OpenAI tests ChatGPT ads as Anthropic, Zhipu and Thinking

OpenAI will begin testing advertisements inside ChatGPT next month, marking the first broad insertion of paid placements into its flagship conversational product as the company seeks new revenue streams amid rising costs. The initial trial will target free-tier users and test a small set of ad formats in the chat interface.

Sequoia Capital is in advanced talks to invest in Anthropic, representing another major venture firm move into large language model startups. Separately, Zhipu AI announced a partnership with Huawei to run its models on domestic Chinese chips, a strategic shift intended to reduce dependence on US semiconductor suppliers. Thinking Machines, a smaller generative AI startup, experienced public leadership turmoil this week as executives and board members clashed over strategy and governance.

OpenAI and ChatGPT ads

OpenAI's ad trial is framed as a pragmatic step to diversify revenue beyond subscription plans and API fees. The company has been operating at a loss while spending heavily on model training and infrastructure. The test will be limited in scope and duration: ads will appear for a subset of free users, and formats will include contextual display units integrated near the chat window rather than full-page takeovers.

Industry observers expect the experiment to measure both short-term revenue uplift and longer-term effects on user engagement and retention. Advertisers may gain access to conversational placements that can be contextually targeted by user queries. At the same time, privacy advocates and some users are likely to question data use and the mixing of ad content into an assistant experience.

OpenAI faces a balancing act: monetize a massive user base without degrading the product that drove adoption. Any negative user reaction could push more people toward paid tiers or alternative chat providers, altering the tradeoffs between ad revenue and subscription growth.

Funding moves, chip geopolitics and startup drama

Sequoia's expected investment in Anthropic would continue a pattern of major venture funds re-entering the AI core infrastructure market after a year of selective deals. The capital would support Anthropic's continued model development and commercial expansion, and it signals persistent investor appetite for differentiated safety-focused model teams.

Zhipu AI's collaboration with Huawei centers on deploying language models on domestically produced accelerators and system software. The partnership aims to create production-capable stacks that rely less on US-designed GPUs and toolchains. For Chinese model providers, that path lowers regulatory and supply-chain risk but raises engineering challenges tied to performance and compatibility.

Thinking Machines' leadership dispute spilled into public view after conflicting statements from executives and board members about hiring, product timelines and fundraising. The friction highlights the governance risks smaller AI startups face as they scale technical teams and seek follow-on capital.

Why it matters

OpenAI testing ads in ChatGPT signals a shift toward ad-supported conversational experiences that could reshape pricing and product segmentation across the industry. Sequoia's move toward Anthropic and Zhipu's Huawei tie-up illustrate divergent responses to capital and supply-chain pressure: more investment into safety-centered models in the West, and a push for hardware independence in China. Governance battles at startups such as Thinking Machines show the fragility of early-stage teams under rapid growth and investor scrutiny.