VentureBeat

Join our daily and weekly newsletters for the latest updates and exclusive content on industry-leading AI coverage. Learn More I have a confession to make: Even though I’ve been a tech journalist for much of my career and consistently rushed to embrace, or at least test out, the latest and greatest in personal technology, I’ve never quite found voice assistants to be useful for me to use regularly. Part of that was because the tech has so far been admittedly pretty clunky: Apple pushed the boundary by acquiring and releasing Siri back in 2011, and was soon joined by the Amazon Echo smart speaker and Alexa voice assistant back in 2014. While I tested and used both, I stopped after a few weeks in both cases because I found myself constantly having to “fight” with the voice interaction — pausing before I spoke a query or repeating myself too often, for example. I know I’m hardly alone, as numerous articles have been written over the last decade-plus about the shortcomings of both of those early voice assistants. But then along came OpenAI with ChatGPT, and its humanlike Advanced Voice Mode audio interaction launched — finally, after a lengthy delay from its first target date — back in September 2024. Many AI power users have remarked about how useful and helpful OpenAI’s ChatGPT Advanced Voice Mode is: able to scour the web for information and carry on full conversations, analyze and react to imagery uploaded to it, even to pause when interrupted and allow the human user to redirect or move the conversation to other topics rapidly — much like real human-to-human conversation. And more recently, OpenAI cofounder and president Greg Brockman reshared a post on X by AI power user and Wharton School of Business professor Ethan Mollick noting that it is possible with newer versions of the Apple iPhone to map its new “action button” to ChatGPT Advanced Voice Mode, allowing users to turn the voice on with one click. I already had Advanced Voice Mode available to me as one of the pinnable and customizable widgets on the top of my iPhone Lock Screen, but this requires you to actually look down at the phone and find the icon. Mapping Advanced Voice Mode to the iPhone’s physical Action Button — the small one, located on the left side of the device — seems like an even more accessible option that requires only feel to activate. I just enabled it and am hopeful it will finally lead me to use this admittedly amazing tech more often. Here’s how I did it: You need a newer iPhone, from 15 Pro and up iPhone 15 Pro iPhone 15 Pro Max iPhone 16 iPhone 16 Plus iPhone 16 Pro iPhone 16 Pro Max The Action Button is a new physical button located on the left edge of newer versions of the iPhone, just above the volume up and down buttons. When setting up a new one of these iPhone models running iOS 18 or later, the phone’s starting process should give you an option to select what you want to use the Action Button for. By default, it is set to “silence” your iPhone’s ringer and notification sounds. However, if you already set up your iPhone and didn’t get this option or use it, don’t fear! You can still adjust it later. Here’s how. Download the official ChatGPT iOS app from the App Store It’s available here. While free ChatGPT users can still access it, there is a variable monthly limit on how many times they can call up Advanced Voice Mode. ChatGPT Plus, Pro, Team, Enterprise and Edu paying subscribers (starting at $20 per month) all have much higher or unlimited caps on the number of Advanced Voice Mode interactions they can access. Once you have the ChatGPT iOS app installed on your iPhone, proceed to the next step. Go to ‘Settings’ to re-assign the iPhone Action Button Tap on the “Settings” gear icon on your home screen. Then, scroll down to the second section of options and you should find “Action Button” listed third, below “General” and “Accessibility.” Tap it. This should open the Action Button assignment application/option on your phone, which is a screen where you can swipe between different options for what the Action Button will do when pressed. Swipe left to move through the options (and right to go back to one) until you reach the option labeled “Shortcut” (it was the ninth screen for me). Tap the up/down arrow selector screen below the text “Shortcut” and it should open yet another selection screen, this time showing a variety of Shortcuts similar to those found in the official Apple iPhone Shortcuts app. Except, if you scroll down below “Get Started” and any custom shortcuts you previously created in the “My Shortcuts” sections, you’ll see a list of third-party app icons that also offer shortcuts. Among them should be the ChatGPT iOS app. Tap this. Finally, this should pull up yet another selection screen showing various different actions within the ChatGPT iOS app that can be mapped to your iPhone’s Action Button. You want to tap to select the one that shows a small headphones icon labeled “Start voice conversation.” Tapping this should bring you back to the main Action Button selection screen, with Shortcut once again displayed prominently and “Start Voice Conversation” now listed as the action. After all that, you can finally swipe up to close the Settings app and long press on the physical action button. It should pull up the familiar Advanced Voice Mode screen showing a blue circle denoting the voice assistant. Begin speaking when you see this circle and you’re off to the races! If you are interested in setting up your iPhone Lock Screen like I have, to also have a tappable onscreen icon to activate ChatGPT Advanced Voice Mode, you can do so pretty easily as well. Read on for how… Once again, you’ll need to have the ChatGPT iOS app downloaded (duh).

Join our daily and weekly newsletters for the latest updates and exclusive content on industry-leading AI coverage. Learn More Nvidia today unveiled Nvidia Project Digits, a personal AI supercomputer that provides AI researchers, data scientists and students worldwide with access to the power of the Nvidia Grace Blackwellplatform. Project Digits features the new Nvidia GB10 Grace Blackwell Superchip, offering a petaflop of AI computing performance for prototyping, fine-tuning and running large AI models. The company made the announcement during CEO Jensen Huang’s opening keynote at CES 2025, the big tech trade show in Las Vegas this week. With Project Digits, users can develop and run inference on models using their own desktop system, then seamlessly deploy the models on accelerated cloud or data center infrastructure. It is based on a “super secret chip called GB110, the smallest Blackwell we can make,” Huang said. Nvidia Project Digits “AI will be mainstream in every application for every industry. With Project Digits, the Grace Blackwell Superchip comes to millions of developers,” said Huang. “Placing an AI supercomputer on the desks of every data scientist, AI researcher and student empowers them to engage and shape the age of AI.” GB10 Superchip provides a petaflop of power-efficient AI performance The GB10 Superchip is a system-on-a-chip (SoC) based on the Nvidia Grace Blackwell architecture. It delivers up to one petaflop of AI performance at FP4 precision. GB10 features an Nvidia Blackwell GPU with latest-generation CUDA cores and fifth-generation Tensor Cores, connected via NVLink-C2C chip-to-chip interconnect to a high-performance Nvidia Grace CPU, which includes 20 power-efficient cores built with the Arm architecture. MediaTek, a market leader in Arm-based SoC designs, collaborated on the design of GB10, contributing to its best-in-class power efficiency, performance and connectivity. The GB10 Superchip enables Project Digits to deliver powerful performance using only a standard electrical outlet. Each Project Digits features 128GB of unified, coherent memory and up to 4TB of NVMe storage. With the supercomputer, developers can run up to 200-billion-parameter large language models. In addition, using Nvidia ConnectX networking, two Project Digits AI supercomputers can be linked to run up to 405-billion-parameter models. Grace Blackwell AI supercomputing within reach There are 72 Blackwell chips on this wafer. With the Grace Blackwell architecture, enterprises and researchers can prototype, fine-tune and test models on local Project Digits systems running Linux-based Nvidia DGX OS, and then deploy them seamlessly on Nvidia DGX Cloud, accelerated cloud instances, or data center infrastructure. This allows developers to prototype AI on Project Digits and then scale on cloud or data center infrastructure, using the same Grace Blackwell architecture and the Nvidia AI Enterprise software platform. Project Digits users can access an extensive library of Nvidia AI software for experimentation and prototyping, including software development kits, orchestration tools, frameworks and models available in the Nvidia NGC catalog and on the Nvidia Developer portal. Developers can fine-tune models with theNvidia NeMo framework, accelerate data science with Nvidia Rapids libraries and run common frameworks such as PyTorch, Python and Jupyter notebooks. To build agentic AI applications, users can also harness Nvidia blueprints and Nvidia NIM microservices, which are available for research, development and testing via the Nvidia Developer Program. When AI applications are ready to move from experimentation to production environments, the Nvidia AI Enterprise license provides enterprise-grade security, support and product releases of Nvidia AI software. Availability Project Digits is coming. Project Digits will be available in May from Nvidia and top partners, starting at $3,000. “By making Llama models open-source, we’re committed to democratizing access to cutting-edge AI technology. With Project Digits, developers can harness the power of Llama locally, unlocking new possibilities for innovation and collaboration,” said Ahmad Al-Dahle, head of GenAI at Meta, in a statement. “Advancing AI requires tools that empower researchers to experiment at scale, speed and precision. Nvidia’s Project Digits represents a significant leap forward. I’m excited to see how 128GB in such a small form factor can advance the future of enterprise AI,” said Silvio Savarese, chief scientist at Salesforce, in a statement. “At Hugging Face, we want to make it easy for developers to build their own AI,” said Je Boudier, head of product at Hugging Face, in a statement. “Nvidia’s Project Digits will empower AI builders to build and run their own Gen AI models and systems at the edge. With 128GB of unified memory, AI builders can run 200B parameter models locally, and connect multiple Project Digits systems to scale from there. I can’t wait to see what the Hugging Face community will build with Nvidia Project Digits.” “Nvidia’s Project Digits is a powerhouse you can hold in the palm of your hand. With two Project Digits units, developers can easily work with AI models up to 405B parameters in size. We can’t wait to see the apps people will build with this,” said Michael Chiang, cofounder of Ollama, in a statement. source

Join our daily and weekly newsletters for the latest updates and exclusive content on industry-leading AI coverage. Learn More OpenAI has taken a more aggressive approach to red teaming than its AI competitors, demonstrating its security teams’ advanced capabilities in two areas: multi-step reinforcement and external red teaming. OpenAI recently released two papers that set a new competitive standard for improving the quality, reliability and safety of AI models in these two techniques and more. The first paper, “OpenAI’s Approach to External Red Teaming for AI Models and Systems,” reports that specialized teams outside the company have proven effective in uncovering vulnerabilities that might otherwise have made it into a released model because in-house testing techniques may have missed them. In the second paper, “Diverse and Effective Red Teaming with Auto-Generated Rewards and Multi-Step Reinforcement Learning,” OpenAI introduces an automated framework that relies on iterative reinforcement learning to generate a broad spectrum of novel, wide-ranging attacks. Going all-in on red teaming pays practical, competitive dividends It’s encouraging to see competitive intensity in red teaming growing among AI companies. When Anthropic released its AI red team guidelines in June of last year, it joined AI providers including Google, Microsoft, Nvidia, OpenAI, and even the U.S.’s National Institute of Standards and Technology (NIST), which all had released red teaming frameworks. Investing heavily in red teaming yields tangible benefits for security leaders in any organization. OpenAI’s paper on external red teaming provides a detailed analysis of how the company strives to create specialized external teams that include cybersecurity and subject matter experts. The goal is to see if knowledgeable external teams can defeat models’ security perimeters and find gaps in their security, biases and controls that prompt-based testing couldn’t find. What makes OpenAI’s recent papers noteworthy is how well they define using human-in-the-middle design to combine human expertise and contextual intelligence on one side with AI-based techniques on the other. “When automated red teaming is complemented by targeted human insight, the resulting defense strategy becomes significantly more resilient,” writes OpenAI in the first paper (Ahmad et al., 2024). The company’s premise is that using external testers to identify the most high-impact real-world scenarios, while also evaluating AI outputs, leads to continuous model improvements. OpenAI contends that combining these methods delivers a multi-layered defense for their models that identify potential vulnerabilities quickly. Capturing and improving models with the human contextual intelligence made possible by a human-in-the-middle design is proving essential for red-teaming AI models. Why red teaming is the strategic backbone of AI security Red teaming has emerged as the preferred method for iteratively testing AI models. This kind of testing simulates a variety of lethal and unpredictable attacks and aims to identify their most potent and weakest points. Generative AI (gen AI) models are difficult to test through automated means alone, as they mimic human-generated content at scale. The practices described in OpenAI’s two papers seek to close the gaps automated testing alone leaves, by measuring and verifying a model’s claims of safety and security. In the first paper (“OpenAI’s Approach to External Red Teaming”) OpenAI explains that red teaming is “a structured testing effort to find flaws and vulnerabilities in an AI system, often in a controlled environment and collaboration with developers” (Ahmad et al., 2024). Committed to leading the industry in red teaming, the company had over 100 external red teamers assigned to work across a broad base of adversarial scenarios during the pre-launch vetting of GPT-4 prior to launch. Research firm Gartner reinforces the value of red teaming in its forecast, predicting that IT spending on gen AI will soar from $5 billion in 2024 to $39 billion by 2028. Gartner notes that the rapid adoption of gen AI and the proliferation of LLMs is significantly expanding these models’ attack surfaces, making red teaming essential in any release cycle. Practical insights for security leaders Even though security leaders have been quick to see the value of red teaming, few are following through by making a commitment to get it done. A recent Gartner survey finds that while 73% of organizations recognize the importance of dedicated red teams, only 28% actually maintain them. To close this gap, a simplified framework is needed that can be applied at scale to any new model, app, or platform’s red teaming needs. In its paper on external red teaming OpenAI defines four key steps for using a human-in-the-middle design to make the most of human insights: Defining testing scope and teams: Drawing on subject matter experts and specialists across key areas of cybersecurity, regional politics, and natural sciences, OpenAI targets risks that include voice mimicry and bias. The ability to recruit cross-functional experts is, therefore, crucial. (To gain an appreciation for how committed OpenAI is to this methodology and its implications for stopping deepfakes, please see our article “GPT-4: OpenAI’s shield against $40B deepfake threat to enterprises.”) Selecting model versions for testing, then iterating them across diverse teams: Both of OpenAI’s papers emphasize that cycling red teams and models using an iterative approach delivers the most insightful results. Allowing each red team to cycle through all models is conducive to greater team learning of what is and isn’t working. Clear documentation and guidance: Consistency in testing requires well-documented APIs, standardized report formats, and explicit feedback loops. These are essential elements for successful red teaming. Making sure insights translate into practical and long-lasting mitigations: Once red teams log vulnerabilities, they drive targeted updates to models, policies and operational plans — ensuring security strategies evolve in lockstep with emerging threats. Scaling adversarial testing with GPT-4T: The next frontier in red teaming AI companies’ red teaming methodologies are demonstrating that while human expertise is resource-intensive, it remains crucial for in-depth testing of AI models. In OpenAI’s second paper, “Diverse and Effective Red Teaming with Auto-Generated Rewards and Multi-Step Reinforcement Learning” (Beutel et al., 2024), OpenAI addresses the challenge of scaling adversarial testing using an automated, multi-pronged approach that combines human insights with AI-generated attack strategies. The core

Join our daily and weekly newsletters for the latest updates and exclusive content on industry-leading AI coverage. Learn More 2025 is poised to be a pivotal year for enterprise AI. The past year has seen rapid innovation, and this year will see the same. This has made it more critical than ever to revisit your AI strategy to stay competitive and create value for your customers. From scaling AI agents to optimizing costs, here are the five critical areas enterprises should prioritize for their AI strategy this year. 1. Agents: the next generation of automation AI agents are no longer theoretical. In 2025, they’re indispensable tools for enterprises looking to streamline operations and enhance customer interactions. Unlike traditional software, agents powered by large language models (LLMs) can make nuanced decisions, navigate complex multi-step tasks, and integrate seamlessly with tools and APIs. At the start of 2024, agents were not ready for prime time, making frustrating mistakes like hallucinating URLs. They started getting better as frontier large language models themselves improved. “Let me put it this way,” said Sam Witteveen, cofounder of Red Dragon, a company that develops agents for companies, and that recently reviewed the 48 agents it built last year. “Interestingly, the ones that we built at the start of the year, a lot of those worked way better at the end of the year just because the models got better.” Witteveen shared this in the video podcast we filmed to discuss these five big trends in detail. Models are getting better and hallucinating less, and they’re also being trained to do agentic tasks. Another feature that the model providers are researching is a way to use the LLM as a judge, and as models get cheaper (something we’ll cover below), companies can use three or more models to pick out the best output to make a decision on.  Another part of the secret sauce? Retrieval-augmented generation (RAG), which allows agents to store and reuse knowledge efficiently, is getting better. Imagine a travel agent bot that not only plans trips but books flights and hotels in real time based on updated preferences and budgets. Takeaway: Businesses need to identify use cases where agents can provide high ROI — be it in customer service, sales, or internal workflows. Tool use and advanced reasoning capabilities will define the winners in this space. 2. Evals: the foundation of reliable AI Evaluations, or “evals,” are the backbone of any robust AI deployment. This is the process of choosing which LLM — among the hundreds now available — to use for your task. This is important for accuracy, but also for aligning AI outputs with enterprise goals. A good eval ensures that a chatbot understands tone, a recommendation system provides relevant options, and a predictive model avoids costly errors. For example, a company’s eval for a customer-support chatbot might include metrics for average resolution time, accuracy of responses, and customer satisfaction scores. A lot of companies have been investing a lot of time into processing inputs and outputs so that they conform to a company’s expectations and workflows, but this can take a lot of time and resources. As models themselves get better, many companies are saving effort by relying more on the models themselves to do the work, so picking the right one gets more important. And this process is forcing clear communication and better decisions. When you “get a lot more conscious of how to evaluate the output of something and what it is that you actually want, not only does that make you better with LLMs and AI, it actually makes you better with humans,” said Witteveen.  “When you can clearly articulate to a human: This is what I want, here’s how I want it to look like, here’s what I’m going to expect in it. When you get really specific about that, humans suddenly perform a lot better.”  Witteveen noted that company managers and other developers are telling him: “Oh, you know, I’ve gotten much better at giving directions to my team just from getting good at prompt engineering or just getting good at, you know, looking at writing the right evals for models.” By writing clear evals, businesses force themselves to clarify objectives — a win for both humans and machines. Takeaway: Crafting high-quality evals is essential. Start with clear benchmarks: response accuracy, resolution time, and alignment with business objectives. This ensures that your AI not only performs but aligns with your brand’s values. 3. Cost efficiency: scaling AI without breaking the bank AI is getting cheaper, but strategic deployment remains key. Improvements at every level of the LLM chain are bringing dramatic cost reductions. Intense competition among LLM providers, and from open-source rivals, is leading to regular price cuts. Meanwhile, post-training software techniques are making LLMs more efficient. Competition from new hardware vendors such as Groq’s LPUs, and improvements by the legacy GPU provider Nvidia, are dramatically reducing inference costs, making AI accessible for more use cases. The real breakthroughs come from optimizing the way models are put to work in applications, which is the time of inference, rather than the time of training, when models are first built using data. Other techniques like model distillation, along with hardware innovations, mean companies can achieve more with less. It’s no longer about whether you can afford AI — you can do most projects much less expensively this year than even six months ago — but how you scale it. Takeaway: Conduct a cost-efficiency analysis for your AI projects. Compare hardware options and explore techniques like model distillation to cut costs without compromising performance. 4. Memory personalization: tailoring AI to your users Personalization is no longer optional — it’s expected. In 2025, memory-enabled AI systems are making this a reality. By remembering user preferences and past interactions, AI can deliver more tailored and effective experiences. Memory personalization isn’t widely or openly discussed because users often feel uneasy about AI applications storing personal information to enhance service. There are privacy

Join our daily and weekly newsletters for the latest updates and exclusive content on industry-leading AI coverage. Learn More Nvidia today announced that its autonomous vehicle (AV) platform, Nvidia Drive AGX Hyperion, has passed industry-safety assessments by TÜV SÜD and TÜV Rheinland — two of the industry’s foremostauthorities for automotive-grade safety and cybersecurity. This achievement raises the bar for AV safety, innovation and performance, Nvidia said during the keynote speech by Nvidia CEO Jensen Huang at CES 2025, the big tech trade show in Las Vegas this week. Drive Hyperion is the industry’s end-to-end autonomous driving platform. It includes the Drive AGX system-on-a-chip (SoC) and reference board design, the Nvidia DriveOS automotive operating system, a sensor suite, and an active safety and level 2+ driving stack. Automotive safety pioneers such as Mercedes-Benz, JLR and Volvo Cars are adopting the platform, which is designed to be modular, so customers can easily use what they need. It is also scalable and built to be upgradeable and compatible across future DRIVE SoC generations. Available in the first half of this year, the latest iteration of Drive Hyperion — designed for both passenger and commercial vehicles — will feature the high-performance Drive AGX Thor SoC built on the Nvidia Blackwell architecture. “A billion vehicles driving trillions of miles each year move the world. With autonomous vehicles — one of the largest robotics markets — now here, the Nvidia Blackwell-powered platform will shift this revolution into high gear,” said Huang. “The next wave of autonomous machines will rely on physical AI world foundation models to understand and interact with the real world, and Nvidia Drive is purpose-built for this new era, delivering unmatched functional safety and AI.” Driving safety forward: certified assurance for next-gen vehicles Nvidia Thor AI chip platform Next-generation vehicles will be increasingly software-defined, capable of receiving new features and functionality over their lifetime. Tapping into Nvidia’s 15,000 engineering years invested in vehicle safety, Drive Hyperion will help ensure advanced automotive systems with rich, AI-based functionalities are compliant with the automotive industry’s stringent functional safety and cybersecurity standards. Nvidia recently received safety certifications and assessments from accredited third parties, including: ● TÜV SÜD, which granted the ISO 21434 Cybersecurity Process certification to Nvidia for automotive SoC, platform and software engineering processes. Additionally, Nvidia DriveOS 6.0 conforms toISO 26262 Automotive Safety Integrity Level (ASIL) D standards, pending certification release.● TÜV Rheinland, which performed an independent United Nations Economic Commission for Europe (UNECE) safety assessment of Nvidia Drive AV related to safety requirements for complex electronicsystems. In addition, Nvidia is now accredited by the ANSI National Accreditation Board (ANAB) to provide safety and cybersecurity inspections for Nvidia Drive ecosystem partners. The new Nvidia Drive AI Systems Inspection Lab will help the Nvidia Drive automotive ecosystem build autonomous driving software that meets the industry’s evolving safety and AI standards. Nvidia is the first platform company to receive a comprehensive set of third-party assessments for its automotive technologies — including the Nvidia Drive end-to-end self-driving platform, spanning SoC, OS, sensor architecture and level 2+ application software — as well as independent accreditation as an AI systems safety and cybersecurity inspection lab for the automotive market. Intelligence powered by industry-leading compute Thus Nvidia has won a key accreditation for its autonomous tech. Nvidia Drive Thor, the core computer for Drive Hyperion, is the successor to the production-proven Nvidia Drive Orin. Its architecture compatibility and scalability means developers can use existing software from earlier Drive product generations, as well as integrate future updates, to achieve seamless development pipelines. Drive Thor is based on the Nvidia Blackwell architecture and is optimized for the most demanding processing workloads, including those involving generative AI, vision language models and large language models. Its simplified architecture enhances generalization, reduces latency and boosts safety by harnessing powerful Nvidia accelerated computing to run the end-to-end AV stack and a proven safety stack in parallel. Drive Thor paves the way for the next era of AV technology, known as AV 2.0, which involves delivering humanlike autonomous driving capabilities for navigating the most complex roadway scenarios. In addition to the Drive AGX in-vehicle computer, two other Nvidia computers serve as the foundation for automotive-grade AV development: Nvidia DGX systems for training advanced AI models and building a robust AV software stack in the cloud, and the Nvidia Omniverse platform running on Nvidia OVX systems for simulation and validation. These three computers, now enhanced with the new Nvidia Cosmos world foundation model platform, are set to accelerate end-to-end AV development and mass deployment. source

Join our daily and weekly newsletters for the latest updates and exclusive content on industry-leading AI coverage. Learn More With 90% of cybersecurity and risk leaders predicting they’ll see budget increases in 2025, many are facing a new era of accountability, with boards wanting to see solid returns on cybersecurity investments. That’s an elusive expectation to deliver on, given that 35.9% of a typical CISO’s budget is going for software. Knowing if, how, when and under what conditions a given cybersecurity software investment delivers a hard-number-based ROI is not easy to do, and such numbers of hard to prove. Clear budget wins do exist, though. They start with automating security operations center (SOC) workflows that are overwhelming analysts with too many conflicting alerts. Automating an endpoint detection and response system is one good place to start, with the goal of reducing alert fatigue in SOCs so analysts can focus on more complex threats and intrusion attempts. Another is automating patch management. CISOs need to move beyond trying to get this done manually with overextended teams, and automate it using the latest AI- and ML-based platforms purpose-built for optimizing patch management network-wide. Forrester’s “Budget Planning Guide 2025: Security and Risk” provides insights into why CISOs are seeing their budgets preserved when other areas of an organization are experiencing layoffs, budget cuts, and, in some cases, new programs being put on hold or canceled altogether. (Note, however, that cybersecurity budgets are, on average, just 5.7% of IT annual spending.) Gartner’s latest forecast update (4Q 2024) of end-user spending for information security reflects the resilience of CISOs’ budgets in the aggregate. These budgets are predicted to grow from $184 billion in 2024 to $294 billion in 2028, and Gartner forecasts the market will grow at a 12.43% compound annual growth rate (CAGR) in four years. Security software is expected to be the fastest-growing segment, consistent with Forrester’s recent findings of CISO spending benchmarks. Gartner predicts spending on security software will grow from $59.9 billion in 2022 to $134.3 billion in 2028, attaining a CAGR of 14.4%. The 10 fastest-growing market segments are outperforming the aggregate market by a slim margin of 12.63%, with cloud security the fastest-growing segment, projected to attain a CAGR of 25.87% from 2024 to 2028.   2025 is shaping up to be the year of CISO fiscal accountability Stephanie Balaouras, Forrester vice president, group director, stated in a recent webinar, “When you think about AI, when you think about some of the novel threats that we’re looking at, when you think about post-quantum encryption, [and] the concerns about that, we are at this inflection point.” Gartner predicts that by 2028, 22% of cyberattacks and data leaks will involve generative AI. Boards aren’t stopping there. While they’re funding the realities of this inflection point by approving security budgets and, in some cases, increasing them, they’re most focused on cutting tech stack sprawl and the expensive licensing fees needed to keep the tech running. Boards’ approval of budgets to improve compliance, reduce AI risks, and reduce tech stack sprawl all hinge on CISOs and their teams delivering this year. Reading between the lines of Forrester’s budget report, we can see that CISOs have entered a new era of accountability. How CISOs are optimizing cybersecurity spending to make the most impact Cloud infrastructure, data, and software are where CISOs are prioritizing their budgets going into 2025, with data-related investments anticipated to make the most significant impact. Forrester sees the increasing adoption of AI and generative AI (gen AI) as driving the needed updates to infrastructure. “Any Gen AI project that we discussed with customers ultimately becomes a data integration project,” says Pascal Matska, vice president and research director at Forrester. “You have to invest into specific capabilities and platforms that run specific AI workloads in the most suitable infrastructure at the right price point, and also drive investments into cloud-native technologies such as Kubernetes and containers and modern data platforms that really are there to help you drive out some of the frictions that exist within the different business silos,” Matska continued. Security and risk leaders are anticipating the most significant changes in their budget next year to be in cloud security, investing in new security technology to run on-premises, and security awareness and training initiatives. Each of those areas is projected to see an increase of 10% or more in 2025 budgets. Protecting revenue is core to CISO accountability One of the most valuable takeaways from Forrester’s cybersecurity planning guide is how essential it is for CISOs to take responsibility for protecting revenue if they want to stand a chance of implementing the guide’s recommendations. VentureBeat continues to see that successful CISOs know how to lead their teams to support and protect revenue, and are often included in board-level discussions and report to the CEO. CISOs who drive gains in revenue advance their careers. “When something touches as much revenue as cybersecurity does, it is a core competency. And you can’t argue that it isn’t,” Jeff Pollard, VP and principal analyst at Forrester, said during his keynote titled “Cybersecurity Drives Revenue: How to Win Every Budget Battle” at the company’s Security and Risk Forum in 2022. Budgeting to protect revenue needs to start with the weakest, most at-risk areas. These include software supply chain security, API security, human risk management, and IoT/OT threat detection. Software supply chains are under siege, with 91% of enterprises falling victim to security incidents in just a year, underscoring the need for better safeguards for continuous integration/continuous deployment (CI/CD) pipelines. Open-source libraries, third-party development tools, and legacy APIs created years ago are just a few threat vectors that make software supply chains and APIs more vulnerable. Persistent attacks on open-source components with wide distribution, including the Log4j vulnerability, are fueling more significant investment in software supply chain security. Where CISOs plan to invest in new technologies Forrester advises CISOs to consider investing in four new technology areas, briefly described below:   Exposure management and cyber

Join our daily and weekly newsletters for the latest updates and exclusive content on industry-leading AI coverage. Learn More Nvidia has launched its Cosmos world foundation model platform to accelerate physical AI development. In a keynote speech at CES 2025 by Nvidia CEO Jensen Huang, the company said the platform includes state-of-the-art generative world foundation models, advanced tokenizers, guardrails and an accelerated video processing pipeline built to advance the development of physical AI systems such as autonomous vehicles (AVs) and robots. Physical AI models are costly to develop, and require vast amounts of real-world data and testing. Cosmos world foundation models, or WFMs, offer developers an easy way to generate massive amounts of photoreal, physics-based synthetic data to train and evaluate their existing models. Developers can also build custom models by fine-tuning Cosmos WFMs. Cosmos models will be available under an open model license to accelerate the work of the robotics and AV community. Developers can preview the first models on the Nvidia API catalog, or download the family of models and fine-tuning framework from the Nvidia NGCTM catalog or Hugging Face. “It is trained on 20 million hours of video,” Huang said. “Nvidia Cosmos. It’s about teaching the AI to understand the physical world.” Cosmos generates synthetic data Leading robotics and automotive companies, including 1X, Agile Robots, Agility, Figure AI, Foretellix, Fourier, Galbot, Hillbot, IntBot, Neura Robotics, Skild AI, Virtual Incision, Waabi, and XPENG, along with ridesharing giant Uber are among the first to adopt Cosmos. “The ChatGPT moment for robotics is coming. Like large language models, world foundation models are fundamental to advancing robot and AV development, yet not alldevelopers have the expertise and resources to train their own,” said Jensen Huang, founder and CEO of Nvidia, in a statement. “We created Cosmos to democratize physical AI and put general robotics in reach of every developer.” Nvidia’s journey to CES 2025 Open world foundation models to accelerate the next wave of AI Nvidia Cosmos’ suite of open models means developers can customize the WFMs with datasets, such as video recordings of AV trips or robots navigating a warehouse, according to the needs of their target application. Cosmos WFMs are purpose-built for physical AI research and development, and can generate physics-based videos from a combination of inputs, like text, image and video, as well as robot sensor or motion data. The models are built for physically based interactions, object permanence, and high-quality generation of simulated industrial environments — like warehouses or factories — and of driving environments, including various road conditions. In his opening keynote at CES, Huang showcased ways physical AI developers can use Cosmos models, including for: Video search and understanding, enabling developers to easily find specific training scenarios, like snowy road conditions or warehouse congestion, from video data. Controllable 3D-to-real synthetic data generation, using Cosmos models to generate photoreal videos from controlled 3D scenarios developed in the Nvidia Omniverse platform. Physical AI model development and evaluation, whether building a custom model on the foundation models, improving the models using Cosmos for reinforcement learning or testing how they perform given a specific simulated scenario. Foresight — the ability to predict the results of a physical AI model’s next potential actions — to help it select the best action to follow. Multiverse simulation, using Cosmos and Omniverse to generate every possible future outcome an AI model could take to help it select the best and most accurate path. Nvidia is marrying tech for AI in the physical world with digital twins. Building physical AI models requires petabytes of video data and tens of thousands of compute hours to process, curate and label that data. To help save enormous costs in data curation, training and model customization, Cosmos features: An Nvidia AI and CUDA-accelerated data processing pipeline, powered by Nvidia NeMo Curator, that enables developers to process, curate and label 20 million hours of videos in 14 days using the Nvidia Blackwell platform, instead of 3.4 years using a CPU-only pipeline. Nvidia Cosmos Tokenizer, a state-of-the-art visual tokenizer for converting images and videos into tokens. It delivers eight times more total compression and 12 times faster processing than today’s leading tokenizers. The Nvidia NeMo framework for highly efficient model training, customization and optimization. World’s largest physical AI industries adopt cosmos Pioneers across the physical AI industry are already adopting Cosmos technologies. 1X, an AI and humanoid robot company, launched the 1X World Model Challenge dataset using Cosmos Tokenizer. XPENG will use Cosmos to accelerate the development of its humanoid robot. And Hillbot and SkildAI are using Cosmos to fast-track the development of their general-purpose robot. “Data-scarcity and variability are key challenges to successful learning in robot environments,” said Pras Velagapudi, chief technology officer at Agility, in a statement. “Cosmos’ text-, image- and video-to-world capabilities allow us to generate and augment photorealistic scenarios in a variety of tasks that we can use to train models without needing as much expensive, real-world data capture.” Transportation leaders are also using Cosmos to build physical AI for AVs. Waabi, a company pioneering generative AI for the physical world, will use Cosmos for the search and curation of video data for AV software development and simulation. Wayve, which is developing AI foundation models for autonomous driving, is evaluating Cosmos as a tool to search for edge and corner case driving scenarios used for safety and validation. AV toolchain provider Foretellix will use Cosmos, alongside Nvidia Omniverse Sensor RTX APIs, to evaluate and generate high-fidelity testing scenarios and training data at scale. Uber is partnering with Nvidia to accelerate autonomous mobility. Rich driving datasets from Uber, combined with the features of the Cosmos platform and Nvidia DGX Cloud, will help AV partners build stronger AI models even more efficiently. “Generative AI will power the future of mobility, requiring both rich data and very powerful compute,” said Dara Khosrowshahi, CEO of Uber. “By working with Nvidia, we are confident that we can help supercharge the timeline for safe and scalable autonomous driving solutions for the

Join our daily and weekly newsletters for the latest updates and exclusive content on industry-leading AI coverage. Learn More Nvidia has announced it has key design wins for autonomous vehicles with car makers such as Toyota, Aurora and Continental. These partners are part of a growing list for Nvidia, and they’re rolling out next-generation highly automated an autonomous vehicle fleets. Jensen Huang, CEO of Nvidia, made the announcements at his opening keynote speech at CES 2025. Nvidia announced today that those companies have joined the list of global mobility leaders developing and building their consumer and commercial vehicle fleets on Nvidia accelerated computing and AI. Toyota, the world’s largest automaker, will build its next-generation vehicles on the high-performance, automotive-grade Nvidia Drive AGX Orin system-on-a-chip (SoC), running the safety-certified Nvidia DriveOS operating system. These vehicles will offer functionally safe, advanced driving assistance capabilities. Nvidia’s Thor chip platform The majority of today’s auto manufacturers, truckmakers, robotaxi and autonomous delivery vehicle companies, tier-one suppliers and mobility startups are developing on Nvidia’s Drive AGX platform and technologies. With cutting-edge platforms spanning from training in the cloud to simulation to compute in the car, Nvidia’s automotive vertical business is expected to grow to approximately $5 billion in fiscalyear 2026. “The autonomous vehicle revolution has arrived, and automotive will be one of the largest AI and robotics industries,” said Huang. “Nvidia is bringing two decades of automotive computing, safety expertise and its CUDA AV platform to transform the multitrillion-dollar auto industry.” Aurora, Continental and Nvidia this week also announced a long-term strategic partnership to deploy driverless trucks at scale, powered by Nvidia Drive. Nvidia’s accelerated compute running DriveOS will be integrated into the Aurora Driver, an SAE level 4 autonomous-driving system that Continental plans to mass-manufacture in 2027. Nvidia Drive Hyperion Other mobility companies adopting Nvidia Drive accelerated compute for their next-generation advanced driver-assistance systems and autonomous vehicle roadmaps include BYD, JLR, Li Auto, Lucid, Mercedes-Benz, NIO, Nuro, Rivian, Volvo Cars, Waabi, Wayve, Xiaomi, ZEEKR, Zoox and many more. Nvidia offers three core computing systems and the AI software essential for end-to-end autonomous vehicle development. One is the Nvidia Drive in-vehicle computer for processing real-time sensor data. The other two are Nvidia DGX systems for training AI models and software stacks, and the Nvidia Omniverse platform running on Nvidia OVX systems for testing and validating self-driving systems in simulation. source

Join our daily and weekly newsletters for the latest updates and exclusive content on industry-leading AI coverage. Learn More Facebook creator and Meta CEO Mark “Zuck” Zuckerberg shook the world again today when he announced sweeping changes to the way his company moderates and handles user-generated posts and content in the U.S. Citing the “recent elections” as a “cultural tipping point,” Zuck explained in a roughly five-minute-long video posted to his Facebook and Instagram accounts this morning (Tuesday, January 7) that Meta would cease using independent third-party fact checkers and fact-checking organizations to help moderate and append notes to user posts shared across the company’s suite of social networking and messaging apps, including Facebook, Instagram, WhatsApp and Threads. Instead, Zuck said that Meta would rely on a “Community Notes” style approach, crowdsourcing information from the users across Meta’s apps to give context and veracity to posts, similar to (and Zuck acknowledged this in his video) the rival social network X (formerly Twitter). Zuck cast the changes as a return to Facebook’s “roots” in free expression, and a reduction in over-broad “censorship.” See the full transcript of his remarks at the bottom of this article. Why this policy change matters to businesses With more than 3 billion users across its services and products worldwide, Meta remains the largest social network to date. In addition, as of 2022, more than 200 million businesses worldwide, most of them small, used the company’s apps and services — and 10 million were active paying advertisers on the platform, according to one executive. Meta’s new chief global affairs officer Joe Kaplan, a former deputy chief of staff for Republican President George W. Bush — who recently took on the role in what many viewed as a signal to lawmakers and the wider world of Meta’s willingness to work with the GOP-led Congress and White House following the 2024 election — also published a note to Meta’s corporate website describing some of the changes in greater detail. Already, some business executives such as Shopify’s CEO Tobi Lutke have seemingly embraced the announcement. As Lutke wrote on X today: “Huge and important change.” Founders Fund chief marketing officer and tech influencer Mike Solana also hailed the move, writing in a post on X: “There’s already been a dramatic decrease in censorship across the [M]eta platforms. but a public statement of this kind plainly speaking truth (the “fact checkers” were biased, and the policy was immoral) is really and finally the end of a golden age for the worst people alive.” However, others are less optimistic and receptive to the changes, viewing them as less about freedom of expression, and more about currying favor with the incoming administration of President-elect Donald J. Trump (to his second non-consecutive term) and the GOP-led Congress, as other business executives and firms have seemingly moved to do. “More free expression on social media is a good thing,” wrote the nonprofit Freedom of the Press Foundation on the social network BlueSky (disclosure: my wife is a board member of the non-profit). “But based on Meta’s track record, it seems more likely that this is about sucking up to Donald Trump than it is about free speech.” George Washington University political communication professor Dave Karpf seemed to agree, writing on BlueSky: “Two salient facts about Facebook replacing its fact-checking program with community notes: (1) community notes are cheaper. (2) the incoming political regime dislikes fact-checking. So community notes are less trouble. The rest is just framing. Zuck’s sole principle is to do what’s best for Zuck.” And Kate Starbird, professor at the University of Washington and cofounder of the UW Center for an Informed Public, wrote on BlueSky that: “Meta is dropping its support for fact-checking, which, in addition to degrading users’ ability to verify content, will essentially defund all of the little companies that worked to identify false content online. But our FB feeds are basically just AI slop at this point, so?” Reached by email, Damian Rollison, Director of Market Insights at AI marketing firm SOCi, also noted that Zuck and Meta appeared by emulating a more libertine approach toward online content moderation championed by X owner Elon Musk: “I think it’s safe to say that no one predicted Elon Musk’s chaotic takeover of Twitter would become a trend other tech platforms would follow, and yet here we are. We can see now in retrospect that Musk established a standard for a newly conservative approach to the loosening of online content moderation, one that Meta has now embraced in advance of the incoming Trump administration. What this will likely mean is that Facebook and Instagram will see a spike in political speech and posts on controversial topics. As with Musk’s X, where ad revenues are down by half, this change may make the platform less attractive to advertisers. It may also cement a trend whereby Facebook is becoming the social network for older, more conservative users and ceding Gen Z to TikTok, with Instagram occupying a middle ground between them.” When will the changes take place? Both Zuck and Kaplan stated in their respective video and text posts that the changes to Meta’s content moderation policies and practices would be coming to the U.S. in “the next couple of months.” Meta will discontinue its independent fact-checking program in the United States, launched in 2016, in favor of a community notes model inspired by X (formerly Twitter). This system will rely on users to write and rate notes, requiring agreement across diverse perspectives to ensure balance and prevent bias. According to its website, Meta had been working with a variety of organizations “certified through the non-partisan International Fact-Checking Network (IFCN) or European Fact-Checking Standards Network (EFCSN) to identify, review and take action” on content deemed “misinformation.” However, as Zuck opined in his video post, “after Trump first got elected in 2016 the legacy media wrote non-stop about how misinformation was a threat to democracy. We tried, in good faith, to address those concerns without becoming the arbiters of

Join our daily and weekly newsletters for the latest updates and exclusive content on industry-leading AI coverage. Learn More Nvidia launched its much-awaited Nvidia GeForce RTX 50 series graphics processing units (GPUs), based on the Blackwell RTX tech. Jensen Huang, CEO of Nvidia, disclosed the news during his opening keynote speech at CES 2025, the big tech trade show in Las Vegas this week. “Blackwell, the engine of AI, has arrived for PC gamers, developers and creatives,” said Huang. “Fusing AI-driven neural rendering and ray tracing, Blackwell is the most significant computer graphics innovation since we introduced programmable shading 25 years ago.” The new RTX Blackwell Neural Rendering Architecture comes with about 92 billion transistors. It has 125 Shader Teraflops of performance 380 RT TFLOPS, 4,000 AI TOPS, 1.8 terabytes per second of memory bandwidth, G7 memory (from Micron) and an AI-management processor. The top SKU has basically over 3,352 trillion AI operations per second (TOPS) of computing power. “The programmable shader is also able to carry neural networks,” Huang said. A neural face rendering. Among the new technologies in this generation are RTX Neural Shaders, DLSS 4, RTX Neural Face rendering to create more realistic human faces, RTX Mega Geometry for rendering environments, and Reflex 2. The DLSS 4 now can generate multiple frames at once thanks to advanced AI technology. That makes for much better frame rates. Nvidia showed that one scene could be rendered at 27 frames per second with the DLSS turned off, with a 71 millisecond PC latency. DLSS 2 can do that scene with its super resolution tech at 71 FPS and PC latency of 34 milliseconds. DLSS 3.5 can do the scene at 140 FPS and 33 milliseconds. But DLSS 4 comes in at a whopping 247 FPS and 34 milliseconds. DLSS 4 is more than eight times better performance than systems that aren’t using AI for the predictive processing. Nvidia’s SKUs include the GeForce RTX 50 Series Desktop Family. It includes the top of the line GPU, the GeForce RTX 5090 coming in at 3,404 AI TOPS and 32GB of G7 memory for $1,999. It also includes the GeForce RTX 5080 at 1,800 AI TOPS and 16GB of G7 memory for $999. The GeForce RTX 5070 Ti (the performance of a 4090) has 1,406 AI TOPS, 16GB of G7 memory for $749 and the GeForce RTX 5070 has 1117 AI TOPS, 12GB of G7 and costs $549. Nvidia also said the GeForce RTX 50 Series will come to laptops with two times efficiency with more performance at half the power compared to the previous generation. It has 40% more battery life with Black Max-Q, two times larger generative AI models, and it is as thin as 14.9 millimeters in terms of laptop thickness. As far as pricing goes, the laptops will come as follows: RTX 5090 at 1,824 AI TOPS and 24GB at $2,899. The RTX 5080 laptops will be at 1,334 AI TOPS, 16GB and $2,199. The RTX 5070 Ti will be 992 AI TOPS, 12GB and $1,599 and the RTX 5070 will be 798 AI TOPS, eight GB and $1,299. Those are steep prices, but they represent the high end of value in GPUs for gaming. Nvidia unveiled its Nvidia GeForce RTX 50 Series graphics chips. Justin Walker, senior director of GeForce products, said in press briefing that Nvidia’s GeForce graphics card brand just celebrated its 25-year anniversary. It was the hit product that helped cement the company’s dominance in the ultra-competitive graphics processing unit (GPU) market and it enabled the company to use graphics as a springboard to AI processing, which is why Nvidia is the most valuable company in the world with a market capitalization of $3.65 trillion. Now, it turns out, Walker said, AI can be used to help accelerate the performance of GPUs. “The great thing about that is that while we are now an AI company, as well as gaming, our gaming side still benefits tremendously from the fact that we are doing AI,” Walker said. And that’s the root of one of the announcements: Nvidia took the wraps of DLSS 4, which uses AI to predict the next pixel that needs to be drawn and then preemptively renders the pixel based on that prediction. The AI TOPS (a measure of AI performance) will be up to 4,000. The new architecture of the 5000 series will have 1.8 terabytes per second of memory bandwidth, and it’s also tapping the Blackwell architecture that is the foundation of Nvidia’s latest AI processors. The new GPU also has neural rendering technologies such as neural shaders. “This is probably the biggest thing to happen in the graphics since programming for shaders, we are actually going to be embedding small neural networks within the shaders itself, and these neural networks can do certain things much more effectively and efficiently than traditional shaders,” Walker said. The tech will enable Nvidia to compress textures eight times to maximize use of memory. The Reflex 2 tech will use predictive shading to reduce the latency between when a gamer creates a movement and it shows up on the screen, so it will be 75% more responsive for gamers. The 5090 series is likely to ship in January and the rest of the systems are going to ship in the March time frame, and the company will say which companies are shipping with the technology later. A number of games like Cyberpunk 2077 can play in 4K resolution at over 200 frames per second. Walker said the company will have a list of games that take advantage of the various features. Nvidia DLSS 4 Boosts Performance by Up to 8 times Nvidia’s DLSS 4 AI tech is paying off. DLSS 4 debuts Multi Frame Generation to boost frame rates by using AI to generate up to three frames per rendered frame. It works in unison with the suite of DLSS technologies to increase performance by up to 8x over traditional rendering, while maintaining responsiveness