😸 Anthropic’s $30B War Chest, the $200M Political Battle, and the Flood of Free Chinese Models

2026-02-12 · Source: The Neuron · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Emerging Technologies & Innovation, Corporate Strategy & Leadership · Depth: Novice, long

Summary

Anthropic recently secured a historic $30 billion funding round, making it the second-largest private tech raise after OpenAI's $40 billion+ round. Both companies are now engaged in a political "arms race," with Anthropic contributing $20 million to a Super PAC advocating for AI regulation, while OpenAI co-founder Greg Brockman invested $25 million in a PAC opposing government intervention. Collectively, AI companies have committed over $200 million to the 2026 midterms. This week also saw significant model releases, including Google's Gemini 3 Deep Think, which excels in math, science, and coding benchmarks; MiniMax M2.5, an open-weight Chinese model matching Claude Opus on coding at 10-20x lower cost; OpenAI's GPT-5.3-Codex-Spark, offering rapid code edits; and GLM-5, another open-weight, MIT-licensed Chinese model that is 5-8x cheaper than Opus and ranks highly in agent and coding benchmarks.

Key takeaway

For AI Engineers evaluating model deployment strategies, the emergence of high-performing, cost-effective open-weight models from China, alongside advancements in proprietary models, necessitates a re-evaluation of your current stack. You should explore integrating diverse models like MiniMax M2.5 or GLM-5 for specific tasks to optimize performance and reduce operational costs, rather than relying solely on established closed-source solutions.

Key insights

The AI landscape is marked by massive investments, political influence, and rapid advancements in both open-source and proprietary models.

Principles

• Open-weight models offer significant cost advantages.
• AI development involves substantial political lobbying.
• Diverse models excel in different specialized tasks.

Method

The article highlights the potential to swap between various AI models (e.g., via Openrouter or Openclaw) in workflows, leveraging their distinct strengths for tasks like coding or complex reasoning.

In practice

• Explore open-weight models like MiniMax M2.5 for cost-effective coding.
• Utilize Gemini 3 Deep Think for complex math and science problems.
• Consider GPT-5.3-Codex-Spark for instant code iteration.

Topics

• AI Funding & Investment
• Large Language Models
• AI Agents & Automation
• AI Policy & Geopolitics
• AI Safety & AGI

Code references

• openclaw/openclaw

Best for: AI Engineer, Machine Learning Engineer, NLP Engineer, AI Product Manager, Tech Journalist, General Interest

Related on AIssential

• not much happened today — The AI landscape is rapidly evolving with open-weight models nearing frontier capabilities, while agentic systems and specialized benchmarks drive practical application and efficiency.
• 😸 Anthropic and OpenAI Both Dropped Their Best AI Models On The Same Day — Leading AI developers are advancing models towards autonomous job execution, emphasizing either broad utility or deep specialization.
• 🗞️ Alibaba released their open-source Qwen3.5-9B model, and it holds its own against OpenAI's 120B param gpt-oss system. — Smaller, efficient AI models are achieving performance comparable to much larger systems through architectural innovations and multimodal training.
• Chinese AI lab Zhipu releases GLM-5 under MIT license, claims parity with top Western models — GLM-5 is an open-source, 744B-parameter Chinese AI model claiming parity with top Western models on agent and coding tasks.
• Z.ai GLM-5: New SOTA Open Weights LLM — Chinese AI labs are rapidly advancing open-weight models, emphasizing cost-efficiency, long-context, and agentic capabilities.
• There has been a situation in AI — GLM52 marks the arrival of a true open-source frontier AI model, challenging the perceived "moat" of proprietary systems like GPT-5.5 and Opus 48.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by The Neuron.