Core Automation, a new artificial intelligence startup founded by former OpenAI Vice President of Research Jerry Tworek, is reportedly seeking between $500 million and $1 billion in fresh funding, just weeks after its formation. The ambitious raise reflects growing investor appetite for next-generation AI systems that move beyond today’s static large language models.

Tworek brings formidable credentials. During his seven-year tenure at OpenAI, he contributed to some of the lab’s most influential breakthroughs, including Chinchilla scaling laws, early coding systems such as Codex and ChatGPT, reinforcement learning for robotics, and advanced reasoning models like o1 and o3 that topped LiveBench benchmarks. He exited OpenAI in January 2026, citing limits on pursuing high-risk research within large organizations.

Core Automation’s core thesis is continual learning. Unlike conventional LLMs that require periodic retraining, the startup aims to build AI systems that learn continuously from real-world interactions. These self-improving agents are designed for enterprise automation, software development, and robotics—areas where adaptability and long-term learning are critical.

The funding target implies a post-money valuation exceeding $5 billion. In today’s capital-intensive AI race, such a war chest could finance frontier-scale training runs rivaling top labs. Potential backers are expected to include elite venture firms with prior ties to Tworek.

If successful, Core Automation could help solve persistent challenges such as model staleness, hallucinations, and slow adaptation. However, competition for compute, talent, and energy remains fierce.

The startup’s progress will test whether elite individual researchers can still leapfrog established AI giants—and whether continual learning can unlock the next phase of intelligent automation.