The $100 Billion Round: Nvidia, Microsoft, and Amazon Join the OpenAI Surge

Reports from The Information and Bloomberg today indicate that OpenAI is finalizing a historic funding round that could reach $100 billion. Following SoftBank’s $30 billion commitment earlier this week, Nvidia is reportedly in talks to invest $30 billion, while Microsoft and Amazon are lining up multi-billion dollar stakes.

• The Valuation: This would value OpenAI between $750B and $830B.

• The Cloud Catch: Amazon’s deal reportedly includes a $38 billion commitment for OpenAI to use AWS cloud power over the next seven years, signaling a shift in OpenAI's hardware dependency. Read more: Nvidia, Microsoft, Amazon in talks to invest up to $60 billion in OpenAI

EU Formally Opens "Specification Proceedings" Against Google

Following yesterday's warnings, the European Commission has officially opened two sets of proceedings under the Digital Markets Act (DMA). The EU is mandating that Google provide "free and effective" interoperability for third-party AI developers on Android.

• Leveling the Field: The goal is to ensure that a user can set Claude, ChatGPT, or an open-source model as their "System AI" with the same hardware permissions (camera, microphone, and screen-reading) that Google’s Geminicurrently enjoys.

• Data Access: Google is also being forced to share anonymized search ranking and query data with rival AI chatbot providers. Read more: Commission opens proceedings to assist Google in complying with DMA obligations

Gartner Predicts "AI Sovereignty" Will Consume 1% of Global GDP

A major report released today by Gartner predicts that by 2027, 35% of countries will be locked into region-specific, sovereign AI platforms. Driven by fears of Western data monopolies, nations are increasingly building domestic "AI stacks" aligned with local laws and cultures. Gartner warns that countries failing to establish a sovereign AI stack may be forced to spend at least 1% of their GDP on foreign AI infrastructure by 2029 to keep up. Read more: Gartner Predicts 35% of Countries Will Use Region-Specific AI by 2027

Model Launch: Arcee AI Debuts "Trinity Large" 400B MoE

US-based startup Arcee AI has disrupted the frontier model space with the release of Trinity Large, a 400-billion-parameter sparse Mixture of Experts (MoE) model. Despite its massive size, it utilizes a "sparse" architecture (only 13B active parameters per token), allowing it to run 2-3x faster than Meta’s Llama 4 or Google’s Gemini 2 on similar hardware. It was trained in just 33 days using 2,048 Nvidia B300 GPUs.

Read more: Arcee AI Launches Trinity Large, a 400B Sparse MoE AI Model

Prompt of the Day: The "Sovereign AI" Localizer

With the rise of "Sovereign AI" (per the Gartner report), models are becoming more specialized for specific regions. You can use this prompt to see how a "Global" model (like Gemini or GPT) handles local nuances compared to how a specialized regional model might.

The Prompt:

"I am drafting a legal/business proposal for a project in [Insert Region, e.g., Switzerland or Singapore]. Analyze the following text for 'Cultural and Regulatory Friction.' Identify three phrases that might be interpreted differently under local labor laws or social customs in that specific region, and suggest 'Sovereign-AI' style adaptations that align with [Specific Local Regulation, e.g., Swiss Data Protection or Singapore's AI Ethics Framework]."