|
|
|
Jul 15, 2025
|
|
|
|
|
Supported by
|
|
|
|
|
|
|
|
|
|
Nvidia CEO Jensen Huang said Monday that the chip designer plans to resume selling an artificial intelligence chip to China that the U.S. government had previously restricted. After meetings in Washington, Nvidia said it had begun filing export license applications to sell the H20 chip, which is less powerful than chips it sells in the U.S. The chip had been restricted due to U.S. concerns it would aid a geopolitical rival. Nvidia said it expects the Commerce Department to grant the license. The company said it hopes to begin shipments to China soon. Nvidia earlier this year stopped issuing forecasts of China-related revenue, and its executives previously warned the firm could take an $8 billion hit this quarter as a result of recent restrictions. Before the restrictions, Chinese companies including ByteDance, Alibaba Group and Tencent Holdings placed at least $16 billion in orders for the H20s.
|
|
|
|
Coding assistant company Cognition has entered a definitive agreement to acquire the remaining staff and assets of Windsurf, according to a person with direct knowledge of the matter. All the remaining employees who did not join Google following a $2.4 billion licensing and recruiting deal with Windsurf last week will join Cognition, the person said. Cognition will have access to the more than $100 million left in Windsurf’s bank as a result of the deal, financial terms of which couldn’t be learned. The deal comes after a tumultuous 72-hours for Windsurf, a four-year-old coding startup based in Mountain View, California. On Friday, Google hired Windsurf CEO Varun Mohan and a group of researchers and engineers from the company, in exchange for paying a $2.4 billion nonexclusive licensing fee, the vast majority of
which went to Windsurf shareholders. The majority of Windsurf’s 250-person staff remained at the company after the Google deal. As a result of the deal with Cognition, the Windsurf employees will see their equity vest sooner than scheduled, the person said.
|
|
|
|
Meta Platforms in recent weeks has discussed developing closed artificial intelligence models which would represent a departure from its current focus on open-source, or free, models, according to four people familiar with the discussions. Some leaders, including Meta’s chief AI officer Alexandr Wang, have suggested that the company shouldn’t open-source its best models, two of those people said. Other executives have said open-sourcing remains advantageous as Meta tries to catch up to rivals, another person said. The discussions come amid a broader overhaul of Meta’s efforts in AI, after stumbles earlier this year. The social media giant last month finalized a deal to invest $14.3 billion in Scale AI and hired Wang, the data labeling company’s CEO. It has also hired former GitHub CEO Nat Friedman and former Safe Superintelligence CEO Daniel Gross and is expected
to partially buy out Friedman and Gross’ venture capital fund. “Our position on open source AI is unchanged,” a spokesperson for Meta said in a statement. “We plan to continue releasing leading open source models. We haven’t released everything we’ve developed historically, and we expect to continue training a mix of open and closed models going forward.” Meta in recent weeks has also paused some work on the largest version of its flagship large language model Llama 4, according to three people familiar with the work, marking the
latest setback for the model. Meta previously delayed two versions of Llama 4 over performance problems during its development. The company later also delayed a reasoning version and the largest version of the model. The New York Times earlier reported some details related to Meta’s discussions about open-source AI and work on Llama 4.
|
|
|
|
The U.S. Department of Defense has awarded contracts worth up to $200 million to artificial intelligence developers xAI, Google and Anthropic to “address critical national security needs,” the agency said Monday. The Pentagon said OpenAI won a similar $200 million contract last month. The government will use the firms’ AI tools for “essential tasks in our warfighting domain as well as intelligence, business, and enterprise information systems,” Chief Digital and AI Officer Doug Matty said in a statement. The contracts will allow other federal agencies such as the General Services Administration to use the AI applications emerge from the DoD work, according to the announcement. AI firms are increasingly jockeying for public sector contracts. OpenAI recently beefed up sales teams focused on winning such deals,
The Information previously reported. And Elon Musk’s xAI on Monday announced a new version of its Grok chatbot meant for government customers.
|
|
|
|
Elon Musk says Tesla shareholders will vote on whether the electric vehicle maker should invest in xAI, noting in a post on X that “if it was up to me, Tesla would have invested in xAI long ago.” Musk’s comments, made on Sunday night, came a day after the Wall Street Journal reported that Musk’s SpaceX would
invest $2 billion in xAI, part of a $5 billion fundraising by xAI already underway. Musk’s comment about the Tesla investment appeared to be off the cuff, however, coming as a response to a Tesla fan arguing that “Tesla needs to be able to invest in xAI.” Musk has also made clear his opposition to combining Tesla and SpaceX, a sign that despite the investments between his companies, he doesn’t want to merge them into one.
|
|
|
|
| |
