|
|
|
Mar 19, 2026
|
|
|
|
|
Supported by
|
|
|
|
|
|
|
Happy Thursday! Microsoft is concerned about OpenAI’s plan to release an AI agent product through AWS. Meta is shutting down its Horizon Worlds platform for virtual reality headset users. JPMorgan halts a $5.3 billion debt deal for Qualtrics.
|
|
|
|
OpenAI’s plan to release an AI agent product through Amazon Web Services last month raised concerns among senior Microsoft executives that the product would violate OpenAI’s contract to run models exclusively on Microsoft’s Azure cloud. It’s now publicly intimating that a rift is brewing. “We are confident that OpenAI understands and respects the importance of living up to [its] legal obligation,” a Microsoft spokesperson said. (The Financial Times earlier reported on Microsoft’s comments.) OpenAI and AWS, as part of a new investment and business partnership, said last month they will develop a new product that lets AWS customers build AI tools using OpenAI models. Microsoft holds the exclusive rights to sell OpenAI’s models directly to cloud customers, but Amazon said the new product—which it’s calling a stateful runtime environment—will rely entirely on versions of OpenAI models that run on AWS without needing the Microsoft-hosted versions. Microsoft executives have been concerned that the potential product, which AWS plans to release in the coming months, could violate Microsoft’s contractual rights, a person with knowledge of the situation previously said. Microsoft’s deal with OpenAI grants it exclusive rights to host OpenAI’s models through what’s known as an application programming interface, while the AWS product aims to sidestep that by instead offering tools for developing custom AI
applications instead of directly selling the model APIs themselves. (Read more about the stateful product here.)
|
|
|
|
Meta Platforms is shutting down Horizon Worlds for its virtual reality headset users, the company announced Tuesday. The VR social network will continue to operate via its mobile app, but users accessing it through the Meta Quest headset will no longer be able to create, publish or update virtual environments starting in June. Horizon Worlds was once a central part of CEO Mark Zuckerberg’s vision for the metaverse, but the platform has faced scrutiny since its launch over concerns about child safety and
moderation, as well as criticism over its cartoon style avatars. In January, the company laid off about 1,500 employees in Reality Labs, the division responsible for its virtual and mixed reality products, including Horizon Worlds and Ray-Ban Meta smart glasses.
|
|
|
|
JPMorgan and other banks have suspended a $5.3 billion debt deal for survey software company Qualtrics because debt investors were concerned over Qualtrics’ exposure to the decline in software valuations, according to Bloomberg. The troubled deal shows how investor fears about AI harming traditional software firms is having an impact beyond corporate share prices. Such news may not bode well for the likes of Salesforce, which last week sold $25 billion in bonds to complete half of its planned stock buybacks and could further
leverage debt for its remaining $25 billion purchases. Qualtrics, which previously traded publicly, was acquired several years ago by private equity firm Silver Lake and other investors. In October last year, Qualtrics said it would acquire a data analytics firm Press Ganey Forsta for nearly $7 billion, and the debt deal with JPMorgan was intended to support that acquisition, Bloomberg previously reported. Qualtrics and JPMorgan did not respond to requests for comment. A person familiar with the deal told The Information that JPMorgan is waiting to
complete a separate debt deal to support the buyout of Electronic Arts before turning focus to Qualtrics. JPMorgan is still in talks with investors on Qualtrics, the person said.
|
|
|
|
Tencent Holdings President Martin Lau said the company wants to create an AI agent within its popular WeChat messaging app in China that can handle a wide range of practical daily tasks for users, confirming an earlier story by The Information. “Our goal for the future is to create an AI agent within WeChat,” Lau said on Wednesday during Tencent’s annual earnings press conference. He said such an AI agent would utilize WeChat’s vast ecosystems of miniprograms, content, social interaction and payments. “This would allow the AI agent not only to chat
with users but also perform many practical tasks.” Lau said building a WeChat AI agent involves numerous challenges, including how to fully protect users’ privacy and security. The company also needs to make sure the operation of AI models powering the agent can support the app’s huge user base, he said. “We don’t have a specific timeline for this, but we are making progress step by step,” Lau added. Tencent also plans to at least double its investment in the development of new AI products this year, from the 18 billion yuan ($2.6 billion) it spent last year. In recent weeks, the company has unveiled multiple AI agent products based on or similar to OpenClaw, for example. Tencent’s solid revenue and profits from its core businesses, including
videogames and social networks, enable the company to invest more in new AI products, Lau said. The company reported fourth quarter revenue up 13% to 194.4 billion yuan, and net profit up 18% to 66.7 billion yuan from a year earlier.
|
|
|
|
|
|
| |
