Reflection AI Founded by Ex-DeepMind Researchers Seeks $1 Billion to Rival Meta, DeepSeek Amid $155B AI Spend by Tech Giants

Reflection AI, a one-year-old startup founded by former DeepMind researchers, is in advanced talks to raise over $1 billion to develop open-source large language models (LLMs) aimed at competing with established AI projects from Meta, DeepSeek, and Mistral. The company, known for creating the AI agent Asimov, has reportedly secured most of the targeted funding. This development comes amid a robust investment climate in artificial intelligence, with major technology firms including Meta, Microsoft, Amazon, and Alphabet collectively spending $155 billion on AI in 2025 and projecting $400 billion in capital expenditures next year to enhance AI infrastructure and data centers. Meanwhile, OpenAI has reached 700 million users and secured $8.3 billion in funding at a $300 billion valuation. Despite its rapid growth and an expected $20 billion revenue by the end of 2025, OpenAI faces an $8 billion cash burn and increasing competition from AI initiatives backed by Google, Meta, and Microsoft.