of a letter with your Claude app and getting a pretty good response in one shot."
Our take: public systems quietly priced in deterrence-by-hassle — if complaining is painful enough, only the desperate or determined bother. When AI drops that cost to near zero, agencies face a triage problem, not an authenticity problem. Schmitz argues this is the moment to "rethink pretty much everything about how this process looks." Bug bounty programs met the same flood and got worthless machine-generated reports; the public-service version is stranger and more hopeful — the queue is full of people who were always entitled to be there.
Data centers are going off-grid — Spark Capital just paid up to prove it. TAR, a startup building off-grid modular power systems for AI data centers, raised a $120 million Series A led by Spark Capital at a valuation of roughly $1 billion, Bloomberg reported. Founded by Pat Becker and Leonhard Soenke, the company combines renewable generation and battery storage for gigawatt-scale facilities, with an automation stack that runs everything from site selection to commissioning.
Why it matters: the grid interconnection queue has become the binding constraint on AI buildout — chips get ordered years out, but plugging them in can take longer. A nine-figure round at a $1 billion valuation for a company whose pitch is essentially "skip the utility" says investors now see power, not compute, as the bottleneck.
Maven Robotics leaves stealth with $100 million — and robots already on warehouse floors. The startup, founded by former Apple autonomy engineers, builds general-purpose wheeled robots that travel up to 10 mph and lift up to 30 kg with two arms. Its $100 million Series A lands with active deployments in real facilities, and comes on top of roughly $18 million raised across its seed and accelerator programs.
Why it matters: most robotics rounds buy a roadmap; Maven's arrives after the prototype stage, where most hardware startups die. General-purpose industrial robots that work in messy human spaces rather than fenced cells remain the sector's hardest open problem, and a lean team with machines already deployed is the strongest crossover signal physical AI has given this month.
What to watch: whether agencies start redesigning intake for the AI era rather than just filtering it — and whether TAR's off-grid model and Maven's deployed fleet turn today's validation into signed contracts.
Would you file that complaint you've been sitting on if a chatbot did the paperwork — and should agencies have to absorb the result? Tell us in the comments.
