AI's newest billionaires, SoftBank's & Meta's new acquisitions & more (December 30, 2025)

Good morning,

AI money is moving fast, regulation is tightening, and content quality is under real pressure. Today’s stories show where capital, control, and credibility are colliding in the AI boom.

Let’s dive in 👇

━━━

Want to engage in conversation about AI news? Join the Byte-Sized Reddit Community

🚀 Big Money & Big Moves

🧠 SoftBank circles $4B AI infrastructure bet

SoftBank nears deal with data center as it moves closer to investing in DigitalBridge, a major data center firm. The deal reinforces SoftBank’s thesis that compute, not apps, is the real AI choke point. Infrastructure is quietly becoming the most defensible layer in the stack.

🤖 Meta buys Manus, the AI startup everyone watched

Meta just bought Manus, an AI startup (for $2B) in a move that signals aggressive talent and IP consolidation. Manus had buzz for its agent-style workflows and research velocity. Meta keeping it semi-independent suggests they want speed without bureaucracy.

💰 AI boom creates a new billionaire class

New billionaires are emerging from the AI boom as founders, early employees, and infra investors cash in. Unlike past cycles, much of this wealth is tied to compute, chips, and platforms. AI is minting fortunes faster, and earlier, than previous tech waves.

⚠️ Models, Media & Market Signals

🌏 Alibaba and Abu Dhabi back MiniMax IPO

Alibaba and Abu Dhabi plan to invest ahead of MiniMax’s $600M IPO. The deal underscores rising Middle East influence in global AI capital flows. Sovereign-backed AI plays are becoming a real force.

📺 Over 21 percent of YouTube is now AI slop

More than one fifth of YouTube content is estimated to be AI-generated spam. Platforms are struggling to balance creator tools with quality control. Expect stronger filters, or a serious trust collapse.

📄 New research pushes model reasoning forward

A new arXiv paper explores advanced AI reasoning with techniques aimed at improving reliability and generalization. The work highlights how architectural tweaks can outperform brute-force scaling. Expect more gains from smarter design, not just bigger models.

🛠️ Tools of the Day

→ Resemble Clone – Create realistic AI voice clones for media workflows
→ Adminder – Generate ad-ready videos optimized for performance channels
→ Influcio – AI-powered influencer discovery and campaign management

⚡ Quick Hits

→ China drafts strict AI rules targeting suicide and violence risks
→ OpenAI safety executive outlines new risk mitigation efforts
→ AI data centers are driving electrician job growth
→ ChatGPT integrations expand into DoorDash, Spotify, and Uber
→ Goldman backs massive Texas AI power sites

🧠 TL;DR

Capital is flowing hard into AI infrastructure, not just flashy apps. Governments and platforms are racing to rein in risk, spam, and unintended harm. The next AI phase will reward control over compute, distribution, and trust.

Cheers,
David