Anthropic Announces Claude for Teachers, DeepSeek Faces Funding Crunch, New York Freezes Data Center Construction — AI News Digest

Anthropic Announces "Claude for Teachers" — Bringing AI to Education in Earnest

On July 14, Anthropic announced "Claude for Teachers," a new plan for educational institutions. It reportedly helps teachers streamline lesson preparation and grading, and includes features to support AI literacy education for students. Google and Microsoft are also investing heavily in AI for education, but Anthropic appears to be differentiating itself by putting safety front and center.

Detailed features and pricing will be announced over time.

DeepSeek Short on Funds Just After Raising $7 Billion — A Reflection of Soaring AI Development Costs

It has come to light that DeepSeek, a Chinese AI developer, needs additional funding just a few weeks after securing $7 billion in its first major funding round. The cost of training and operating AI models is climbing faster than originally projected, laying bare the reality that even enormous fundraising makes it hard to keep up in the development race.

DeepSeek has drawn industry attention with its low-cost, high-performance models, but as the race to develop next-generation models intensifies, building a sustainable financial foundation has become a key challenge.

New York State Fully Freezes New Data Center Construction

New York State has decided to impose a comprehensive freeze on the construction of new data centers. The move is a regulatory response, from the standpoint of power consumption and environmental impact, to the surge in data center demand driven by the AI boom. Because data centers consume enormous amounts of electricity and water, their impact on local power grids and the environment has become a growing concern in many regions.

This decision stands as a symbolic case of local governments' wariness toward the rapid expansion of AI infrastructure. Similar moves could spread to other states and countries.

Meta Used AI to Select Layoff Targets Based on Medical Information — Sparking Ethical Debate

According to a Reuters report, testimony from former employees has revealed that Meta (formerly Facebook) used AI to analyze employees' medical information and narrowed down workers with medical conditions as layoff targets. Using AI for personnel evaluation and downsizing decisions is not uncommon in itself, but the inclusion of medical information has triggered a new wave of ethical debate.

While AI adoption in HR advances on the efficiency front, risks of discrimination and privacy violations have been pointed out, and regulatory scrutiny is expected to intensify.

An Open-Model Rush Is Here — Kimi K3, DeepSeek V4, and a New GLM Model Lined Up for Release

The open-weight AI model scene is buzzing. Moonshot AI's "Kimi K3" is expected to be released within hours, and multiple sources report that general availability (GA) of DeepSeek V4 will begin this week. Moreover, rumors indicate that a new GLM model release is drawing near, with some observers expecting GLM 5.5 to arrive in August.

There are also reports that an open coding model called KAT-Coder-Air V2.5 will be released soon, suggesting that competition among coding-focused models is about to heat up as well.

With multiple major models being released in a short period, users have no shortage of options — but it is also becoming harder to decide which model to use.

NVIDIA Announces "Nemotron Labs" — Customizable Open Models for Enterprises and Nations

NVIDIA has announced "Nemotron Labs," an initiative that enables companies and countries to build and customize AI models under their own control. Built on open models, it aims to meet requirements for security, reliability, and custom tuning.

For government agencies that prioritize AI sovereignty and companies that cannot hand their data to outside parties, this NVIDIA approach could prove to be an attractive option. NVIDIA has traditionally centered its business on supplying GPUs, but its move to deepen involvement in the model layer is becoming increasingly clear.