Claude Code Red: Anthropic Cuts Off OpenAI’s Access Amid Model Rivalry
Anthropic has decided to cut off OpenAI’s access to its Claude AI models, citing violations of its terms of service. The move comes after OpenAI reportedly used Claude’s coding tools to compare them against their own models in preparation for the upcoming GPT-5 release. Anthropic emphasized that OpenAI’s actions went against their commercial terms, which prohibit using their models to build or enhance competing products.
This decision highlights the increasing competition in the AI industry, where access to cutting-edge models like Claude is critical for staying ahead. While Anthropic is still offering OpenAI limited access for safety evaluations, the full extent of the restrictions remains unclear, further intensifying the rivalry between AI giants.
OpenAI Raises $8.3B, Valued at $300B
OpenAI has raised $8.3 billion in its latest funding round, pushing its valuation to $300 billion, surpassing major companies like AMD, Coca-Cola, and General Electric. The funding, which came in ahead of schedule, is part of the company’s plan to raise $40 billion this year. Key investors include Dragoneer Investment Group, Blackstone, TPG, Fidelity, and other top firms.
With annual recurring revenue hitting $13 billion, OpenAI is on track to reach $20 billion by the end of the year. The company now has 5 million paying business customers for its ChatGPT product. This financial boost comes at a crucial time as OpenAI faces intense competition, ongoing legal battles, and restructuring efforts related to its transition from a non-profit to a for-profit organization.
Amazon’s New Plan: Ads in Alexa Conversations
Amazon’s CEO Andy Jassy recently shared plans to introduce ads in Alexa+ during the Q2 earnings call. He mentioned that as more users interact with Alexa+, ads could become a new revenue stream and a way for product discovery. These ads could appear during shopping recommendations or multi-step conversations. Alexa+ is free for Prime members, while non-Prime users will need to pay a $19.99 monthly subscription.
Ads on Alexa have been limited so far, such as visual ads on Echo Show devices or pre-recorded ads on smart speakers. However, the concept of AI-generated ads in Alexa+ is new. These ads could help users discover new products, but there may be trust issues if users are not informed about which suggestions are sponsored. Competitors like Google and OpenAI are also considering adding ads to their AI assistants, though the model is still unclear.
Meta’s $72 Billion AI Infrastructure Investment in 2025
Meta Platforms Inc. is set to invest up to $72 billion in AI infrastructure for 2025, a significant jump from previous years. CEO Mark Zuckerberg stated that this investment is vital for developing cutting-edge AI models like Llama 4 and Behemoth, which aim to push the boundaries of superintelligence. The company plans to build a vast data center with a capacity of over 2 gigawatts, equivalent to the size of Manhattan, and deploy around 1.3 million GPUs by the end of the year. This is part of Meta’s strategy to compete with tech giants like Google and Amazon in the escalating AI arms race.
This massive investment reflects Meta’s commitment to staying at the forefront of the AI revolution, but it has raised concerns among investors about the long-term profitability of such spending. Despite these worries, Meta’s stock rose by 8.5% after the announcement, showing strong investor confidence in the company’s AI strategy. As the competition in AI heats up, Meta’s aggressive approach highlights the high stakes of securing leadership in this rapidly growing field.
Meta’s Shift: Superintelligence AI Models May Not Be Open-Source
Meta CEO Mark Zuckerberg has indicated that the company might not open-source all of its advanced AI models, particularly those related to superintelligence. In a recent communication, Zuckerberg acknowledged the potential benefits of superintelligence but emphasized the novel safety concerns it introduces. He stated that Meta would need to be rigorous in mitigating these risks and be cautious about what it chooses to open-source.
This marks a significant shift from Meta’s previous stance. Since 2022, Meta has made its powerful Llama AI models available as open-weight, allowing developers broad access. However, with the emergence of AI systems that can potentially improve themselves, Zuckerberg highlighted the need for a more cautious approach. He noted that as AI models become more powerful, the safety concerns become more serious, and it may not be responsible to open-source them. This change reflects a balancing act between fostering innovation and ensuring safety in the rapidly evolving field of AI.