AI News Roundup: SoftBank's $40B OpenAI Bet and the Year's Biggest Deals

The final days of 2025 are serving up blockbuster AI news. SoftBank just completed the largest private funding round in history, Meta snapped up the hottest AI agent startup in a deal struck in just 10 days, and regulators on three continents are making moves that will shape how AI develops in 2026.

Here’s what you need to know.

The Big Story: SoftBank Completes $40 Billion OpenAI Investment

SoftBank has finalized its $40 billion investment in OpenAI, making it the largest private funding round ever recorded. The Japanese investment giant wired the final $22-22.5 billion last week to complete the transaction.

This deal gives SoftBank approximately an 11% stake in OpenAI, making it the second-largest investor behind Microsoft (27.5%). The investment valued OpenAI at around $260 billion on a pre-money basis.

For context: $40 billion is more than the entire GDP of many countries. It’s larger than the market cap of companies like Snowflake, Atlassian, or Palantir. Masayoshi Son is betting that OpenAI will define the next era of computing - and he’s putting historic capital behind that conviction.

Combined with Nvidia’s $20 billion acquisition of Groq earlier this week, the AI industry is closing 2025 with over $60 billion in major deals in December alone.

Sources: CNBC, Yahoo Finance, SiliconAngle

Today’s Top Stories

Meta Acquires Manus for $2+ Billion

Meta has acquired Manus, the Singapore-based AI agent startup that became the talk of Silicon Valley after debuting last spring. The deal was struck in approximately 10 days for more than $2 billion.

Manus - which we covered in depth in our analysis of the acquisition - had raised just $75 million in its Series B and hit $100M+ annualized revenue within 8 months of launch. The approximately 100-person team will join Meta, with the technology integrating into Meta AI across Facebook, Instagram, and WhatsApp.

Meta confirmed there will be no continuing Chinese ownership interests following the transaction - a notable clarification given Manus’s Chinese roots and the current geopolitical climate.

Sources: CNBC, TechCrunch, Bloomberg

OpenAI Hiring “Head of Preparedness” at $555K

OpenAI is hiring a new Head of Preparedness with a salary of approximately $555,000 per year plus equity. CEO Sam Altman announced the role on X, warning it will be “a stressful job” where the hire will “jump into the deep end pretty much immediately.”

The role involves running OpenAI’s technical safety strategy, coordinating capability evaluations, threat models, and mitigations across risk domains including cybersecurity, biological/chemical threats, and AI self-improvement.

This is one of the most critical AI safety roles in the industry - and OpenAI is paying accordingly.

Sources: Fortune, Quartz

Google DeepMind Open Sources Gemma Scope 2

Google DeepMind has released Gemma Scope 2, a massive interpretability toolkit that allows researchers to trace how AI models “think.” Positioned as a “microscope for large language models,” the tools let researchers inspect how decisions form inside AI systems rather than treating them as opaque black boxes.

The release involved storing approximately 110 Petabytes of data and training over 1 trillion total parameters. This is a significant contribution to AI safety research and mechanistic interpretability.

Source: Open Source For You

Trump Executive Order to Preempt State AI Laws

The White House issued an executive order titled “Ensuring a National Policy Framework for Artificial Intelligence” that directs the Department of Justice to establish an AI Litigation Task Force within 30 days. The task force will challenge state AI laws deemed inconsistent with federal policy.

This is a significant move toward federal preemption of state AI regulation - potentially nullifying laws in states like Colorado, California, and elsewhere that have passed their own AI governance frameworks.

Sources: White House, Skadden Analysis

China Proposes World’s First Regulations for Human-Like AI

China’s Cyberspace Administration released draft “Interim Measures for the Management of Artificial Intelligence Anthropomorphic Interactive Services” on December 27. These would be the world’s first regulations specifically governing AI with human or anthropomorphic characteristics.

The rules target AI chatbots that could influence users emotionally or impersonate humans - a recognition that as AI becomes more human-like, new guardrails are needed.

Sources: CNBC, Bloomberg

Quick Hits

  • Funding: Octopus Energy is spinning off its AI unit Kraken Technologies at an $8.65 billion valuation following a $1 billion funding round led by D1 Capital Partners - CNBC

  • Regulation: Texas Responsible Artificial Intelligence Governance Act takes effect January 1, 2026, banning AI-driven pornography bots and chatbots impersonating minors - KUT Austin

  • Models: SK Telecom unveiled ‘A.X K1,’ Korea’s first 500B-scale hyperscale AI model with 519 billion parameters - PR Newswire

  • Legal: Six authors filed lawsuits against Anthropic, OpenAI, Google, Meta, xAI, and Perplexity, seeking $150,000 per title per defendant - the first copyright lawsuit against Elon Musk’s xAI - Publishers Weekly

  • Markets: Japanese memory chipmaker Kioxia has become the world’s best-performing stock with shares rising 540% YTD on AI memory demand - Bloomberg

  • Hardware: Chinese scientists at Shanghai Jiao Tong University claim a breakthrough in developing the first all-optical computing chip capable of supporting large-scale AI models (single source - unverified)

  • Enterprise: VCs surveyed by TechCrunch predict enterprises will increase AI budgets in 2026 but concentrate spending on fewer vendors - consolidation is coming - TechCrunch

What This Means

The final week of 2025 reinforces three trends that will define AI in 2026:

1. Capital concentration is accelerating. SoftBank’s $40B into OpenAI, Nvidia’s $20B for Groq, Meta’s $2B+ for Manus - the biggest players are consolidating access to the best talent and technology. The barrier to competing at the frontier keeps rising.

2. AI agents are the battleground. Meta didn’t just acquire a chatbot - they acquired Manus, an autonomous agent platform. OpenAI’s Operator, Google’s Agentspace, and now Meta’s enhanced AI assistant are all racing toward agents that don’t just answer questions but actually do work. This is the shift from AI as a tool to AI as a worker.

3. Regulation is fragmenting and consolidating simultaneously. China is proposing first-of-their-kind rules for human-like AI. The EU is amending its AI Act. The US is trying to preempt state laws. Texas is going live with new requirements. Every major jurisdiction is taking a different approach - but within the US, the federal government is clearly moving to establish national standards.

For businesses, the implications are clear: AI budgets will grow, but expect to spend more with fewer vendors. For individuals, AI agents are ready to take work off your plate - from scheduling to research to web automation. And for everyone, pay attention to the regulatory landscape - the rules are being written right now.

2025 AI Year in Numbers

As we close out the year, here’s where the industry stands:

Metric2025 Value
Total AI Investment$202.3 billion (up 75% from 2024)
AI Share of Global VC~50% (up from 34% in 2024)
US Share of AI Funding$159 billion (79% of global)
San Francisco Bay Area$122 billion (76% of US AI funding)
Generative AI Spending$37 billion (up from $11.5B in 2024)

Sources

Ready to automate your busywork?

Carly schedules, researches, and briefs you—so you can focus on what matters.

Get Carly Today →