Amazon Circumvents Gilroy Community Vote for AI Data Center Using 45-Year-Old Rules

Amazon Web Services (AWS) has started construction on a 56-acre AI data center in Gilroy, California, circumventing a community vote by adhering to local zoning rules set 45 years ago. The project, located between a Walmart Supercenter and the Gilroy Premium Outlets, has been in the approval process since 2020, with the public comment period ending in 2024—before data centers became a hot topic nationwide. This has left residents like Rosa Rodriguez locked out of the discussion, despite concerns about drought and environmental impact.
Key Details
- Location: 56-acre farmland in Gilroy, CA, about 30 miles southwest of San Jose.
- Zoning loophole: The site adheres to industrial zoning rules established 45 years ago, which do not require a public vote for projects of this scale.
- Approval timeline: Application started in 2020; permitting was delayed at least one year due to disagreements between Amazon and city officials. The public comment period closed in 2024.
- City position: Gilroy City Mayor Greg Bozzo stated that other similar-scale industrial projects, like a food-distribution center, have been approved without resistance.
- Amazon response: Roger Wehner, AWS VP of economic development, said the site went through a long approval process with public notices and comment periods.
This case highlights how legacy zoning rules can allow large infrastructure projects to bypass modern community engagement. For developers tracking AI data center expansion, this is a precedent that could affect planning and community relations in other regions.
📖 Read the full source: HN LLM Tools
👀 See Also

Claude Code System Prompts v2.1.51/52: New Prompts, SDK Updates, and GA Features
Claude Code system prompts v2.1.51 and v2.1.52 add six new prompts, update SDK/API references across seven languages, and promote code execution and memory to GA. The Python Agent SDK has been reworked with async changes and new interfaces.

Anthropic-xAI Compute Deal: Beyond Claude Code Limits
Anthropic signed a 300MW / 220k GPU compute deal with competitor xAI. This signals tighter GPU supply and structural cross-lab compute sharing, with implications for inference pricing and multi-provider routing.

Training GPT-1 on an RTX 2060 Super 8GB – Proof of Concept for Local Pretraining
A developer trained the original GPT-1 (117M params) on a local RTX 2060 Super 8GB VRAM in 1 hour, proving that anyone with a gaming GPU can pre-train a 1B-sized model. Code and weights are open-sourced.

OpenClaw 2026.4.2 and 2026.3.31 break local LLM connections
OpenClaw versions 2026.4.2 and 2026.3.31 are causing connection timeouts to locally hosted Ollama instances. The issue appears when connecting to Ubuntu boxes running locally, with error logs showing LLM request timeouts and failover decisions.