Broadcom ships latest networking chip to speed AI
By Max A. Cherney
SAN FRANCISCO (Reuters) -Broadcom has begun to ship its latest networking chip that aims to speed AI, the company said on Tuesday.
The chip, called the Tomahawk 6, boasts double the performance compared with the prior version and other traffic control features that make the networking chip significantly more efficient, Ram Velaga, a Broadcom senior vice president, told Reuters in a Monday interview.
The speed boost means that fewer networking switches are needed to perform the same task, Velaga said.
Broadcom's networking chips have gained increased importance because of AI. When constructing the necessary data centers for AI applications, infrastructure builders must string together hundreds or thousands of chips.
Building large-scale clusters of networked chips requires specialized networking gear and chips, of which the Tomahawk series of processors is one such component.
With the Tomahawk 6, Broadcom's engineers have boosted its speed and capabilities to the point where it can be used to construct the larger data centers that are necessary for AI, which can be over 100,000 graphics processors (GPUs) strung together, Velaga said.
"In a couple of years, you will start to see a million GPUs housed inside a physical building," he said.
Broadcom's networking chips use the Ethernet networking protocol, which has been a networking standard for decades. Nvidia produces hardware that uses a rival tech called InfiniBand and several products based on Ethernet.
"All of these networks can be very simply done on Ethernet, you don't need esoteric technologies," Velaga said.
The Tomahawk 6 is the first product in that line that will use several chips combined into a single package, a tech known as chiplets that is widely adopted by other chip designers such as Advanced Micro Devices. Adding chiplets roughly doubled the amount of silicon area used in the design, Velaga said.
Broadcom is producing the Tomahawk switch on Taiwan Semiconductor Manufacturing Co's three nanometer process.
Hashtags
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
Yahoo
16 minutes ago
- Yahoo
WWDC to focus on redesigns as Apple remains sidelined on AI, Bloomberg says
Apple's (AAPL) upcoming Worldwide Developers Conference will do little to assuage fears that the iPhone maker is a laggard in AI, Blomberg's Mark Gurman reports. Instead, the event will focus on design and productivity enhancements for its long-established operating system franchises. The company's keynote address will introduce redesigned software interfaces for the iPhone, iPad, Mac, Apple TV and Apple Watch, in addition to more minor tweaks to the Vision Pro headset. As part of the end-to-end overhaul, the company is also making a sweeping change to its software branding, which will shift from version numbers to a year-based system. That means Apple will introduce iOS 26, iPadOS 26, tvOS 26, visionOS 26, macOS 26 and watchOS 26 – named for 2026. Internally, the operating systems are known as Luck, Charisma, Discovery, Cheer and Nepali, respectively, the author notes. The AI changes will be surprisingly minor are unlikely to impress industry watchers, especially considering the rapid pace of innovation by Alphabet's (GOOG) (GOOGL) Google, Meta Platforms (META), Microsoft (MSFT) and OpenAI, the publication adds. Easily unpack a company's performance with TipRanks' new KPI Data for smart investment decisions Receive undervalued, market resilient stocks right to your inbox with TipRanks' Smart Value Newsletter Published first on TheFly – the ultimate source for real-time, market-moving breaking financial news. Try Now>> See the top stocks recommended by analysts >> Read More on AAPL: Disclaimer & DisclosureReport an Issue Apple's growing list of issues hinders AI reboot, WSJ says Apple expands partnership in India with Tata, Reuters reports Morning News Wrap-Up: Thursday's Biggest Stock Market Stories Apple says App Store ecosystem facilitated $1.3T in developer sales in 2024 This Is How Much Analysts Expect Apple's (AAPL) EPS to Fall after Court Ruling
Yahoo
16 minutes ago
- Yahoo
Will $50,000 Invested in Nvidia Stock Be Worth $1 Million in 10 Years?
Nvidia shares are up 850% since ChatGPT sparked the artificial intelligence (AI) boom, but most Wall Street analysts still recommend buying the stock. The company is the market leader in AI accelerator chips, but its true strength lies in vertical integration that spans hardware and software products. Seven stocks in the S&P 500 generated such colossal returns in the last decade that they would have turned $50,000 into $1 million. 10 stocks we like better than Nvidia › Nvidia (NASDAQ: NVDA) has been a cornerstone of the artificial intelligence (AI) trade for several years. Its share price has increased 850% since January 2023, a period that roughly coincides with the launch of ChatGPT. But Wall Street is still overwhelmingly bullish on the semiconductor company. Angelo Zino at CFRA Research thinks Nvidia "will be the most important company to our civilization over the next decade." More broadly, among 73 analysts following Nvidia, the median 12-month target price is $175 per share. That implies 25% upside from its current share price of $140. Could Nvidia stock turn $50,000 into $1 million over the next decade? Here are my thoughts. What sets Nvidia apart is vertical integration. The company has over 90% market share in data center graphics processing units (GPUs), chips that accelerate complex workloads such as artificial intelligence (AI). But the company supplements its GPUs with adjacent hardware like CPUs, interconnects, and networking equipment. Nvidia also develops software products. AI Enterprise is a suite of tools, code libraries, and pretrained models that streamline the development of AI applications for use cases like autonomous robots, conversational agents, and optimization systems. CrowdStrike uses those tools to power threat detection capabilities on its cybersecurity platform. Similarly, Omniverse is a software platform that supports 3D application development. It also serves as a simulation engine that lets engineers generate synthetic data for developing machine learning models. Amazon uses the Omniverse platform to optimize warehouse design and train fulfillment center robots. Nvidia frequently sets performance records at the MLPerf benchmarks, objective tests that evaluate AI systems on training and inference workloads. That is an important competitive advantage: Nvidia builds the best AI accelerators on the market. But vertical integration reinforces that advantage by letting the company design entire data center systems with the "lowest total cost of ownership," according to CEO Jensen Huang. Grand View Research says spending on AI hardware, software, and services will increase at 35.9% annually through 2030. Nvidia has a good shot at matching that growth rate. Indeed, Wall Street expects earnings to grow at 40% annually through the fiscal year ending January 2027. That makes the current valuation of 44 times earnings seem fair. Nvidia shares would need to increase 1,900% (20-fold) in the next decade to turn $50,000 into $1 million. Returns of that magnitude are theoretically possible in that time frame. In fact, seven stocks currently in S&P 500 (SNPINDEX: ^GSPC) hit that mark in the last decade, as listed: Nvidia: +25,700% Advanced Micro Devices: +4,980% Axon Enterprise: +2,380% Texas Pacific Land: 2,110% Arista Networks: 1,950% Tesla: 1,920% Fair Isaac: 1,900% However, while 20-fold returns are theoretically possible, Nvidia has virtually no chance of hitting that mark in the next decade. The company is already worth $3.4 trillion, meaning its market value would hit $68 trillion if the stock increased 20 times. That seems highly unlikely when the entire S&P 500 is only worth $48 trillion today. Nevertheless, Nvidia is still a worthwhile investment. AI will likely be the most transformative technology in history, and the company is well positioned to benefit as demand for AI infrastructure increases. Potential catalysts include generative AI, autonomous vehicles, and humanoid robots. Also, Nvidia has a burgeoning software business that may evolve into a significant source of revenue as those catalysts take shape. Before you buy stock in Nvidia, consider this: The Motley Fool Stock Advisor analyst team just identified what they believe are the for investors to buy now… and Nvidia wasn't one of them. The 10 stocks that made the cut could produce monster returns in the coming years. Consider when Netflix made this list on December 17, 2004... if you invested $1,000 at the time of our recommendation, you'd have $674,395!* Or when Nvidia made this list on April 15, 2005... if you invested $1,000 at the time of our recommendation, you'd have $858,011!* Now, it's worth noting Stock Advisor's total average return is 997% — a market-crushing outperformance compared to 172% for the S&P 500. Don't miss out on the latest top 10 list, available when you join . See the 10 stocks » *Stock Advisor returns as of June 2, 2025 John Mackey, former CEO of Whole Foods Market, an Amazon subsidiary, is a member of The Motley Fool's board of directors. Trevor Jennewine has positions in Amazon, Arista Networks, Axon Enterprise, CrowdStrike, Nvidia, and Tesla. The Motley Fool has positions in and recommends Advanced Micro Devices, Amazon, Arista Networks, Axon Enterprise, CrowdStrike, Nvidia, and Tesla. The Motley Fool recommends Fair Isaac. The Motley Fool has a disclosure policy. Will $50,000 Invested in Nvidia Stock Be Worth $1 Million in 10 Years? was originally published by The Motley Fool Error in retrieving data Sign in to access your portfolio Error in retrieving data Error in retrieving data Error in retrieving data Error in retrieving data
CNET
24 minutes ago
- CNET
NotebookLM Was Already My Favorite AI Tool, but New Features Keep Making It Even Better
NotebookLM has always been a fun idea -- it's kind of a mini-LLM for all of your personal documents, or really any documents you want to feed it. After taking another look recently, it's definitely more than a diversion. It's become my favorite AI tool ever and something I use nearly every day. Powered by Google's Gemini AI, NotebookLM breaks down complex subjects into an easy-to-understand format and helps you brainstorm new ideas. There's now a mobile app for iOS and Android, and new features were just announced during Google I/O earlier this month. It keeps getting better without feeling like it's becoming overstuffed with features just for the sake of it. NotebookLM isn't just Google Keep stuffed with AI, nor is it just a chatbot that can take notes. It's both and neither. Instead of asking questions to Gemini, only for it to find an answer from the ether of the internet, NotebookLM will only search through the sources that you provide it. It's a dead simple concept that feels like one of the most practical uses of AI, giving way to the perfect study buddy for classes or work. And Google didn't stop there. Now it can do so much more, and will reward your poking around to see what it can do for you. And features like its impressive Audio Overviews have since trickled down into Gemini itself, allowing it to be used in a much wider set of Google's products. Below, I'll cover some of NotebookLM's most interesting features (including the newly announced ones) and how it became one of my favorite AI tools to use. For more, check out Google's smart glasses plans with AndroidXR. What is NotebookLM? NotebookLM is a Gemini-powered note-taking and research assistant tool that can be used in a multitude of ways. It all starts with the sources you feed it, whether they're webpage URLs, YouTube videos or audio clips, allowing you to pull multiple sources together into a cohesive package and bring some organization to your scattered thoughts or notes. The most obvious use case for NotebookLM is using it for school or work. Think of it -- you've kept up with countless classes and typed notes down for every one and even perhaps recorded some lectures. Sifting through everything individually can eventually get you to some semblance of understanding, but what if you could get them to work together? Once you've uploaded your sources, Gemini will get to work to create an overall summary of the material. From there, you can begin asking Gemini questions about specific topics on the sources and information from the sources will be displayed in an easy-to-understand format. This alone may be enough for some people just looking to get the most out of their notes, but that's really just scratching the surface. Available for desktop and mobile NotebookLM's three panel layout NotebookLM/Screenshot by Blake Stimac NotebookLM has been available for a while now on the desktop and is broken into a three-pane layout, consisting of Source, Chat and Studio panels. Both the Source and Studio panels are collapsible, so you can have a full-screen chat experience if you prefer. While the Source and Chat panels are pretty self-explanatory, the Studio panel is where magic can happen (though some of the features can also be created directly from the Chat panel). This is where you can get the most out of your NotebookLM experience. The NotebookLM app is like having a data alchemist in your pocket The mobile app for Android and iOS launched the day before Google I/O 2025. Blake Stimac/CNET Those familiar with the desktop experience will feel right at home with the new mobile apps for iOS and Android. The streamlined app allows you to switch between the Source, Chat and Studio panels via a menu at the bottom. When you go to the view that shows all of your notebooks, you'll see tabs for Recent, Shared, Title and Downloaded. While not everything is on the app yet, it's likely just a matter of time before it matches the web version's full functionality. Audio Overviews If you didn't hear about NotebookLM when it was first announced, you likely did when Audio Overviews were released for it. Once you have at least one source uploaded, you can then opt to generate an Audio Overview, which will provide a "deep dive" on the source material. These overviews are created by none other than Gemini, and it's not just a quick summary of your material in audio format -- it's a full-blown podcast with two "hosts" that break down complex topics into easy-to-understand pieces of information. They're incredibly effective, too, often asking each other questions to dismantle certain topics. The default length of an Audio Overview will vary depending on how much material there is to go over and the complexity of the topic -- though I'm sure there are other factors at play. In my testing, a very short piece of text created a five-minute audio clip, whereas two lengthier and more dense Google Docs documents I uploaded created an 18-minute Overview. If you want a little more control on the length for your Audio Overview, you're in luck. Announced in a blog post during Google I/O earlier this month, users now have three options to choose from: shorter, default and longer. This is perfect if you either want to have a short and dense podcast-like experience of if you want to get into the nitty gritty on a subject with a longer Audio Overview. You can interact with your AI podcasters It gets even better. Last December, NotebookLM got a new design and new ways to interact with Audio Overviews. The customize button allows you to guide the conversation so that key points are covered. Type in your directive and then generate your Audio Overview. Now, if you want to make this feature even more interactive, you can choose the Interactive mode, which is still in beta, to join the conversation. The clip will play, and if you have a particular question in response to something that's said, you can click the join button. Once you do, the speakers will pause and acknowledge your presence and ask you to chime in with thoughts or questions, and you'll get a reply. I wanted to try something a little different, so I threw in the lyrics of a song as the only source, and the AI podcast duo began to dismantle the motivations and emotions behind the words. I used the join feature to point out a detail in the lyrics they didn't touch on, and the two began to dissect what my suggestion meant in the context of the writing. They then began linking the theme to other portions of the text. It was impressive to watch: They handled the emotional weight of the song so well, and tactfully at that. Mind Maps Generating a Mind Map is just one of several powerful features from NotebookLM Google/Screenshot by Blake Stimac I'd heard interesting things about NotebookLM's Mind Map feature, but I wanted to go in blind when I tried it out, so I did a separate test. I took roughly 1,500 words of Homer's Odyssey and made that my only source. I then clicked the Mind Map button, and within seconds, an interactive and categorical breakdown of the text was displayed for me to poke around in. Many of the broken-down sections had subsections for deeper dives, some of which were dedicated to single lines for dissection. Clicking on a category or end-point of the map will open the chat with a prefilled prompt. I chose to dive into the line, "now without remedy," and once clicked, the chat portion of NotebookLM reopened with the prefilled prompt, "Discuss what these sources say about Now without remedy, in the larger context of [the subsection] Alternative (worse)." The full line was displayed, including who said it, what it was in response to and any motivations (or other references) for why the line was said in the text. Study guides and more If the combination of all that Audio Overviews and Mind Maps could do sounds like everything a student might need for the perfect study buddy, NotebookLM has a few other features that will solidify it in that place. Study guides After you've uploaded a source, you can create a quick study guide based on the material that will automatically provide a document with a quiz, potential essay questions, a glossary of key terms and answers for the quiz at the bottom. And if you want, you can even convert the study guide into a source for your notebook. FAQs Whether you're using it for school or want to create a FAQ page for your website, the NotebookLM button generates a series of potentially common questions based on your sources. Timeline If you're looking for a play-by-play sort of timeline, it's built right in. Creating a timeline for the Odyssey excerpt broke down main events in a bulleted list and placed them based on the times mentioned in the material. If an event takes place at an unspecified time, it will appear at the top of the timeline, stating this. A cast of characters for reference is also generated below the timeline of events. Briefing document The briefing document is just what it sounds like, giving you a quick snapshot of the key themes and important events to get someone up to speed. This will include specific quotes from the source and their location. A summary of the material is also created at the bottom of the document. How NotebookLM really 'sold' me I already really liked NotebookLM's concept and execution during its 1.0 days, and revisiting the new features only strengthened my appreciation for it. My testing was mostly for fun and to see how this tool can flex, but using it when I "needed" it helped me really get an idea of how powerful it can be, even for simple things. During a product briefing, I did my typical note-taking: Open a Google Doc, start typing in fragmented thoughts on key points, and hope I could translate what I meant when I needed to refer back to them. I knew I would also receive an official press release, so I wasn't (too) worried about it, but I wanted to put NotebookLM to the test in a real-world situation when I was using it for real -- and not just tinkering, when nearly anything seems impressive when it does what you tell it to. I decided to create a new notebook and make my crude notes (which looked like a series of bad haikus at first glance) the only source, just to see what came out on the other end. Not only did NotebookLM fill in the blanks, but the overall summary read almost as well as the press release I received the following day. I was impressed. It felt like alchemy -- NotebookLM took some fairly unintelligible language and didn't just turn it into something passable, but rather, a pretty impressive description. Funny enough, I've since become a more thorough note-taker, but I'm relieved to know I have something that can save the day if I need it to. Video Overviews are on the way Another feature that was announced during Google I/O was Video Overviews, and it's exactly what it sounds like. There's currently no time frame outside of "coming soon" from the blog post, but it should be a good way to get a more visual experience from your notebooks. We'd previously heard that Video Overviews might be on the way, thanks to some sleuthing from Testing Catalog. The article also mentioned that the ability to make your notebooks publicly available and view an Editor's Picks list of notebooks will eventually make their way to NotebookLM. The Editors Picks feature has yet to rear it's head, but you can indeed now share notebooks directly or make them publicly available for anyone to access. While we're waiting for the new features, here's a preview of a Video Overview below. If you need more from NotebookLM, consider upgrading Most individuals may never have the need to pay for NotebookLM, as the free version is robust enough. But if you're using it for work and need to be able to add more sources or the option to share your notebook with multiple people, NotebookLM Plus is worth considering. It gives you more of everything while introducing more customization, additional privacy and security features as well as analytics. It's worth noting that NotebookLM Plus will also be packaged in with Google's new AI subscriptions. For more, don't miss Google's going all-in on AI video with Flow and Veo 3
