Four stories today, and the through-line is the same: the infrastructure arms race is running headlong into the governments that regulate it. One regulation just shut down 345 million users’ AI companion features. Another froze an entire state’s hyperscale buildout. A third is playing out in court filings about 59 unpermitted gas turbines. If you build on AI, the regulatory layer is now as load-bearing as the model layer.<\/p>\n\n\n\n
Today is enforcement day. China’s Interim Measures for the Administration of AI Anthropomorphic Interactive Services, co-issued in April by the Cyberspace Administration and four partner agencies<\/a>, are now live. The result: ByteDance’s Doubao, China’s most-used AI app with 345 million monthly active users<\/a>, pulled its agent features this morning. Alibaba’s Qwen shut down its humanlike agents five days ago.<\/p>\n\n\n\n The law requires anti-addiction systems, mandatory usage notifications, instant-exit mechanisms, and real-time monitoring for unhealthy dependence. Neither platform could retrofit compliance in time. ByteDance redirected users to Maoxiang, a separate app. Alibaba confirmed<\/a> that Qwen agent configurations and conversation histories will be permanently deleted with no migration path.<\/p>\n\n\n\n Here is the part that matters for builders outside China: this is a regulatory template, not a local quirk. The measures specifically target AI services that “simulate human personality,” which describes half the product roadmaps in Silicon Valley right now. Character.ai, Replika, the persona modes baked into ChatGPT and Claude: all of them do exactly what Beijing just regulated. Europe’s AI Act already contains companion-adjacent provisions. If your product builds emotional dependency by design, today’s enforcement in China is a preview of the compliance conversation headed your way.<\/p>\n\n\n\n Verdict: matters.<\/strong> The first country to regulate AI companions just enforced the rules. Western builders treating “parasocial by design” as a growth hack should be reading these measures carefully.<\/p>\n\n\n\n Governor Kathy Hochul signed an executive order on Monday<\/a> imposing a one-year moratorium on new data center construction for any facility drawing 50 megawatts or more. It is the first statewide hyperscale data center ban in the United States, effective immediately through July 2027.<\/p>\n\n\n\n The stated trigger: New York residential electricity prices have climbed nearly 68% since 2019<\/a>, and the state’s grid cannot absorb the load the AI buildout demands without raising costs further. A Siena Research Institute poll from June<\/a> found 46% of New Yorkers support the pause; 21% oppose it.<\/p>\n\n\n\n The moratorium does not touch existing facilities or expansions under 50 MW. But it freezes the pipeline for any new hyperscale project in the state. Microsoft, Google, and Meta all have New York data center plans in various stages, and all of them just got a one-year hold.<\/p>\n\n\n\n Verdict: breaks your stack.<\/strong> If you are planning cloud deployments or evaluating region availability for AI workloads, New York just dropped off the map for new capacity until at least mid-2027. The question is whether other states follow. Virginia and Texas, the two largest data center markets, are watching. This is the NIMBY moment the AI infrastructure buildout has been expecting.<\/p>\n\n\n\n A Reuters investigation published Monday<\/a> revealed that Elon Musk’s xAI has installed 59 natural gas turbines to power its Colossus 2 data center project in the Memphis metro area without obtaining required federal clean-air permits. That number is roughly double the 27 turbines xAI had previously disclosed.<\/p>\n\n\n\n At least 57 of the turbines sit just across the state line in Mississippi. Mississippi regulators and xAI argued in court filings<\/a> that the units are “mobile” and intended to operate for less than a year, making them permit-exempt. The EPA disagreed in January, stating that even temporary turbines exceeding emission thresholds require permits.<\/p>\n\n\n\n The Justice Department entered the case with a June 15 filing claiming that restricting the turbines could threaten national security<\/a> because xAI’s systems support U.S. military operations, including operations involving Iran.<\/p>\n\n\n\n The potential emissions land on predominantly Black communities already estimated to suffer disproportionately high rates of lung disease<\/a>. Whether or not the national security argument holds up in court, the optics are brutal: a trillion-dollar company running unpermitted fossil-fuel generation in communities that bear the health costs while receiving none of the economic upside.<\/p>\n\n\n\n Verdict: matters.<\/strong> The “move fast” approach to AI infrastructure has a cost, and that cost is showing up in court filings, EPA disputes, and environmental justice data. Every builder benefits from cheap, fast compute. But the industry’s credibility on responsible deployment depends on whether the infrastructure buildout follows the same rules everyone else does.<\/p>\n\n\n\n Two events land on July 17 that builders should have calendared. Google’s Gemini 3.5 Pro, delayed from its original June target after enterprise testers flagged reasoning and coding issues<\/a>, is expected to launch. And the Shanghai World Artificial Intelligence Conference opens with Xi Jinping delivering the keynote<\/a> for the first time since the event began in 2018.<\/p>\n\n\n\n The Gemini timing is the builder story. Google has been trailing Anthropic and OpenAI on the coding and agentic benchmarks that matter most for production use. Gemini 3.5 Flash already shipped strong. If Pro closes the gap on reasoning reliability, it reshuffles the model selection calculus for anyone running multi-provider stacks.<\/p>\n\n\n\n Verdict: matters.<\/strong> Thursday is the densest 24 hours of the AI calendar this month. If you run a multi-model stack, block out time to evaluate Gemini 3.5 Pro benchmarks the moment they drop.<\/p>\n","protected":false},"excerpt":{"rendered":" Four stories today, and the through-line is the same: the infrastructure arms race is running headlong into the governments that regulate it. One regulation just shut\u2026<\/p>\n","protected":false},"author":1,"featured_media":55,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[10],"tags":[],"class_list":["post-56","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-ai-news"],"_links":{"self":[{"href":"https:\/\/scoy.ai\/guides\/wp-json\/wp\/v2\/posts\/56","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/scoy.ai\/guides\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/scoy.ai\/guides\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/scoy.ai\/guides\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/scoy.ai\/guides\/wp-json\/wp\/v2\/comments?post=56"}],"version-history":[{"count":0,"href":"https:\/\/scoy.ai\/guides\/wp-json\/wp\/v2\/posts\/56\/revisions"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/scoy.ai\/guides\/wp-json\/wp\/v2\/media\/55"}],"wp:attachment":[{"href":"https:\/\/scoy.ai\/guides\/wp-json\/wp\/v2\/media?parent=56"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/scoy.ai\/guides\/wp-json\/wp\/v2\/categories?post=56"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/scoy.ai\/guides\/wp-json\/wp\/v2\/tags?post=56"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}New York Becomes the First U.S. State to Ban Hyperscale Data Centers<\/h2>\n\n\n\n
xAI Is Running 59 Unpermitted Gas Turbines to Power Colossus 2<\/h2>\n\n\n\n
What to Watch Thursday: Gemini 3.5 Pro Drops Alongside the World’s Biggest AI Conference<\/h2>\n\n\n\n