AI News Roundup – Nvidia and OpenAI pare down investment deal, India hosts AI summit, ByteDance video-generation model worries Hollywood, and more

To help you stay on top of the latest news, our AI practice group has compiled a roundup of the developments we are following.

• • Nvidia and OpenAI are abandoning their $100 billion investment deal in favor of a separate $30 billion arrangement, according to the Financial Times. As this roundup covered earlier this month, the original deal struck in September 2025 has been on thin ice as of late, as investor concern over the AI and software sectors has roiled markets. The original deal tied together two of the largest players in the AI space and sparked a wave of similar deals between hardware and software companies throughout the industry, knitting them together into a complex circular web of capital flows. However, such arrangements also sparked fears of a growing AI bubble, and since the beginning of 2026 tech stocks have been down nearly 17 percent. The new arrangement between Nvidia and OpenAI will involve a $30 billion investment from Nvidia in return for stock in OpenAI, much pared back from the previous deal, which included plans from OpenAI to purchase Nvidia hardware to power its AI models. OpenAI is currently in the midst of a massive fundraising round aimed at raising more than $100 billion, and Nvidia’s recent investment is considered part of this series, along with other investments expected from Japan’s SoftBank Group and Amazon.
• • The New York Times reports on a recent AI summit hosted by India as the nation seeks to boost its domestic AI industry. The India AI Impact Summit, held in recent days in New Delhi, is the fourth annual summit in a series that began in the UK in 2023. Previous summits had focused on AI safety measures and regulations, but this year’s event had an air of optimism about the technology, and the host nation sought to promote its own AI investments and entrepreneurs. Hundreds of exhibits, including those from many India-based startups, showcased AI-based solutions for everything from aircraft-bird collisions to fashion design. Several attendees noted that India had little domestic AI computing capacity, a resource that is dominated by large American companies like Microsoft, but several Indian conglomerates have pledged to support the construction of new data centers in the country. The summit also attracted foreign dignitaries, including French President Emmanuel Macron and the CEOs of Google and Microsoft, but it was Indian Prime Minister Narendra Modi who set the tone for the event with a call for regulated AI development: “[w]e must give AI an open sky, and keep the command in our hands.”
• • China’s ByteDance is alarming Hollywood with the release of its latest video-generation model, according to CNN. This past week, the company released Seedance 2.0, its latest video-generation model, which has been noted for its extremely realistic-looking outputs. However, it was particularly unrealistic videos that began appearing on social media that drew attention to the model, including those featuring major Hollywood celebrities such as Tom Cruise and Brad Pitt. Media giants Paramount and Disney both sent cease-and-desist letters to ByteDance, accusing them of copyright infringement, and ByteDance was also criticized by SAG-AFTRA, a major Hollywood labor union. ByteDance responded to these concerns by saying that they would implement measures to better protect intellectual property, but the model stoked fears in the entertainment industry of replacement by AI models, as well as broader concerns over privacy and the possibility of more realistic “deepfakes.” Further legal action from copyright holders may follow against ByteDance if the alleged infringement continues, but for now videos generated by Seedance 2.0 are likely to continue proliferating on the internet.
• • The Financial Times reports that an outage at Amazon Web Services (AWS) was caused by an AI tool acting autonomously. In December 2025, AWS, one of the world’s largest cloud providers, experienced a 13-hour outage to a system after Amazon engineers allowed the Kiro AI coding tool to make changes to the cloud system. According to people familiar with the matter speaking to the FT, the AI tool decided to delete and recreate a software environment, which caused the outage. Several Amazon employees told the FT that this was the second such occasion that an AI tool had caused a service disruption, and that the outages were “small but entirely foreseeable.” The company told the FT that “[i]n both instances, this was user error, not AI error,” and that AWS implemented further safeguards into its development process to prevent similar incidents from occurring in the future.
• • The U.S. Defense Department has threatened to sever its relationship with AI maker Anthropic, according to Axios. As this roundup covered earlier this month, the Pentagon and Anthropic have been involved in a spat over the proper amount of safeguards that should be included in AI technology used for military applications. In the Pentagon’s view, AI makers should allow their models to be used for “all lawful purposes,” including weapons development and battlefield operations. However, Anthropic has pushed back on these terms, saying that the use of AI for surveillance of Americans and autonomous weapons systems should be off-limits. Due to these disagreements, a source in the Pentagon told Axios that the department was considering limiting or terminating its partnership with Anthropic altogether. However, an Anthropic spokesman said that the company had “not discussed the use of Claude for specific operations with the Department,” and indicated that Anthropic was still committed to bringing its AI technology to national security applications.