Exclusive: Meta’s Muse Tops 3 Million Weekly Users -- Revenue Quintupled at AI Memory Maker Micron -- DeepSeek Open-Sources Software Tools for Huawei AI Chips -- Microsoft Science President Peter Lee to Step Down  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏  ͏ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ ­ 

Now streaming → →

Oct 01, 2026

The Information AM

Supported by Sponsor Logo

Save 25% on an annual subscription to read the most important news about technology and business first. For even more, Save $250 on The Information Pro for unlimited access to our proprietary org charts, databases and surveys.

Happy Thursday! Apple plans an aggressive new push into the smart-home market. Meta's Muse now has more than 3 million weekly users. Micron's revenue nearly quintuples on strong demand for memory hardware for AI chips.

Read more briefings
1.
Apple Plans Aggressive New Push Into Smart Home
By Aaron Tilley Source: Bloomberg 

Apple plans to launch a smart home hub device on October 13th, Bloomberg reported.

The hub has a roughly 6-inch square display that can sit on countertops or be mounted on walls, the news outlet reported. The device will showcase the capabilities of Apple’s overhauled Siri, which was recently released after a delay of two years, and be able to recognize people who interact with it. The screen will display the identified user’s personal information, such as messages and upcoming calendar appointments.

With the new product, Apple is moving into a crowded market. Google and Amazon launched their own smart home displays years ago. Apple has dipped its toe into the smart home over the years, with features like HomeKit to let devices hook into Apple’s software, but has limited success compared to rivals.

As part of this push, Apple is also releasing an updated HomePod mini and a new Apple TV.

2.
Exclusive: Meta’s Muse Tops 3 Million Weekly Users
By Jyoti Mann Source: The Information 

Meta Platforms’ Muse now has more than 3 million users who give the AI agent at least one prompt a week, according to internal data reviewed by The Information, giving the product a sizable active audience just weeks after its launch.

That figure is a subset of a broader group totaling more than 4 million who use the AI assistant in some way at least once a week, according to the figures. The gap is attributable to people who review something Muse did, approve an action or otherwise interact with it without giving the assistant a new task or query.

The number of daily active users who have sent at least one prompt to Muse surpassed 1 million, the internal data showed, with the majority of Muse’s prompts coming through the app rather than the web.

The figures provide the clearest picture yet of Muse’s nascent popularity, and mark a sharp increase from figures previously reported by The Information that showed Muse had about 500,000 weekly users and about 250,000 daily active users about a week after it launched.

The early numbers point to strong initial adoption, but they remain tiny compared with Meta’s broader reach. Meta’s family of apps reaches billions of people globally, meaning the millions of users currently active on Muse represent only a fraction of the company’s potential consumer base.

Muse is also entering an increasingly competitive market for AI agents that can take actions on behalf of users. SpaceXAI launched Grok Bot in beta in August and OpenAI introduced Dots this week, its own always-on agents. Instinct, another AI assistant app launched by a fledgling startup, has also gained popularity in tech circles.

3.
Revenue Quintupled at AI Memory Maker Micron
By Amir Efrati Source: The Information 

Memory maker Micron said revenue nearly quintupled to $54.2 billion in the quarter ending Sept.3, as demand for high-bandwidth memory hardware for AI chips showed no sign of slowing down.

Micron’s gross margin nearly doubled to $86.8%, thanks to its price hikes, allowing it to generate an astounding $33 billion in free cash flow during the quarter, up from about $17.5 billion in free cash flow in the prior fiscal quarter. That’s higher than Apple’s $32 billion in free cash flow in the July quarter, and also far higher than even Nvidia’s most recent cash result.

High-bandwidth memory makers like Micron are effectively sold out through next year, and the real test will come in a few years when long-term agreements it signed this year with large cloud providers and other customers lapse. Still, those agreements have turned such memory firms into guaranteed money makers for the foreseeable future. Micron and its ilk are also funding demand for their products, as each of the three major HBM providers invested in Anthropic’s latest funding round.

Micron’s gain is seemingly everyone else’s pain. Nvidia has tested reduced-memory versions of its next-generation Rubin Ultra chip, and Nvidia’s latest chip system prices are rising about 17%, driven partly by memory costs, The Information has reported. Meta and Apple, whose consumer hardware also uses products from memory firms like Micron, have also flagged price hikes due in part to memory prices.

4.
DeepSeek Open-Sources Software Tools for Huawei AI Chips
By Juro Osawa Source: The Information  

DeepSeek said Wednesday that it would open-source software tools that are customized for Huawei’s AI chips, as the Chinese AI firm steps up its efforts to use domestic semiconductors for training and running its models.

The software tools include a Huawei-compatible version of TileLang—a high-level programming language that DeepSeek has also used for Nvidia chips—as well as computing and communication libraries, the company said in a social media post on WeChat. DeepSeek, which previously open-sourced similar tools for Nvidia’s platform, is now offering all the equivalent tools for Huawei’s platform.

DeepSeek’s latest move shows how the company is accelerating its adoption of Huawei chips. CEO Liang Wenfeng told investors in a recent meeting that a major priority for his company is to use more domestic chips to train its models, adding that he expects Huawei to start delivering training chips to DeepSeek as early as the fourth quarter this year, The Information reported.

For DeepSeek and other Chinese AI firms, one of the biggest challenges ahead is how to secure enough compute to train and run their powerful new models in light of U.S. export restrictions on Nvidia’s advanced chips.

In the WeChat post, DeepSeek said it has been working closely with Huawei to build advanced “supernode” AI computing system using Huawei’s hardware.

During the research and development process, “Huawei’s team provided full and unwavering support,” DeepSeek said in the post.

DeepSeek first started experimenting with Huawei chips for AI model training, and Huawei only started to collaborate with DeepSeek after that, The Information previously reported.

5.
Microsoft Science President Peter Lee to Step Down
By Aaron Holmes Source: The Information 

Microsoft veteran Peter Lee, who for years led the company’s internal research unit, will leave the company, he announced Wednesday.

Lee is the latest in a string of senior executive departures from Microsoft announced this year. So far this year, Office and Windows boss Rajesh Jha, Xbox CEO Phil Spencer, chief communications officer Frank Shaw and security chief Charlie Bell have announced plans to leave the company.

Lee joined Microsoft in 2010 to help lead its research division of roughly 1,500 scientists, which has conducted research on fields relevant to Microsoft’s business such as AI, quantum computing, search, robotics and cryptography, among others. That research has resulted in new Microsoft products, including its launch of a quantum computing chip last year that Microsoft said could shape the future of computing.

Under Lee, Microsoft Research initially led Microsoft’s efforts to develop AI, but the company was ultimately outpaced by younger labs like OpenAI, a fact CEO Satya Nadella once bemoaned before making Microsoft’s $10 billion investment in the lab in exchange for the rights to reuse its technology in 2023. Lee later led research that aimed to evaluate OpenAI’s models, including how they can be distilled into smaller models and how they could be applied to healthcare.

Lee in February stepped away from leading Microsoft Research into a new role as President of Microsoft Science. Since then, executive vice president Igor Tsyganskiy has led the research unit.

6.
Google Unveils Gemini 4 Argon, Pricing It Well Below Rivals
By Jason Dean Source: The Information 

Google launched Gemini 4 Argon, its latest flagship AI model, touting its abilities in cybersecurity defense and coding, legal and finance work. The company hasn’t launched such a frontier model in nearly a year, prompting questions about whether it could stay in the race with Anthropic and OpenAI.

The search giant is pricing the new model well below leading offerings from those rivals. Argon will start at an introductory price of $2 per million input tokens and $10 per million output tokens, then double that rate after the teaser period expires. Even the higher rate would be significantly less than the price of GPT-6 Astra, OpenAI’s most powerful model, and Claude Fable 5.1, the top model from Anthropic.

Similar to phased rollouts that other leading AI labs have done for new frontier models, Google said it is releasing Argon first to cybersecurity partners that are part of its Fairwind Program, which enables participants to use the technology to discover and fix potential software vulnerabilities in their systems. The company said it is developing so-called safeguards to prevent misuse before releasing the model more broadly, though it didn’t specify a timeline.

Koray Kavukcuoglu, Google’s chief AI architect and leader of its Google DeepMind unit, told The Information last week at the AI Agenda Live conference that the company was nearing the release of Gemini 4 and that it would “continue the fast-paced iterations” after that.

Google’s last next-generation flagship model was Gemini 3, released in November of last year, though it has since updated that and released a series of smaller, Flash models.

7.
OpenAI Accuses Moonshot of Distillation Campaign
By Juro Osawa Source: The Information  

OpenAI said it recently identified and mitigated a “coordinated model-distillation campaign” involving individuals associated with Moonshot AI, the Chinese developer of Kimi models.

OpenAI said the operators of the campaign tried to extract protected reasoning—the model’s hidden thinking process before it generates an answer—by interacting with models in ways that could make their reasoning visible. This included “copying encrypted reasoning from one conversation and asking a model in another conversation to decrypt and transcribe the hidden reasoning content,” the company said on Thursday.

While it was unclear whether all operators originated from a single actor, OpenAI attributed a core cluster of the activity to individuals linked to Moonshot.

Distillation, a common technique used in AI training, has become a highly controversial topic due to the intensifying tech rivalry between the U.S. and China. OpenAI has previously accused DeepSeek of distilling its models, while Anthropic has accused many Chinese AI labs, including Moonshot and DeepSeek, of engaging in industrial-scale distillation attacks against its Claude models.

OpenAI said low-volume distillation activity began on July 1. The volume jumped on July 24 and July 25 to 16,000 requests from more than 4,000 users. The company eventually identified similar activity from more than 15,000 users, and fully disrupted the activity by July 28.

“Adversarial distillation poses safety and national security risks. Extracted reasoning could be used to train another model without preserving the safeguards applied to the original model’s user-facing outputs,” OpenAI said.

8.
Banks Including SocGen, MUFG Pull Back From Data Center Loan Deals
By Dakin Campbell Source: The Information 

Some of the largest lenders to the artificial intelligence build out are becoming more choosy about the deals they take on as market volatility and rising yields give lenders reasons to choose safer deals over more speculative ones.

Lenders including Société Générale, Sumitomo Mitsui Banking Corp. and Mitsubishi UFJ Financial Group have become more selective in lending to data center projects, people arranging the deals say. That narrows the pool of banks available to syndicate large multibillion-dollar loans needed for data center projects and other AI infrastructure.

Those banks had previously funded high-profile deals, with Société Générale leading a $7.1 billion debt deal for the first site in OpenAI and Oracle’s Project Stargate. MUFG helped lead a $38 billion financing for two more Oracle projects, while SMBC helped lead the $18 billion financing for Oracle’s New Mexico project.

9.
Sydney Sweeney-Backed Novig Raised at $2 Billion Valuation
By Yueqi Yang Source: Wall Street Journal 

Novig, a sports-only prediction market backed by Sydney Sweeney, has raised new funding at a $2 billion valuation, the Wall Street Journal reported, citing people familiar with the matter.

Founded in 2021 by Harvard graduates Jacob Fortinsky and Kelechi Ukah, Novig only offers trading on sports contracts, such as NFL game outcomes, under a license from the Commodity Futures Trading Commission that it received in June. Earlier this month, it launched a racy ads campaign featuring the actress Sweeney, an investor in Novig, that drew controversy.

In February, Novig raised a $75 million series B, led by Pantera Capital, at a $500 million valuation after the investment. The company declined to comment on the new funding round.

A message from Lambda

Stop overpaying for 90% of your AI.

Frontier APIs are worth it for the hardest 10% of work. Run the other 90% on open-weight models with reserved compute and fixed costs. Buy the frontier. Build the rest.

See the math.

This week on AI Deep Dive

Check out the most recent episode of AI Deep Dive. An upclose look into how frontier AI actually gets built — the models, the money, and the people behind them.

Can AI Make AI Smarter? | TITV’s AI Deep Dive

Popular articles





Opportunities

Group subscriptions

Empower your teams to stay ahead of market trends with the most trusted tech journalism.

Learn more


Brand partnerships

Reach The Information's influential audience with your message.

Connect with our team

Follow us
X
LinkedIn
Facebook
Threads
Instagram
Sent to you@example.com | Manage your preferences or unsubscribe | Help
The Information · 251 Rhode Island Street, Suite 107, San Francisco, CA 94103