Welcome to Daily Zaps, your regularly-scheduled dose of AI news ⚡️ 

Here’s what we got for ya today:

• 💻 Largest AI data center coming to Texas

• ❌ U.S. eyes AI chip export limits

• 🤖 Anthropic updates policy to address rogue AI

• ✅ EU AI Act compliance checker tool

Let’s get right into it!

Sign Up | Advertise | Tools

STARTUPS

Largest AI data center coming to Texas

The surge in demand for artificial intelligence is propelling the growth of nimble data center startups, notably Crusoe Energy, which recently secured a $3.4 billion loan to develop a massive data center in Texas, projected to be one of the largest globally. This facility will host Nvidia's GB200 AI chips for OpenAI, with plans for around 50,000 chips to be operational by early 2025, expanding to 100,000 by fall 2025.

Crusoe, founded in 2018 and previously focused on cryptocurrency mining, is raising equity financing led by Founders Fund and aims to enhance its valuation from $1.4 billion to over $2 billion.

GOVERNMENT

U.S. eyes AI chip export limits

Nvidia's stock saw a slight decline in pre-market trading, dropping about 5% by midday due to concerns over potential U.S. restrictions on the sale of advanced AI chips to certain countries. The Biden administration is reportedly considering capping export licenses for these chips, citing national security issues.

Despite recent volatility, Nvidia’s stock has risen approximately 173% this year and remains the second-most valuable public company after Apple. Meanwhile, Dell is set to launch servers using Nvidia’s Blackwell AI chips next month.

BIG TECH

Anthropic updates policy to address rogue AI

Anthropic has updated its Responsible Scaling Policy to enhance safety protocols for its AI systems, introducing specific Capability Thresholds that trigger additional safeguards as models grow more powerful. This proactive approach aims to prevent misuse in high-risk areas, such as bioweapons and autonomous AI research.

The policy also outlines expanded responsibilities for the Responsible Scaling Officer, emphasizing accountability. By establishing a tiered safety framework, Anthropic seeks to set a standard for AI governance, encouraging industry-wide adoption of similar safety measures.

GLOBAL NEWS

EU AI Act compliance checker tool

AI systems, including those from Meta and OpenAI, fall short of upcoming European Union regulations, particularly regarding cybersecurity resilience and discriminatory outputs. The testing framework, designed to align with the EU's AI Act, scores models on various criteria, with many scoring above 0.75; however, significant gaps were identified.

The LLM Checker tool will be available to developers for assessing compliance, and while the European Commission has not verified the tool, it acknowledges it as a crucial step toward implementing the new laws.