| |  | Mar 23, 2026 | | | | | Supported by |  | | | | | | Welcome back! Elon Musk says Tesla and SpaceX will build a "Terafab" that will manufacture chips. OpenAI expects to increase its headcount to around 8,000 employees by year end, up from about 4,500. Chinese humanoid maker Unitree files for a $610 million IPO.
| | | | Elon Musk said Tesla and SpaceX will build a "Terafab" together in Austin that will design and manufacture semiconductors for both companies, though he did not provide a timeline for the project. The factory, located near Tesla's current Austin headquarters, will eventually produce two series of chips: one tailored to projects including Tesla's Cybercab and Optimus robot, and another for space, Musk said at an event in Austin Saturday, which was livestreamed. Tesla currently uses AI4 chips manufactured by Samsung in its vehicles, while xAI, now part of SpaceX, is a big Nvidia customer. Musk had previously said that Tesla was exploring building a chip fab because output from existing suppliers like TSMC isn't enough to power Tesla's future plans. While Tesla's electric vehicle business has suffered two years in a row of declining sales, Musk has repeatedly projected gigantic growth for new projects like the Cybercab and Optimus robots, as well as demand for data centers in space. Saturday's presentation included a slide showing that Tesla expects to manufacture 1 billion Optimus robots. Musk also showed a rendering of a satellite, which he called an "AI Sat Mini," that he presented as part of SpaceX's plans to eventually build a gigantic orbital data center. Musk said the AI Sat Mini could use solar power to deliver 100 kilowatts of power. That means SpaceX would have to launch thousands of such satellites to match the computing power Musk's AI company, xAI, already has at its data centers on Earth. | | | | OpenAI expects to increase its headcount to around 8,000 employees by year end, up from about 4,500 employees, according to a person familiar with the plans. The hiring plans come as it shifts some of its resources from new initiatives to focus on its flagship ChatGPT as well as its coding product, and accelerates its efforts to go after large corporate customers. It's planned to hire hundreds of consultants to expand a technical consulting team that helps large corporations develop custom AI applications and agents to automate employee tasks, The Information previously reported. The new hires will largely work across product development, engineering, research and sales, according to the Financial Times, which first reported on the plans. | | | | Chinese robotics startup Unitree, one of the world's largest humanoid makers, has filed for an initial public offering in Shanghai, looking to raise 4.2 billion yuan ($610 million). In the filing submitted on Friday, Hangzhou-based Unitree disclosed that its operating revenue last year more than quadrupled to 1.7 billion yuan, while its adjusted net profit grew more than seven-fold to 600 million yuan. Unitree said it sold 3,551 humanoids in the first nine months of 2025, up from 410 units in all of 2024. Unitree is the first among a new generation of Chinese humanoid startups to file for an IPO. Beijing views the development of AI-powered robots as a strategic effort to address labor shortages and move ahead in the country's technology race with the U.S. Last month, Unitree took center stage at state broadcaster CCTV's annual Spring Festival Gala, China's most-watched TV program, as its humanoids performed complex martial arts routines on stage. Unitree said it plans to use the proceeds from the IPO mainly for research and development spending on AI models as well as hardware. | | | | Anthropic filed a brief on Friday showing how close the Department of Defense was to reaching an agreement with it about using its artificial intelligence even after Defense Secretary Pete Hegseth Feb. 27 said he would direct the agency to declare the company a supply chain risk. Anthropic in the filing included a March 4 email from Emil Michael—the undersecretary of war who was leading negotiations with Anthropic—to Anthropic CEO Dario Amodei saying that Michael believed the two sides were "very close" to an agreement. The company argued that the Pentagon's willingness to negotiate after Hegseth's Feb. 27 announcement, showed that the Pentagon did not have "genuine concerns that Anthropic would sabotage military operations." The filings also showed that Anthropic had been working with the U.S. military for longer than previously known. In a supplementary declaration, Thiyagu Ramaswamy, Anthropic's head of public sector, said that Claude was available on a third-party cloud provider's top secret cloud and being used by the intelligence community and Department of Defense customers by May 2024. The cloud provider is Amazon Web Services, according to a person with knowledge of the situation. Anthropic announced a deal to serve the U.S. military through Palantir and AWS in November 2024. In the Friday filing, Anthropic reiterated that it should receive a preliminary injunction halting the Pentagon's designation. The company said it does not have the ability to cause its Claude AI to stop working or alter its operations after it is deployed by the Pentagon, pushing back on the Pentagon's argument. | | | | A jury found Elon Musk liable on Friday for some Twitter investor losses relating to his threat to back out of his acquisition of the social media site in 2022. Musk had struck a deal to buy Twitter for $54.20 per share in April of 2022. But in July of that year, he tried to back out of it, blaming what he said were a high number of bots and spam accounts on the site. The statements led to fluctuations in Twitter's share price that led to losses for some investors. Musk ultimately purchased Twitter for $44 billion, after Twitter sued to enforce the acquisition agreement. The decision in San Francisco federal court is likely to be appealed by Musk. Total damages could reach $2.6 billion, CNBC reported, citing attorneys for the plaintiffs. | | | | The Trump Administration on Friday released a framework detailing its policy approach to the AI industry, calling on Congress to pass federal laws for AI regulation that would override state legislation. A "patchwork of conflicting state laws would undermine American innovation and our ability to lead in the global AI race," the White House said. The administration's framework envisions giving parents tools like account controls to manage their children's online activity and usage of AI; streamlining permitting for data centers to generate power on-site; ensuring that American consumers aren't paying higher electricity prices due to data centers; and introducing AI-focused skills programs for workers. The Trump Administration also said that it doesn't believe AI model training on copyrighted material violates copyright laws, but added that the courts should be the ones to resolve this issue. It said that Congress should defend free speech in AI models and shouldn't overly regulate the AI industry, which could discourage innovation. Instead, Congress should make federal datasets available for model training and should encourage individual industries to regulate their applications of AI, rather than create a new federal body to regulate AI broadly. The framework is likely to set off intense political debate, as some Republican governors, like Florida's Ron DeSantis, have been pushing for state regulation of AI. | | | | Jeff Bezos' space firm Blue Origin asked the U.S. government for permission to launch up to 51,600 satellites that would one day handle artificial intelligence computing from orbit. The plan could put Blue Origin in competition with Elon Musk's SpaceX, which said in a January filing that it wants to put as many as one million satellites into space to form a gigantic data center. Blue Origin outlined its rival plan, called "Project Sunrise," in a Federal Communications Commission filing late Thursday. While both Musk and Bezos have talked up the idea of data centers in space, their companies' filings have been light on detail, such as the proposed design of the data center satellites or when each company would expect to launch them. Musk said this year that it will be cheaper to run AI compute in space than on Earth by 2028 or 2029, while Bezos said last October that it could take "a couple of decades" to build large data centers in space. | | | | Amazon is working on a smartphone that incorporates artificial intelligence capabilities, more than a decade after the e-commerce giant tried and failed to crack the market with its Fire phone, Reuters reported on Friday. The plans indicate Amazon is taking a different approach to new AI devices than OpenAI, which is focused on a smart speaker and other home and wearable devices. Amazon announced an initial $15 billion investment in OpenAI last month, and will receive customized versions of OpenAI models to power Amazon consumer products through that deal. Code-named Transformer internally, the effort comes out of a unit within Amazon's devices division called ZeroOne, which aims to create breakthrough gadgets and is led by J Allard, according to the report. One vision for a device includes a "dumbphone" with limited features, the report said, which could help Amazon promote the device as more of an accessory as opposed to a primary smartphone. A release timeline isn't clear, and it's possible Amazon could scrap the project due to strategy shifts, the report said. | | | | | Popular articles
By Sri Muppidi and Aaron Holmes
By Stephanie Palazzolo and Aaron Tilley
| | | | | Opportunities Empower your teams to stay ahead of market trends with the most trusted tech journalism.
Learn more
Reach The Information's influential audience with your message. Connect with our team | | | | | |
0 comentários:
Postar um comentário