Keep pulling the thread on AI Defense.
CoreWeave is larger than the next 15 competing neoclouds combined.
CoreWeave exited the previous year with a revenue run rate equivalent to a Fortune 500 company.
CoreWeave provided guidance that it will exit the current year with a revenue run rate equivalent to a Fortune 300 company.
CoreWeave projects it will achieve a revenue run rate equivalent to a Fortune 100 company by the end of next year, approximately 18 months from now.
CoreWeave has acquired the MLOps platform Weights and Biases.
Cisco's AI Defense product was written entirely by AI coding agents.
By the end of the current year, it is likely that a combination of AI models and agents will be able to accomplish a full 24 hours of human work independently.
Anthropic is expanding its product offerings up the stack to compete directly with application-layer companies, such as Cursor, that are built on its models.
Healthcare AI startup Exacare grew its revenue run rate from under $400,000 to $16 million between the start of 2025 and the end of February 2026.
Andrej Karpathy built and released an "auto researcher" AI agent capable of autonomously building a GPT-2 model from scratch.
Tobi Lütke used an AI agent to adapt Andrej Karpathy's auto-researcher project for a different task, resulting in a new model with 19% better performance within 24 hours.
Amazon Web Services' initial go-to-market strategy for foundation models involved launching with Anthropic because Amazon lacked its own competitive frontier model.