Microsoft is accelerating its push to compete with OpenAI, its longtime collaborator, by developing its own powerful AI models and exploring alternatives to power products like Microsoft’s Copilot bot.
Microsoft has developed its own AI “reasoning” models comparable to models like OpenAI’s o1 and o3-mini, the The Information reports. OpenAI is said to have refused Microsoft’s requests for technical details about how o1 works — stoking tensions between the firms.
Microsoft has also developed a family of models called MAI that are competitive with OpenAI’s own, Bloomberg reports, and is reportedly considering offering them through an API later this year. Parallel to those efforts, Microsoft is said to be testing alternative AI models from xAI, Meta, Anthropic, and DeepSeek as possible replacements for OpenAI technology in Copilot.
Microsoft, which has invested around $14 billion in OpenAI to date, has looked to hedge its bets in a number of ways, including hiring DeepMind and Inflection co-founder Mustafa Suleyman to lead the tech giant’s AI efforts.
Four days before he leaves office, US president Joe Biden has issued a sweeping cybersecurity directive ordering improvements to the way the government monitors its networks, buys software, uses artificial intelligence, and punishes foreign hackers.
The 40-page executive order unveiled on Thursday is the Biden White House’s final attempt to kickstart efforts to harness the security benefits of AI, roll out digital identities for US citizens, and close gaps that have helped China, Russia, and other adversaries repeatedlypenetrate US government systems.
The order “is designed to strengthen America’s digital foundations and also put the new administration and the country on a path to continued success,” Anne Neuberger, Biden’s deputy national security adviser for cyber and emerging technology, told reporters on Wednesday.
Looming over Biden’s directive is the question of whether president-elect Donald Trump will continue any of these initiatives after he takes the oath of office on Monday. None of the highly technical projects decreed in the order are partisan, but Trump’s advisers may prefer different approaches (or timetables) to solving the problems that the order identifies.
Trump hasn’t named any of his top cyber officials, and Neuberger said the White House didn’t discuss the order with his transition staff, “but we are very happy to, as soon as the incoming cyber team is named, have any discussions during this final transition period.”
The core of the executive order is an array of mandates for protecting government networks based on lessons learned from recent major incidents—namely, the security failures of federal contractors.
The order requires software vendors to submit proof that they follow secure development practices, building on a mandate that debuted in 2022 in response to Biden’s first cyber executive order. The Cybersecurity and Infrastructure Security Agency would be tasked with double-checking these security attestations and working with vendors to fix any problems. To put some teeth behind the requirement, the White House’s Office of the National Cyber Director is “encouraged to refer attestations that fail validation to the Attorney General” for potential investigation and prosecution.
The order gives the Department of Commerce eight months to assess the most commonly used cyber practices in the business community and issue guidance based on them. Shortly thereafter, those practices would become mandatory for companies seeking to do business with the government. The directive also kicks off updates to the National Institute of Standards and Technology’s secure software development guidance.
Another part of the directive focuses on the protection of cloud platforms’ authentication keys, the compromise of which opened the door for China’s theft of government emails from Microsoft’s servers and its recent supply-chain hack of the Treasury Department. Commerce and the General Services Administration have 270 days to develop guidelines for key protection, which would then have to become requirements for cloud vendors within 60 days.
To protect federal agencies from attacks that rely on flaws in internet-of-things gadgets, the order sets a January 4, 2027, deadline for agencies to purchase only consumer IoT devices that carry the newly launched US Cyber Trust Mark label.
Xbox has announced a few more sustainability efforts, including an expansion of its repairability program. You’ll be able to buy official replacement parts for Xbox Series X/S systems from the and so you can repair your console yourself, even if it’s out of warranty. Until now, Microsoft had only through its own store, but now you can buy components for the Series S and both the and disc drive editions of the Series X. Along with various parts and step-by-step repair guides, iFixit also has an Xbox toolkit for sale.
Meanwhile, as of January 20, uBreakiFix by Asurion will be the first Xbox Authorized Service Provider. It will repair the consoles at nearly 700 participating locations in the US. Previously, authorized in-person Xbox repairs were only available at the Microsoft Store, so gamers will have many more places where they can go to get their console fixed without breaking the warranty.
These are positive steps forward for the right-to-repair movement. They should help reduce the number of games consoles that are consigned to the scrapheap.
“By expanding the number of ways players can get support and repair for their Xbox consoles, we also help extend the longevity of Xbox devices, reduce waste, promote re-use and ultimately reduce our environmental impact,” Trista Patterson, director of gaming sustainability at Xbox, wrote in an . In addition, there are no longer any single-use plastics in Xbox Series X/S packaging following a switch to fully paper- and fiber-based materials.
Elsewhere on the sustainability front, Microsoft says that, through recent efficiencies in how the Xbox Series S handles video content, there’s been an average power reduction of 10 percent while media apps are in use. As for the company’s game studios, they’ve been using the Xbox Sustainability Toolkit to make more energy-efficient games, Microsoft says.
If you buy something through a link in this article, we may earn commission.
Microsoft has a beta test that will finally bring cloud streaming to Xbox consoles. Participants in the Alpha Skip-Ahead and Alpha tiers of the Xbox Insiders program can start using this feature now on their Xbox Series X|S and Xbox One consoles.
This news is an extension of the “stream your own game” feature that Microsoft in November. That initial launch allowed Game Pass Ultimate subscribers to stream select games they’ve digitally purchased to their televisions, Meta Quest VR headsets and to some supported browser setups. The company said at the time that it planned to also bring streaming to Xbox consoles and to the Windows Xbox app in 2025.
While this update is a welcome addition to the “stream your own game” hardware, there are still some caveats on the feature. First, it’s limited to Game Pass Ultimate members. Second, the game needs to support cloud streaming. There’s a short of titles included in the program for now, but several of them are excellent ones that are well worth a look: Baldur’s Gate 3, Balatro, Cyberpunk 2077, Animal Well, Stray and the first six Final Fantasy games, to name a few highlights. Once this goes live to the whole Xbox audience, it should be a useful way to streamline game downloads and to access your whole library without needing to shell out for external storage.
In related Microsoft news, the Windows Xbox app is getting . The new Home screen for the app will highlight curated game collections and suggested titles, as well as recent game news, releases and sales.
Elon Musk’s lawsuit against OpenAI accusing the company of abandoning its non-profit mission was withdrawn in July, only to be revived in August. Now, in an amended complaint, the suit names new defendants including Microsoft, LinkedIn co-founder Reid Hoffman, and former OpenAI board member and Microsoft VP Dee Templeton.
The amended filing also adds new plaintiffs: Neuralink exec and ex-OpenAI board member Shivon Zilis and Musk’s AI company, xAI.
Musk was one of the original founders of OpenAI, which was meant to research and develop AI for the benefit of humanity, and was established as a non-profit originally. He left the company in 2018 after disagreements about its direction.
In the complaint, lawyers for Musk argue that OpenAI is now “actively trying to eliminate competitors” such as xAI by “extracting promises from investors not to fund them.” It’s also allegedly unfairly benefitting from Microsoft’s infrastructure and expertise in what Musk’s counsel describes in the filing as a “de facto merger.”
“xAI has been harmed by, without limitation … an inability to obtain compute from Microsoft on terms anywhere near as favorable as OpenAI receives … and the exclusive exchange between OpenAI and Microsoft of competitively sensitive information,” reads the complaint, filed late Thursday in federal court in Oakland, California.
Hoffman’s position on the boards of both Microsoft and OpenAI while also a partner at Greylock, the investment firm, gave Hoffman a privileged — and illicit — view into the companies’ dealings, the complaint alleges. (Hoffman stepped down from OpenAI’s board in 2023.) Greylock invested in Inflection, Musk’s counsel notes, the AI startup that Microsoft acqui-hired earlier this year — and which could reasonably be considered an OpenAI competitor, according to the complaint.
As for Templeton, whom Microsoft briefly appointed as a non-voting board observer at OpenAI, the amended filing alleges that she was in a position to facilitate agreements between Microsoft and OpenAI that would violate antitrust rules.
“The purpose of the prohibition on interlocking directorates is to prevent sharing of competitively sensitive information in violation of antitrust laws and/or providing a forum for the coordination of other anticompetitive activity,” the complaint reads. “Allowing Templeton and Hoffman to serve as members of OpenAI’s …. board undermined this purpose. “
Alongside Microsoft, Hoffman, and Templeton, California attorney general Rob Bonta is named as a defendant in Musk’s complaint. Bloomberg reported this month that OpenAI is in talks with Bonta’s office over the process to change its corporate structure.
Per the amended complaint, Zilis, who stepped down from OpenAI’s board in 2023 after serving as a member for roughly four years, has standing as an “injured employee” under California Corporations Code. Zilis repeatedly raised concerns over OpenAI’s dealmaking internally that fell on deaf ears — concerns substantially similar to Musk’s, according to the complaint.
Zilis has close ties to Musk, having worked as a project director at Tesla from 2017 to 2019 in addition to directing Neuralink research. (Neuralink is Musk’s brain-computer interface venture.) She’s also the mother of three of Musk’s children, Techno Mechanicus and twins Strider and Azure.
The 107-page amended complaint includes the unusual detail that OpenAI CEO Sam Altman proposed that OpenAI sell its own cryptocurrency in January 2018, before it ultimately decided to transition to a capped-profit structure.
“Heads up, spoke to some of the safety team and there were a lot of concerns about the ICO and possible unintended effects in the future,” Altman wrote in an email to Musk dated January 21, 2018, an exhibit filed with the amended complaint shows. An ICO, or initial coin offering, is an unregulated means by which funds are raised for cryptocurrency businesses. “Going to emphasize the need to keep this confidential, but I think it’s really important we get buy-in and give people the chance to weigh in early.”
Image Credits:Toberoff & Associates
Musk supposedly shot down the crypto sale idea. “I have considered the ICO approach and will not support it,” he wrote in an email reply to Altman and OpenAI co-founders Greg Brockman (now OpenAI’s president) and Ilya Sutskever (OpenAI’s ex-chief scientist), shows an exhibit. “In my opinion, that would simply result in a massive loss of credibility for OpenAI and everyone associated with the ICO.”
The thrust of the lawsuit remains the same on the plaintiffs’ side: that OpenAI profited from Musk’s early involvement in the company yet reneged on its nonprofit pledge to make the fruits of its AI research available to all. “No amount of clever drafting nor surfeit of creative dealmaking can obscure what is happening here,” reads the complaint. “OpenAI, Inc., co-founded by Musk as an independent charity committed to safety and transparency … [is] fast becoming a full for-profit subsidiary of Microsoft.”
OpenAI has sought to dismiss Musk’s lawsuit, calling it “blusterous” and baseless.
Elon Musk’s lawsuit against OpenAI accusing the company of abandoning its nonprofit mission was withdrawn in July, only to be revived in August. Now, in an amended complaint, the suit names new defendants including Microsoft, LinkedIn co-founder Reid Hoffman, and former OpenAI board member and Microsoft VP Dee Templeton.
The amended filing also adds new plaintiffs: Neuralink exec and ex-OpenAI board member Shivon Zilis and Musk’s AI company, xAI.
Musk was one of the original founders of OpenAI, which was meant to research and develop AI for the benefit of humanity, and was established as a non-profit originally. He left the company in 2018 after disagreements about its direction.
In the complaint, lawyers for Musk argue that OpenAI is now “actively trying to eliminate competitors” such as xAI by “extracting promises from investors not to fund them.” It’s also allegedly unfairly benefitting from Microsoft’s infrastructure and expertise in what Musk’s counsel describes in the filing as a “de facto” merger.
“xAI has been harmed by, without limitation … an inability to license OpenAI technology given Microsoft’s exclusive license … an inability to obtain compute from Microsoft on terms anywhere near as favorable as OpenAI receives … and the exclusive exchange between OpenAI and Microsoft of competitively sensitive information.”
Hoffman’s position on the boards of both Microsoft and OpenAI while also a partner at Greylock, the investment firm, gave Hoffman a privileged — and illicit — view into the companies’ dealings, the complaint alleges. (Hoffman stepped down from OpenAI’s board in 2023.) Greylock invested in Inflection, Musk’s counsel notes, the AI startup that Microsoft acqui-hired earlier this year — and which could reasonably be considered an OpenAI competitor, according to the complaint.
As for Templeton, whom Microsoft briefly appointed as a non-voting board observer at OpenAI, the amended filing alleges that she was in a position to facilitate agreements between Microsoft and OpenAI that would violate antitrust rules.
“The purpose of the prohibition on interlocking directorates is to prevent sharing of competitively sensitive information in violation of antitrust laws and/or providing a forum for the coordination of other anticompetitive activity,” the complaint reads. “Allowing Templeton and Hoffman to serve as members of OpenAI’s …. board undermined this purpose. “
Per the amended complaint, Zilis, who stepped down from OpenAI’s board in 2023 after serving as a member for roughly four years, has standing as an “injured employee” under California Corporations Code. Zilis repeatedly raised concerns over OpenAI’s dealmaking internally that fell on deaf ears — concerns substantially similar to Musk’s, according to the complaint.
Zilis has close ties to Musk, having worked as a project director at Tesla from 2017 to 2019 in addition to directing Neuralink research. (Neuralink is Musk’s brain-computer interface venture.) She’s also the mother of three of Musk’s children, Techno Mechanicus and twins Strider and Azure.
The 107-page amended complaint includes the unusual detail that OpenAI CEO Sam Altman proposed that OpenAI sell its own cryptocurrency in September 2017, before it ultimately decided to transition to a capped-profit structure. Musk supposedly shot down the crypto sale idea.
The thrust of the lawsuit remains the same on the plaintiffs’ side: that OpenAI profited from Musk’s early involvement in the company yet reneged on its nonprofit pledge to make the fruits of its AI research available to all. “No amount of clever drafting nor surfeit of creative dealmaking can obscure what is happening here,” reads the complaint. “OpenAI, Inc., co-founded by Musk as an independent charity committed to safety and transparency … [is] fast becoming a full for-profit subsidiary of Microsoft.”
Three weeks later, on Friday, Apple released the first update to macOS 15, and it claims to have fixed those issues. In the macOS 15.0.1 release notes, Apple says that the update “improves compatibility with third-party security software.”
Apple flagged the update in an email to TechCrunch on Thursday, and a spokesperson did not respond to a follow-up asking for more information.
Patrick Wardle, the founder of macOS and iOS security startup DoubleYou, and a longtime expert on Apple security and the developer of several free security tools for macOS, wrote on X that the macOS update includes “a fix for the networking issues that plagued the initial macOS 15 release.”
“And to any Apple apologist who blamed 3rd-party vendors, you deserve to be slapped with a large trout as this was an Apple bug reported before [golden master],” Wardle wrote, referring to the first public release of the macOS 15 software.
When Apple first released macOS 15, several cybersecurity professionals said they were unable to use some security tools, such as CrowdStrike’s Falcon and Microsoft Defender, because of an apparent bug in the new macOS operating system.
At the time, CrowdStrike spokesperson Kevin Benacci said that the company was “waiting for a macOS Sequoia update” to provide official support for its cybersecurity products on Apple’s operating system.
Ugur Koc, a developer who works as a cloud engineer for for cloud managed service provider Glueckkanja, said on X that the new macOS update “resolves the issue with [Microsoft] Defender for Endpoint and other antivirus software, where the network filter was causing issues with the internet connectivity.”
Neither CrowdStrike nor Microsoft responded to a request for comment.
Microsoft has joined an exclusive club of tech giants that are subject to a special abuse control regime in Germany. The country’s Federal Cartel Office (FCO) confirmed on Monday that the software giant could face restrictions if the competition authority deems an intervention is necessary.
The designation, which lasts for five years, is important, as it lets the German authority take a close interest in how Microsoft wields its influence through its activities around generative AI.
However, the regulator said it has yet to take any decisions on “possible proceedings.”
In recent years, Microsoft’s influence on OpenAI has landed the pair on antitrust regulators’ radars. The cozy relationship even saw Microsoft briefly hire OpenAI front-man Sam Altman and other key staffers last fall during a board dispute.
Although Altman ended up staying at OpenAI, the episode underscored the closeness between the two companies, and Microsoft even got a board observer seat at OpenAI (it gave it up this summer). However, careful structuring of their arrangement appears to have kept it flying for now.
The FCO has already looked at the two companies’ partnership, and it found last November that their relationship did not meet the threshold for a traditional merger review. However, now that the regulator is armed with more proactive and wide-ranging powers to regulate Big Tech, Microsoft’s dealings with OpenAI could face closer scrutiny in Germany going forward.
The FCO’s press release highlights how Microsoft’s Copilot AI assistant is used “in many parts” of its ecosystem. It also links the company’s strength in cloud computing to helping it enter partnerships with “highly innovative suppliers,” as it can “offer their AI models as services on Azure and integrate them into its own products.”
Commenting in a statement, Andreas Mundt, president of the FCO, also highlighted Microsoft’s long history of software dominance, adding: “Today, Microsoft’s ecosystem is stronger and more closely interconnected than ever before, because overarching all of its activities is the increasing use of the cloud and AI, key technologies in which Microsoft has consolidated its strong position by developing its own products and entering into cooperations.“
The FCO began investigating whether the tech giant’s market power met the bar for the special abuse controls regime back in March 2023. And this confirmation that the company has “paramount significance for competition across markets” unlocks a range of powers contained in the 2021 update to Germany’s antitrust rulebook. The reform aims to counteract concerns that Big Tech’s market power is hampering rivals’ ability to innovate and compete.
The German law already applies to Amazon, Apple, Google and Meta, and predates the European Union’s Digital Markets Act (DMA), a similar ex-ante competition reform that is also being used to clip Big Tech’s wings.
However, the DMA applies operational controls only to named platforms, while the FCO has designated Microsoft as a whole. This means the German authority has greater freedom to impose controls on Microsoft’s activities, including around AI, if it judges the company’s actions are crimping competition.
The EU’s DMA was drafted before the boom in generative AI tools made ChatGPT a household name. Microsoft is designated as a gatekeeper, but only two of its platforms are directly regulated: the Windows operating system and its social network LinkedIn. That limits the European Commission’s ability to intervene in Microsoft’s activities in AI unless they specifically fall within these two “core platform services.”
“Our decision applies to Microsoft as a whole, not only to individual services or products,” Mundt emphasized. “Based on our decision, we can stop anti-competitive practices which are not covered by the DMA.”
Microsoft spokeswoman Sophie Thomas said in an emailed statement: “We recognize our responsibility to support a healthy competitive environment and we will strive to be proactive, collaborative and responsible in working with the Bundeskartellamt [FCO]. Microsoft is partnering with Germany’s most innovative companies, and we’re committed to investing in the growth of its digital economy.”
Microsoft’s Windows Recall feature, which stores a timeline of activity snapshots on your PC, has a new release date for Windows Insiders. Microsoft unveiled the feature to much fanfare in May, only to delay it indefinitely (after blowback from security researchers) a few weeks later. After taking time to recalibrate, the company said on Wednesday it will roll out Recall to beta testers using Copilot+ PCs in October.
Windows Recall stores snapshots of everything you do on your PC. Designed as a “photographic memory” for your PC activity, it lets you revisit things like products, emails, documents or chats shown on your screen. The feature’s perks are easy to see, especially for those who spend long hours on their PC (or those with foggy memories).
But if that also sounds like a privacy nightmare, security researchers thought so, too. Despite safety assurances from Microsoft during its announcement at Build 2024, cybersecurity and privacy experts sounded the alarm. The fundamental problem was that intruders wouldn’t only get goodies from your traditional file system if they accessed your PC. In addition, they could see anything you’ve done on your computer from the moment you activated Recall to the present. That’s because Microsoft — for reasons we can’t quite comprehend (other than put AI in all the things as quickly as possible) — left Recall’s data unencrypted.
As security expert Kevin Beaumont detailed, Recall didn’t hide sensitive information like passwords or banking details. Sure, your timeline was theoretically safe as long as nobody could access your PC. But if you accidentally installed malware or let an intruder in through other means, they would find a motherlode of sensitive — unencrypted — data.
Microsoft
In response to the blowback, Microsoft added some common-sense security features that left us wondering why they weren’t there in the first place. Again, it’s hard to decipher the company’s motives for that omission when the feature was announced — other than speculating that it wanted to prioritize a seamless user experience over tight security.
These security changes included making the feature opt-in instead of enabled by default when setting up a Copilot+ PC. In addition, Microsoft said the feature would require Windows Hello — a face or fingerprint scan — and deploy “just in time” decryption (only unlocked through Hello). That means if a hacker gains access to your computer, your screenshot timeline should remain encrypted unless you lend your face or finger to unlock it (or they somehow find a way around Hello’s encryption).
The global outage caused by a faulty update from cybersecurity provider CrowdStrike on Friday affected some 8.5 million Windows devices, Microsoft said in a blog post. The update triggered a blue screen of death, bringing systems used by hospitals, airlines, banks and other major services temporarily to a standstill. Only machines running Windows were affected.
While the issue was mostly resolved by Friday afternoon, Microsoft and CrowdStrike are still dealing with the fallout. In the blog post on Saturday, Microsoft’s VP of Enterprise and OS Security, David Weston, wrote that the company is working with CrowdStrike to “develop a scalable solution that will help Microsoft’s Azure infrastructure accelerate a fix for CrowdStrike’s faulty update.” Microsoft has also called in help from Amazon Web Services (AWS) and Google Cloud Platform (GCP).
CrowdStrike said in its own blog post on Saturday that the update — a sensor configuration update — “was designed to target newly observed, malicious named pipes being used by common C2 frameworks in cyberattacks.” Unfortunately, for devices running Windows 7.11 and above that use CrowdStrike’s Falcon sensor, it instead “triggered a logic error that resulted in an operating system crash.” The total number of devices affected worked out to be “less than one percent of all Windows machines,” according to Weston.
