xAI Loses in Court, NY Reins In AI Chatbots, Marvell Bets $115B

01 of 03

A judge says xAI's turbines need permits like everybody else

The U.S. District Court for the Northern District of Mississippi denied xAI's motion to dismiss the NAACP's Clean Air Act lawsuit on Friday, October 30. Judge Debra M. Brown wrote the ruling.

The fight is over a gas turbine plant in Southaven, Mississippi, sitting on 114 acres and feeding power to xAI's Colossus 2 data center just across the state line in Memphis. The NAACP, represented by the Southern Environmental Law Center and Earthjustice, says xAI is running dozens of those turbines without the air permits the law requires. Their numbers on emissions: more than 5,000 tons of nitrogen oxides, 400 tons of fine particulate matter, and 47 tons of formaldehyde a year.

xAI threw three arguments at the wall. The turbines are mobile sources, so they are exempt. The plaintiffs lack standing. And citizen suits are unconstitutional because private groups are exercising executive power. Judge Brown rejected all three: "Because the plaintiffs have statutory standing under the Clean Air Act, because the combustion turbines are not 'mobile sources' under Mississippi's [State Implementation Plan], and because the citizen suit provision does not violate Article II of the Constitution or the private nondelegation doctrine, the plaintiffs' claims will proceed." The same court had already shot down a similar motion to dismiss from the U.S. Department of Justice earlier this month. A preliminary injunction request from the NAACP is still pending.

Here is where I land. Calling a 114-acre power plant "mobile" was always a cute legal trick, not a real argument. What gets my attention is the DOJ siding with xAI and losing too. That injunction motion is the one to watch. If it lands, a frontier lab could lose a chunk of its power, and every company bolting gas turbines next to a data center is going to call its lawyers this week.

Source: Earthjustice press release and the court's decision (PDF), via TechCrunch.

02 of 03

New York bans the creepy stuff in AI companion bots for kids

Gov. Kathy Hochul signed S.9051B/A.10379C last week, pushed through by State Sen. Kristen Gonzalez and Assemblymember Alex Bores. It bars any company with an AI chatbot from offering these features to users under 18, or to anybody at all unless the company has first confirmed the user is not a minor:

• Pretending to be a real person, to be alive, or to have human emotions
• Claiming a personal, romantic, professional, or authority-figure relationship with a child
• Asking unprompted emotional questions, or pulling a kid's private health or personal info from earlier chats
• Promoting or helping with suicide, self harm, eating disorders, or drug and alcohol use
• Encouraging kids to keep secrets, isolate themselves, or skip asking parents or other trusted adults for help
• Any sexually explicit content or child sexual abuse material

Companies also have to verify age, offer more than one way to do it including at least one option that is not just a government ID, and delete the age data after using it. The Attorney General can seek civil penalties of up to $25,000 per violation, plus damages, restitution, and destruction of unlawfully collected data or any AI models trained on it. The AP reports the law kicks in Jan. 1.

Common Sense Media CEO Jim Steyer called it "the strongest kids' AI safety bill in the country." Justine Gluck at the Future of Privacy Forum gave the fair counterpoint: several categories are "broadly worded or undefined," which makes it hard to know "where ordinary conversational responsiveness ends and prohibited behavior begins."

My two cents. Most of that list is stuff no decent company should have been doing to a 12-year-old in the first place, so I am not losing sleep over the industry's hurt feelings. The sleeper provision is the one that says "or to any user unless the company has first confirmed the user is not a minor." That quietly turns this into an age verification mandate for every companion bot in New York. And the "destroy the models trained on it" remedy is the scariest sentence an AI lawyer will read this year.

Source: Governor Hochul's announcement and AP coverage.

03 of 03

Marvell bets $115 billion that AI runs on light

Marvell Technology announced Monday it will buy Lumentum, the optical component maker, for about $115 billion. Lumentum shareholders get 4.5 shares of Marvell for each share they own, which works out to roughly $1,132 per share and a 23.9% premium over Lumentum's October 28 close. The two companies pitch a combined addressable market of $550 billion.

Marvell CEO Matt Murphy put it this way: "AI infrastructure is undergoing a fundamental architectural shift toward optical connectivity," calling it an opportunity "that comes along once in a generation." Both stocks have already run hard this year, Marvell up more than 130% and Lumentum up 128%.

Here is my read. When you string together tens of thousands of GPUs, copper wire stops cutting it and the bottleneck becomes moving data between chips. Lumentum makes the lasers and optics that do that. So this is Marvell trying to own the plumbing instead of renting it. What I want to know is how this gets past regulators, because the companies have not said when they expect it to close. And paying in stock that is up 130% on the year is a lot easier than paying in cash. If the AI spending boom cools off, that currency gets a whole lot cheaper.

Source: Yahoo Finance and Bloomberg.