Sony has announced it will discontinue the AirPeak S1 drone next year. The decision comes three years after the company started taking orders for the remotely controlled flying machine.

In a message on its Japanese website, the tech giant said it will stop selling the drone and most of its related products at the end of March 2025. However, battery packs for the AirPeak will continue to be available through March 2026, and repairs and software maintenance will be offered until the end of March 2030.

Sony put its decision down to “changes in the business environment,” which feels like company-speak for: “We weren’t selling enough of them.”

In what seemed like a brave move to take on drone giant DJI, Sony launched the AirPeak S1 in 2021, targeting professional filmmakers and photographers with a system that enabled the use of Sony’s high-end mirrorless cameras. In an even braver move, the company charged $9,000 for the aircraft, a hefty price that may well have been its undoing.

Sony unveiled the AirPeak S1 to great fanfare, trumpeting its top speed of 55 mph and its dual-operation mode that lets one person fly the drone while someone else focuses on controlling the camera. It also has the ability to stay steady in winds as strong as 44.7 mph, allowing filmmakers to capture smooth footage in challenging conditions.

But one early review of the machine was particularly damning. Photography specialist site PetaPixel criticized the AirPeak S1 for its complex set-up process, poor performance, camera feed issues, and app limitations. It titled its review: “A rare, utter flop for Sony.”

The company had been rolling out gradual improvements via software updates, but evidently they were not enough to save the machine. Sony has no other drones among its myriad of product offerings, and it looks likely to stay that way for the foreseeable future.

The Pixel 9 Pro arrived this year for users looking for a compact flagship. Previously, fans of the Pixel lineup had to turn to the less capable vanilla models if they wanted a phone with a smaller screen. Buyers of the device seem happy so far, with generally positive reviews. A new report claims that the Pixel 9 Pro’s cost of materials is just over $400.

After the release of new smartphones from big brands, the Bill of Materials (BOM) usually comes out, giving an idea of how much it costs to manufacture it. However, it is important to note that the bill of materials does not factor in several additional expenses associated with the commercialization of a device. This includes local/regional taxes, logistics, licensing, and more. You must also take into account the necessary margin for future discounts and offers.

It costs Google about $406 to make the Pixel 9 Pro, report claims

According to Nikkei, the cost of materials for the Pixel 9 Pro totals $406. For reference, the iPhone 16 Pro’s BOM is around $485, according to analysts. Additionally, the Pixel 9 Pro’s BOM is around 11% lower compared to the Pixel 8 Pro. However, the Pixel 8 generation did not feature a compact-sized Pro model. Components like the display and battery on the new-gen model are smaller, which can influence the cost. A more fair comparison for last year’s Pro model would be the BOM of the Pixel 9 Pro XL, which is not available yet.

According to the report, the new Tensor G4 chip costs Google around $80. This is significantly less than what other brands pay for the Snapdragon 8 Elite, which costs nearly $200. This is one of the advantages that brands that design their own chips have. The Samsung M14 display on the Pixel 9 Pro costs $75, the source claims. On the other hand, the camera hardware costs around $61.

Google optimizing production costs of Pixel devices

It would be nice to see the BOM of the Pixel 9 Pro XL emerge as well. This way, we could make a fairer comparison with the previous-generation Pixel 8 Pro. In any case, it appears that Google is also striving to optimize the production costs of its mobile devices. For instance, the camera hardware of the new model is reportedly cheaper than that of its predecessor.

Crypto fans are celebrating the results of the Ohio Senate race, where blockchain entrepreneur Bernie Moreno has defeated Senate Banking Chair Sherrod Brown, a three-term incumbent, in a contest that was key in the battle for control of the U.S. Senate.

Some $40 million of crypto money was directed at defeating Brown, with one PAC paying for five ads designed to boost awareness of Moreno, a businessman who worked as a luxury car dealer and had virtually no name recognition going into the contest.

The race was also a litmus test for whether the more than $245 million raised by the crypto industry this cycle would prove effective at the ballot box. The Ohio contest drew more ad spending than any Senate race in history, and was the biggest single target of crypto money this cycle.

Brown was unpopular with crypto fans, in part because he backed Sen. Elizabeth Warren, D-Mass., in holding hearings on whether digital tokens were tied to terrorism. He voted against pro-crypto legislation, called for more regulation of the sector, and regularly posted anti-crypto rhetoric on social media.

Ripple’s billionaire co-founder Chris Larsen told CNBC Tuesday night that Brown’s loss is “more fallout from the disastrous decision by President Joe Biden to outsource financial regulation to Sen. Warren.”

“Tonight I’m sad, but I’m never giving up,” Brown said in brief concession remarks Tuesday evening.

In December, Brown told journalists that he wasn’t concerned about the crypto industry’s rumblings against him.

“Bring ’em on,” Politico quoted Brown as saying to a crowd of reporters last year.

His antagonists are now taking a victory lap.

Tyler Winklevoss, one of the top individual crypto contributors this election cycle, called Brown a “crypto public enemy,” a “co-conspirator” to Sen. Warren, and a “Gary Gensler crony,” referring to the chair of the SEC. In a post on X, Winklevoss wrote, “The crypto army is striking!”

Coinbase CEO Brian Armstrong is equally excited.

“Tonight the crypto voter has spoken decisively — across party lines and in key races across the country,” Armstrong wrote in a post to X.

Armstrong called it the “most pro-crypto Congress ever” with more than 219 crypto-friendly candidates elected to the House and Senate.

The Stand With Crypto Alliance, launched by Coinbase last year, started a “Live election results” lander for crypto investors to keep track of the results. According to the tracker, 224 pro-crypto candidates have been elected to the House, against 106 anti-crypto House candidates that have won. In the Senate, 14 pro-crypto candidates have been elected, while nine anti-crypto candidates have been victorious.

NBC News hasn’t yet called all of these races.

Coinbase gave more than $75 million to Fairshake and its affiliated PACs, including a fresh pledge of $25 million to support the group in the 2026 midterms. Armstrong personally was among crypto’s top individual donors, giving over $1.3 million to a mix of candidates up and down the ballot.

Coinbase has been battling Gensler in court for more than a year over claims it sells unregistered securities.

The industry hopes that a pro-crypto Congress will pass rules that will apportion more of the regulatory responsibilities to the Commodities Future Trading Commission, which has traditionally been softer on its approach to policing the sector.

“Americans disproportionately care about crypto and want clear rules of the road for digital assets,” Armstrong wrote. “We look forward to working with the new Congress to deliver it. Thank you to everyone who stood with crypto today. We did it!”

The Fairshake affiliated PAC Defend American Jobs, which donated more than $40 million to Moreno, released a statement on Tuesday night, saying that Brown was a “top opponent of cryptocurrency.”

“Senator-Elect Moreno’s come-from-behind win shows that Ohio voters want a leader who prioritizes innovation, protects American economic interests, and will ensure our nation’s continued technological leadership,” the group said.

The political victories were reflected in the market.

Bitcoin surged to a new record above $75,000 earlier Tuesday night as the NBC News Decision Desk projected a good night for Republican nominee Donald Trump.

— CNBC’s Ece Yildirim contributed to this report.

You don’t have to take great pains to preserve your Switch to ensure that you can still play your favorite games on it years and years from now. Nintendo has revealed in its earnings report that the upcoming successor to the Switch will have backwards compatibility and will be able to run games made for the current console. In addition, Switch Online will also be available on the Switch 2, which means your saves stored on the cloud will be carried over and you’ll be able to play NES, SNES and Game Boy titles on the new console.

Nintendo explained that it’s making Switch Online available on the upcoming console, because it thinks it’s important for the company’s future to “carry over the good relationship” it has built with its more than 100 million annual playing users to the new device. The main way to do so is to make use of the Nintendo Account, which ties a user’s history to one account and enables the company to “maintain a continuous relationship” with them across console generations. Before the Nintendo Account was introduced, Nintendo had no easy way to carry a user’s history and purchases over to the next console. “As a result, our relationship with the consumers was interrupted when a new system was purchased,” it said.

The company promised to reveal more information about the Switch 2 “at a later date,” though it didn’t say when exactly. In a recent event where we thought the new Switch could be announced, Nintendo launched an alarm clock instead. Based on its earnings results, it looks like people could be choosing to wait for the new console instead of buying the current Switch: The company had to downgrade its sales forecast for the fiscal year due to a big decline in console sales compared to the same periods last year.

Enterprises looking to deploy multiple AI agents often need to implement a framework to manage them. 

To this end, Microsoft researchers recently unveiled a new multi-agent infrastructure called Magnetic-One that allows a single AI model to power various helper agents that work together to complete complex, multi-step tasks in different scenarios. Microsoft calls Magnetic-One a generalist agentic system that can “fully realize the long-held vision of agentic systems that can enhance our productivity and transform our lives.”

The framework is open-source and available to researchers and developers, including for commercial purposes, under a custom Microsoft License. In conjunction with the release of Magnetic-One, Microsoft also released an open-source agent evaluation tool called AutoGenBench to test agentic systems, built atop its previously released Autogen framework for multi-agent communication and cooperation.

The idea behind generalist agentic systems is to figure out how autonomous agents can solve tasks that require several steps to finish that are often found in the day to day running of an organization or even an individual’s daily life. 

From the examples Microsoft provided, it looks like the company hopes Magnetic-One fulfills almost mundane tasks. Researchers pointed Magnetic-One to tasks like describing trends in the S&P 500, finding and exporting missing citations, and even ordering a shawarma. 

How Magnetic-One works

Magnetic-One relies on an Orchestrator agent that directs four other agents. The Orchestrator not only manages the agents, directing them to do specific tasks, but also redirects them if there are errors.

The framework is composed of four types of agents other than the Orchestrator:

• Websurfer agents can command Chromium-based web browsers and navigate to websites or perform web searches. It can also click and type, similar to Anthropic’s recently released Computer Use, and summarize content. 
• FIleSurfer agents read local files list directories and go through folders.
• Coder agents write codes, analyze information from other agents and create new artifacts.
• ComputerTerminal provides a console where the Coder agent’s programs can be executed. 

The Orchestrator directs these agents and tracks their progress. It starts by planning how to tackle the task. It creates what Microsoft researchers call a task ledger that tracks the workflow. As the task continues, the Orchestrator builds a progress ledger “where it self-reflects on task progress and checks whether the task is completed.” The Orchestrator can assign an agent to complete each task or update the task ledger. The Orchestrator can create a new plan if the agents remain stuck. 

“Together, Magentic-One’s agents provide the Orchestrator with the tools and capabilities that it needs to solve a broad variety of open-ended problems, as well as the ability to autonomously adapt to, and act in, dynamic and ever-changing web and file-system environments,” the researchers wrote in the paper. 

While Microsoft developed Magnetic-One using OpenAI’s GPT-4o — OpenAI is after, all a Microsoft investment — it is LLM-agnostic, though the researchers “recommend a strong reasoning model for the Orchestrator agent such as GPT-4o.” 

Magnetic-One supports multiple models behind the agents, for example, developers can deploy a reasoning LLM for the Orchestrator agent and a mix of other LLMs or small language models to the different agents. Microsoft’s researchers experimented with a different Magnetic-One configuration “using OpenAI 01-preview for the outer loop of the Orchestrator and for the Coder, while other agents continue to use GPT-4o.”

The next step in agentic frameworks

Agentic systems are becoming more popular as more options to deploy agents, from off-the-shelf libraries of agents to customizable organization-specific agents, have arisen. Microsoft announced its own set of AI agents for the Dynamics 365 platform in October. 

Tech companies are now beginning to compete on AI orchestration frameworks, particularly systems that manage agentic workflows. OpenAI released its Swarm framework, which gives developers a simple yet flexible way to allow agents to guide agentic collaboration. CrewAI’s multi-agent builder also offers a way to manage agents. Meanwhile, most enterprises have relied on LangChain to help build agentic frameworks. 

However, AI agent deployment in the enterprise is still in its early stages, so figuring out the best multi-agent framework will continue to be an ongoing experiment. Most AI agents still play in their playground instead of talking to agents from other systems. As more enterprises begin using AI agents, managing that sprawl and ensuring AI agents seamlessly hand off work to each other to complete tasks is more crucial. 

Swiggy has raised around $606 million from a set of more than 75 anchor investors as part of its $1.35 billion initial public offering, as the Indian food delivery and quick commerce startup prepares for the country’s second-largest listing of the year next week.

The Bengaluru-based startup, which is seeking an $11.3 billion valuation in the IPO, received bids worth $15 billion for the $600 million portion. Indian institutional investors received about 56% of the overall anchor allocation, sources familiar with the matter said. Eight of the top 10 Indian mutual funds have invested in the anchor round.

The anchor investors include BlackRock, Fidelity, Norges Bank, BNP Paribas, HSBC, Nomura, Jane Street, Citadel, Motilal Oswal, Kotak, and 360 One, as well as mutual funds and insurance units operated by Indian lenders SBI, ICICI, Kotak, and HDFC, the sources said, requesting anonymity.

In an exchange filing, following the publication of the story, Swiggy confirmed the fundraise.

Swiggy, which counts Prosus, SoftBank, Accel, and Coatue among its backers, competes with firms including Zomato and Nexus-backed Zepto.