Perplexity launches Sonar API, taking aim at Google and OpenAI with real-time AI search

Join our daily and weekly newsletters for the latest updates and exclusive content on industry-leading AI coverage. Learn More Perplexity has launched an aggressive bid to capture the enterprise AI search market, unveiling Sonar, an API service that outperforms offerings from Google, OpenAI and Anthropic on key benchmarks while also undercutting their prices. The move signals a significant shift in the AI landscape, as Perplexity — now valued at $9 billion — directly challenges larger competitors by making its real-time, web-connected search capabilities available to developers and enterprises. The company’s dual-tier strategy — offering both a lightweight Sonar service and a more robust Sonar Pro version — targets different segments of the growing AI integration market. Perplexity’s Sonar Pro outperforms major AI competitors in the SimpleQA benchmark, which measures response accuracy. (Credit: Perplexity) Sonar’s real-time advantage: Bringing fresh data to enterprises Zoom has already integrated Sonar into its AI Companion 2.0 product, allowing users to access real-time information without leaving video conferences — a capability that could reshape how businesses conduct remote meetings and research. The pricing structure appears to be designed to disrupt the market. Sonar’s base tier costs $5 per 1,000 searches plus minimal token fees, while Sonar Pro, despite higher token costs, offers doubled citation density and multi-search capabilities for complex queries. What sets Sonar apart is its real-time web connection, a feature absent in many competing APIs that rely solely on training data. This approach could prove particularly valuable for enterprises requiring current information, although it may face challenges in applications requiring deterministic outputs. Perplexity’s two-tier API offering shows the feature differences between Sonar Pro (left) and the base Sonar service (right), with Pro featuring enhanced citation capability and support for complex queries. (Credit: Perplexity) Disruptive pricing: Affordable AI search for the enterprise market The launch comes at a pivotal moment in the AI industry, when companies are increasingly seeking ways to integrate AI search capabilities into their products. With recent benchmarks showing Sonar Pro achieving an 85.8 F-score on the SimpleQA benchmark — significantly outperforming GPT-4o and Claude — Perplexity appears positioned to capitalize on growing enterprise demand for accurate, citation-backed AI responses. The timing of this launch comes as Perplexity demonstrates significant market momentum, having just secured a $500 million funding round led by Institutional Venture Partners, which valued the company at $9 billion. This strategy could prove particularly effective as enterprises increasingly prioritize AI tools that provide verifiable, current information over black-box solutions. For technical decision makers, Sonar’s launch represents a new option in the AI toolkit, particularly for applications requiring real-time information access and citation tracking. However, the true test will be whether Perplexity can maintain its performance edge and pricing advantage as larger competitors inevitably adjust their strategies. source

Perplexity launches Sonar API, taking aim at Google and OpenAI with real-time AI search Read More »

Consumers Crave More Than Discounts From Loyalty Programs

Last week, my dentist invited me to join their loyalty program. It’s official: The loyalty program is the “go-to” customer relationship marketing tactic. Most global consumers belong to at least one loyalty program, including 90% of online adults in the US, in Europe-5 (88%), and in Australia (93%). According to Forrester’s Consumer Benchmark Survey, 2024, 54% of US online adults agree that loyalty programs influence what they buy, and 64% agree that programs influence where they make purchases. Most agree that loyalty programs make them feel more connected to the brand. To keep members engaged in their program, loyalty marketers must appeal to what consumers really want: Financial rewards. Year after year, consumers rank monetary benefits like instant discounts, loyalty currencies, and exclusive deals from partners at the top of their list of loyalty program perks. Points and discounts are the hallmark of loyalty programs for a reason. They incentivize customers to join and drive incremental behavior that benefits the brand. VIP treatment. Members prioritize financial benefits regardless of region, but they also want a loyalty program that makes them feel special. B2C marketers can do this by providing members with exclusive access to benefits such as limited-release products, first access to deals, and member-only events. Simple loyalty experiences. With so many memberships, it’s easy for consumers to feel overwhelmed by all the rules, offers, and benefits. Maximize consumer participation in a loyalty program with intuitive experiences and personalized updates. For more insights into how consumers feel about loyalty programs, check out our recently published data overview. Questions? We’d love to help you with your loyalty initiatives. Connect with us by scheduling a guidance session. source

Consumers Crave More Than Discounts From Loyalty Programs Read More »

Price Drop: Get Lifetime 1TB of Cloud Storage for Just $130

Pretty much all of the tech giants offer cloud storage nowadays. However, you can easily find yourself shelling out serious money to store your digital data. As a more affordable alternative, Koofr is earning some serious plaudits. This innovative platform lets you upload and access your files with no size limit, and you can even hook up your other online accounts. In a unique offer from TechRepublic Academy, you can pick up a lifetime 1TB subscription for only $129.97 with coupon code KOOFR to be used at checkout. That’s a massive 84% off. Cloud storage is really an essential tool in running any business. Whether it’s simple spreadsheets, promo videos, company logos or even customer data, having a secure online backup of your files is vital. Putting your files in the cloud also means you can work on any device. About Koofr Cloud Storage Koofr provides these benefits and more. This platform allows you to upload and view files on pretty much any device with a browser. This means you can log in on Windows, macOS, Linux, and Chrome laptops along with iOS and Android mobile devices. You can even connect via WebDAV. Koofr’s desktop app makes it easy to manage your data, with smart features like duplicate removal and batch file renaming. The service uses absolutely no trackers, and you can easily connect other online accounts to import your files. Another useful feature for businesses is the ability to share files via custom branded links. This means you can easily go above the file size limit on your email, with the ability to share the same link over and over again. Order today for only $129.97 with code KOOFR to get your lifetime 1TB subscription, normally sold for $810. Prices and availability are subject to change. source

Price Drop: Get Lifetime 1TB of Cloud Storage for Just $130 Read More »

Trump AI plan exposes threat of Europe ‘surrendering’ to big tech

Donald Trump’s big AI announcement has turned heads on both sides of the Atlantic. Trump revealed this week that OpenAI, SoftBank, and Oracle have formed a joint venture — called Stargate — that will invest $500bn in AI infrastructure. The companies said $100bn of the funding was available immediately. The rest would be deployed over the next four years. Trump billed Stargate as “the largest AI infrastructure project by far in history.” He added that the project would ensure “the future of technology” is in the US. Masayoshi Son, the CEO of SoftBank, had another bold prediction. He said the venture would drive  “artificial superintelligence.” The 💜 of EU tech The latest rumblings from the EU tech scene, a story from our wise ol’ founder Boris, and some questionable AI art. It’s free, every week, in your inbox. Sign up now! European tech leaders echoed the sentiment — but fear that the continent will become beholden to American power. David Villalón, the CEO and co-founder of Spanish AI startup Maisa, exemplified this blend of excitement and alarm. “This massive investment shows that the next stage in the growth of AI — Artificial Superintelligence — is no longer a fringe concept but an inevitable reality requiring unprecedented investment in infrastructure, akin to laying down the tracks for the next Industrial Revolution,” he said. Villalón added that the transition requires powerful new computing capacity. Stargate will bring a big dose of this to the US. European tech leaders have called for the continent to react “Without large-scale capital commitments and a bold approach to AI infrastructure, Europe risks surrendering its future to global players who control the fuel of tomorrow,” Villalón said. Europe’s AI worries Villalón pointed to the example of his home country. He believes Spain has “immense potential” in renewables, but needs a major funding boost to remain globally competitive and strategically independent. “Spain is spending peanuts on vacuous publicity AI projects, while ignoring, or not understanding, what’s needed — compute,” he said. Similar anxieties have reverberated across Europe. Jan Marquardt, the CEO of German startup Zivee, warned that AI companies need strong infrastructure, big funding, and minimal regulation, “all of which is available in the USA — and not in Europe.” Christian Klein, the CEO of German tech SAP, added that Stargate should be “a wake-up call” for the continent. Villalón shares their concerns. “To use a football analogy, Europe is currently in the relegation zone while USA and China — with their budgets, quality and ambition — are operating in the Champions League,” he said. source

Trump AI plan exposes threat of Europe ‘surrendering’ to big tech Read More »

Why Enterprises Struggle to Drive Value with AI

Artificial Intelligence is virtually everywhere, whether enterprises have an AI strategy or not. As AI capabilities continue to get more sophisticated, businesses are trying to capitalize on it, but they haven’t done enough foundational work to succeed. While it’s true that companies have been increasing their AI budgets over the last several years, it’s become clear that the ROI of such efforts varies significantly, based on many dynamics, such as available talent, budget, and a sound strategy. Now, organizations are questioning the value of such investments to the point of pulling back in 2025.  According to Anand Rao, distinguished service professor, applied data science and artificial Intelligence at Carnegie Mellon University, the top three challenges are ROI measurement, realization, and maintenance.   “If the work I’m doing takes three hours and now it takes a half an hour, that’s easily quantifiable, [but] human performance is variable,” says Rao. “The second way is having a baseline. We don’t [understand] human performance, but we are saying AI is 95% better than a human, but which human? The top-most performer, an average performer, or the new employee?”  When it comes to realizing ROI, there are different ways to look at it. For example, if AI saves 20% of five peoples’ time, perhaps one could be eliminated. However, if those five people are now spending more time on higher value tasks, then it would be unwise to let any of them go because they are providing more value to the business.  Related:Why Every Employee Will Need to Use AI in 2025 The other challenge is maintenance because AI models need to be monitored and maintained to remain trustworthy. Also, as humans use AI more frequently, they get more adept at doing so while AI is learning from the human, which may increase performance. Enterprises are not measuring that either, Rao says.  “[T]here’s a whole learning curve happening between the human and the AI, and independently the two. That might mean that you may not be able to maintain your ROI, because it may increase or decrease from the base point,” says Rao.   Anand Rao, Carnegie Mellon University There’s also a time element. For example, ChatGPT-4 was introduced in March 2023, but enterprises weren’t ready for it, but in six months or less, businesses had started investing systematically to develop their AI strategy. Nevertheless, there’s still more to do.  [T]he crucial fact is that we are still in the very early days of this technology, and things are moving very quickly,” says Beatriz Sanz Saiz, global consulting data and AI Leader at business management consulting firm EY. “Enterprises should become adept at measuring value realization, risk and safety. CIOs need to rethink a whole set of metrics because they will need to deliver results. Many organizations have a need for a value realization office, so that for everything they do, they can establish metrics upfront to be measured against, whether that is cost savings, productivity, new revenue growth, market share, employee satisfaction [or] customer satisfaction.”  Related:Demand and Supply Issues May Impact AI in 2025 The GenAI Impact  While many enterprises have had plenty of success with traditional AI, Kjell Carlsson, head of AI strategy at enterprise MLOps platform Domino Data Lab, estimates that 90% of GenAI initiatives are not delivering results that move the needle on a sustained basis, nor are they on track to do so.   “[M]ost of these organizations are not going after use cases that can deliver transformative impact, nor do they have the prerequisite AI engineering capabilities to deliver production-grade AI solutions,” says Carlsson. “Many organizations are under the misconception that merely making private instances of LLMs and business apps with embedded GenAI capabilities available to business users and developers is an effective AI strategy. It is not. While there have been productivity gains from these efforts, in most cases, these have been far more modest than expected and have plateaued quickly.”  Related:What Happens if AI No Longer Has Access to Good Data to Train On? Though GenAI has many similarities to driving business value with traditional AI and machine learning, it requires expert teams that can design, develop, operationalize and govern AI applications that rely on complex AI pipelines. These pipelines combine data engineering, prompt engineering, vector stores, guardrails, upstream and downstream ML and GenAI models, and integrations with operational systems.   “Successful teams have evolved their existing data science and ML engineering capabilities into AI product and AI engineering capabilities that allow them to build, orchestrate and govern extremely successful AI solutions,” says Carlsson.  Kjell Carlsson, Domino Data Lab Sound tech strategies identify a business problem and then select the technologies to solve it, but with GenAI, users have been experimenting before they define a problem to solve or expected payoff.   “[W]e believe there is promise of transformation with AI, but the practical path is unclear. This shift has led to a lack of focus and measurable outcomes, and the derailment of plenty of AI efforts in the first wave of AI initiatives,” says Brian Weiss, chief technology officer at hyperautomation and enterprise AI infrastructure company Hyperscience. “In 2025, we anticipate a more pragmatic or strategic approach where generative AI tools will be used to deliver value by attaching to existing solutions with clearly measurable outcomes, rather than simply generating content. [T]he success of AI initiatives hinges on a strategic approach, high-quality data, cross-functional collaboration and strong leadership. By addressing these areas, enterprises can significantly improve their chances of achieving meaningful ROI from their AI efforts.”  Andreas Welsch, founder and chief AI strategist at boutique AI strategy firm Intelligence Briefing, says early in the GenAI hype cycle, organizations were quick to experiment with the technology. Funding was made available, and budgets were consolidated to explore what the technology could offer, but they didn’t need to deliver ROI. Times have changed.  “Organizations who have been stuck in the exploration phase without assessing the business value first, are now caught off guard when the use case does not deliver a measurable return,” says

Why Enterprises Struggle to Drive Value with AI Read More »

Some Good News In The World Of IoT Security: The FCC Launches The US Cyber Trust Mark Program

The US government is doing something positive around IoT security. With the launch of the US Cyber Trust Mark program, the Federal Communications Commission (FCC) authorized a program and developed rules that bring forward a voluntary labeling standard to inform consumers about the cybersecurity impact of wireless IoT devices they may bring into their homes. Consumer IoT devices are scattered everywhere, from doorbell cameras to smart appliances, baby monitors, and streaming devices. And unless consumers take the time to review all the available information online about what these device manufacturers are doing with regards to cybersecurity, they have no idea how a given device manages aspects such as authentication, cryptography, data security, or even device lifespan. This new labeling program gives buyers a quick view on the label of the key cybersecurity functions and a QR code (still need to be careful with those!) that can provide details on how the device manufacturer is addressing the security of the device and the associated data. You may be thinking, “Paddy, this is a good step for consumer devices. How does this impact the security of my business?” That’s a great question. What does the home network of your employees look like? Unless you are security-conscious by nature (or experience), segmented home networks that isolate different devices into their own secured grids are rarer than properly segmented business networks. Compromised IoT devices on the networks of your home/hybrid workers can be used to attack the business devices that your remote employees are attaching to the same home network. Even if you have no remote employees and you don’t allow BYOD laptops, you still must consider mobile device security. Unless every employee has a cell tower in their backyard and/or unlimited data on all mobile device plans, a majority of employees will still connect their smartphone to their home network to save mobile data charges and have a better experience using these mobile devices. While deploying mobile threat defense solutions onto the mobile devices accessing your business resources is a great way to reduce the impact of a compromised IoT device in this manner, security is all about layers, and having more secure IoT devices is a way to assist here. Within your business networks, how many consumer-grade smart appliances are connected? Refrigerators, coffeemakers, microwaves, or even smart assistants litter the networks of many businesses because of their availability and ease of replacement. This type of device labeling can provide security and risk leaders with more details on the impact of these devices on the overall cybersecurity posture of the corporate network. When it comes to commercial IoT devices, there are other initiatives around the world, from guidance to regulation, and within certain markets, there are other requirements and standards that need to be met, such as in healthcare or related to connected vehicles, but like any standard, guideline, or regulation, these should all be seen as the floor to establishing your secured IoT device environment, not the ceiling. With IoT security listed as one of Forrester’s top 10 emerging technologies for 2024, we have a lot of research initiatives going on to bring you, S&R practitioners and leaders, more insight on how to better protect your business when it comes to IoT devices. If you are looking to better protect your organization’s IoT assets, whether they are on your corporate network or your employees’ home network, please schedule an inquiry or guidance session with me to discuss further. source

Some Good News In The World Of IoT Security: The FCC Launches The US Cyber Trust Mark Program Read More »

Private Internet Access VPN Review: How Good Is PIA VPN?

Private Internet Access Fast facts Our rating: 4.5 stars out of 5Pricing: Starts at $3.33 (annual plan)Key features: 10,000-35,000 servers across 91 countries. Customizable VPN experience. Unlimited device connections. Private Internet Access has been a long-time player in the VPN space. It has a massive server fleet that spans across 91 countries and offers fast speeds through its customizable application. Supported platforms include Windows, macOS, Linux, Android, iOS, Smart TVs, and routers. While its operation in the surveillance-heavy United States may sway some users, PIA’s balance of security, speed, and usability make it a strong VPN solution this year. Semperis Employees per Company Size Micro (0-49), Small (50-249), Medium (250-999), Large (1,000-4,999), Enterprise (5,000+) Small (50-249 Employees), Medium (250-999 Employees), Large (1,000-4,999 Employees), Enterprise (5,000+ Employees) Small, Medium, Large, Enterprise Features Advanced Attacks Detection, Advanced Automation, Anywhere Recovery, and more ESET PROTECT Advanced Employees per Company Size Micro (0-49), Small (50-249), Medium (250-999), Large (1,000-4,999), Enterprise (5,000+) Any Company Size Any Company Size Features Advanced Threat Defense, Full Disk Encryption , Modern Endpoint Protection, and more ManageEngine Log360 Employees per Company Size Micro (0-49), Small (50-249), Medium (250-999), Large (1,000-4,999), Enterprise (5,000+) Micro (0-49 Employees), Small (50-249 Employees), Medium (250-999 Employees), Large (1,000-4,999 Employees), Enterprise (5,000+ Employees) Micro, Small, Medium, Large, Enterprise Features Activity Monitoring, Blacklisting, Dashboard, and more Private Internet Access VPN pricing Duration Price 1 year $3.33 per month 3 years (+4 months free) $1.98 per month 1 month $11.95 per month 7-day free trial Free via Android and iOS (account can be used in desktop apps after mobile sign-in) All three subscriptions for PIA come with the same set of features, so you won’t have to worry about missing any key features if you choose one plan over the other. PIA VPN’s one-year subscription is the best value at $3.33 per month and is very affordable compared to other one-year plans offered by competing VPNs. For example, ExpressVPN’s annual subscription costs $6.25 per month, while NordVPN’s Starter one-year plan costs $4.59. Of course, these VPNs bring their own standout features to justify the price, but at face value, PIA VPN’s one-year plan is a bargain. The three-year plan is also very affordable at $1.98 per month, but the $2-dollar spend may not be worth it if you don’t want the three-year-long time investment. Lastly, PIA VPN’s one-month plan falls along the same price range as other VPNs and doesn’t provide the same cost-savings as two former subscriptions. Like most modern VPNs, PIA offers a 30-day money-back guarantee for the three plans. While I wish that PIA had a full-fledged free version (which would be great for first-time VPN users), it provides a seven-day trial through its mobile application (iOS and Android). In my testing, I found that I could use the same account I got through the seven-day mobile free trial and then log into a desktop version of PIA using the same credentials. Having a dedicated desktop free trial is preferred, but the workaround for the free trial was painless and easy to set up. My recommendation would be to fully utilize the seven-day mobile free trial and use it on your desktop or device of choice. This way, you get some hands-on time with the service without spending on an initial payment. Is Private Internet Access VPN safe? PIA VPN has two of the most important security protocols today: OpenVPN and WireGuard. This gives users a good balance of security (with OpenVPN) and speed (with WireGuard). It also uses 256-bit AES encryption, has a reliable kill switch, and protects against DNS leaks. With PIA VPN’s company background, there are drawbacks. First, PIA operates in the United States, which can be a red flag for users who are wary of the country’s surveillance practices. Next is PIA VPN’s ownership. Like CyberGhost VPN, it’s owned by Kape Technologies, which acquired PIA in 2019. Kape (formerly Crossrider) had been associated with distributing malware and adware before it rebranded and started moving into the cybersecurity industry. PIA itself addressed the acquisition and emphasized that the VPN service operates as a separate entity independent of Kape. In terms of addressing security concerns, PIA VPN does have a no-logs policy, which states that it does not keep records of user IP addresses, browsing history, session timestamps and the like. In 2022, this no-logs policy was independently confirmed and verified by Deloitte. Fortunately, PIA VPN has continued its commitment to independent testing as it recently completed a second audit back in April 2024. This audit was also conducted by Deloitte and looked into PIA VPN’s security infrastructure, which includes its server network and network and incident management systems. PIA VPN’s second third-party audit. Image: PIA VPN In my view, third-party audits are essential when choosing a VPN. While providers can promise to keep your data secure, the only real way to confirm security claims is through things like third-party audits. In this respect, I commend PIA VPN for not shying away from independent assessments and continuing to have their service audited. unskippable In addition, PIA VPN utilizes RAM-only servers, which means any possible traces of user data are automatically erased upon reboot. PIA VPN is also an open-source service, providing public access to its source code and allowing privacy enthusiasts to spot vulnerabilities in the code themselves. Lastly, the company has a public Transparency Report that outlines court orders and requests for logs. It reiterates that because the service doesn’t log any information, it doesn’t hand over any data to law enforcement. In my opinion, while PIA’s ownership and US operations are valid concerns, the service has done enough to show that it is a safe and viable VPN service in 2024. Key features of Private Internet Access VPN PIA VPN comes with both industry-standard and unique VPN features for prospective VPN buyers. Let’s take a look at some of PIA’s highlight features. Impressive server fleet PIA VPN’s server selection. Image: Luis Millares PIA VPN offers servers spread out across 91

Private Internet Access VPN Review: How Good Is PIA VPN? Read More »

AI factories are factories: Overcoming industrial challenges to commoditize AI

This article is part of VentureBeat’s special issue, “AI at Scale: From Vision to Viability.” Read more from this special issue here. This article is part of VentureBeat’s special issue, “AI at Scale: From Vision to Viability.” Read more from the issue here. If you were to travel 60 years back in time to Stevenson, Alabama, you’d find Widows Creek Fossil Plant, a 1.6-gigawatt generating station with one of the tallest chimneys in the world. Today, there’s a Google data center where the Widows Creek plant once stood. Instead of running on coal, the old facility’s transmission lines bring in renewable energy to power the company’s online services. That metamorphosis, from a carbon-burning facility to a digital factory, is symbolic of a global shift to digital infrastructure. And we’re about to see the production of intelligence kick into high gear thanks to AI factories.  These data centers are decision-making engines that gobble up compute, networking and storage resources as they convert information into insights. Densely packed data centers are springing up in record time to satisfy the insatiable demand for artificial intelligence.  The infrastructure to support AI inherits many of the same challenges that defined industrial factories, from power to scalability and reliability, requiring modern solutions to century-old problems. The new labor force: Compute power In the era of steam and steel, labor meant thousands of workers operating machinery around the clock. In today’s AI factories, output is determined by compute power. Training large AI models requires massive processing resources. According to Aparna Ramani, VP of engineering at Meta, the growth of training these models is about a factor of four per year across the industry. That level of scaling is on track to create some of the same bottlenecks that existed in the industrial world. There are supply chain constraints, to start. GPUs — the engines of the AI revolution — come from a handful of manufacturers. They’re incredibly complex. They’re in high demand. And so it should come as no surprise that they’re subject to cost volatility.  In an effort to sidestep some of those supply limitations, big names like AWS, Google, IBM, Intel and Meta are designing their own custom silicon. These chips are optimized for power, performance and cost, making them specialists with unique features for their respective workloads. This shift isn’t just about hardware, though. There’s also concern about how AI technologies will affect the job market. Research published by Columbia Business School studied the investment management industry and found the adoption of AI leads to a 5% decline in the labor share of income, mirroring shifts seen during the Industrial Revolution.  “AI is likely to be transformative for many, perhaps all, sectors of the economy,” says Professor Laura Veldkamp, one of the paper’s authors. “I’m pretty optimistic that we will find useful employment for lots of people. But there will be transition costs.” Where will we find the energy to scale? Cost and availability aside, the GPUs that serve as the AI factory workforce are notoriously power-hungry. When the xAI team brought its Colossus supercomputer cluster online in September 2024, it reportedly had access to somewhere between seven and eight megawatts from the Tennessee Valley Authority. But the cluster’s 100,000 H100 GPUs need a lot more than that. So, xAI brought in VoltaGrid mobile generators to temporarily make up for the difference. In early November, Memphis Light, Gas & Water reached a more permanent agreement with the TVA to deliver xAI an additional 150 megawatts of capacity. But critics counter that the site’s consumption is straining the city’s grid and contributing to its poor air quality. And Elon Musk already has plans for another 100,000 H100/H200 GPUs under the same roof. According to McKinsey, the power needs of data centers are expected to increase to approximately three times current capacity by the end of the decade. At the same time, the rate at which processors are doubling their performance efficiency is slowing. That means performance per watt is still improving, but at a decelerating pace, and certainly not fast enough to keep up with the demand for compute horsepower.  So, what will it take to match the feverish adoption of AI technologies? A report from Goldman Sachs suggests that U.S. utilities need to invest about $50 billion in new generation capacity just to support data centers. Analysts also expect data center power consumption to drive around 3.3 billion cubic feet per day of new natural gas demand by 2030. Scaling gets harder as AI factories get larger Training the models that make AI factories accurate and efficient can take tens of thousands of GPUs, all working in parallel, months at a time. If a GPU fails during training, the run must be stopped, restored to a recent checkpoint and resumed. However, as the complexity of AI factories increases, so does the likelihood of a failure. Ramani addressed this concern during an AI Infra @ Scale presentation.  “Stopping and restarting is pretty painful. But it’s made worse by the fact that, as the number of GPUs increases, so too does the likelihood of a failure. And at some point, the volume of failures could become so overwhelming that we lose too much time mitigating these failures and you barely finish a training run.” According to Ramani, Meta is working on near-term ways to detect failures sooner and to get back up and running more quickly. Further over the horizon, research into asynchronous training may improve fault tolerance while simultaneously improving GPU utilization and distributing training runs across multiple data centers.  Always-on AI will change the way we do business Just as factories of the past relied on new technologies and organizational models to scale the production of goods, AI factories feed on compute power, networking infrastructure and storage to produce tokens — the smallest piece of information an AI model uses. “This AI factory is generating, creating, producing something of great value, a new commodity,” said Nvidia CEO Jensen Huang during his Computex 2024 keynote. “It’s

AI factories are factories: Overcoming industrial challenges to commoditize AI Read More »

Enhancing data backup and recovery with AI and ML

In today’s digital age, the need for reliable data backup and recovery solutions has never been more critical. Cyberthreats, hardware failures, and human errors are constant risks that can disrupt business continuity. Addressing these challenges by integrating advanced Artificial Intelligence (AI) and Machine Learning (ML) technologies into data protection solutions can enhance data backup and recovery, providing real-world applications and highlighting the benefits of these technologies. The role of AI and ML in modern data protection AI and ML transform data backup and recovery by analyzing vast amounts of data to identify patterns and anomalies, enabling proactive threat detection and response. According to the Veeam 2024 Data Protection Trends Report, integrating AI and ML into cybersecurity tools is crucial for modern data protection. This integration facilitates real-time monitoring, anomaly detection, and automated responses to potential threats, significantly enhancing an organization’s security posture. For example, Veeam’s AI-driven solutions monitor data environments in real-time, detecting unusual activities that may indicate a cyberthreat, such as unauthorized access attempts or abnormal data transfers. This proactive approach enables swift responses, mitigating potential damage. Additionally, ML algorithms optimize the backup process by learning from historical data, ensuring critical data is protected and readily available for recovery. Impact on backup and recovery efficiency AI and ML can automate many manual tasks traditionally associated with backup and recovery, reducing the risk of human error. Automated systems handle routine tasks such as data validation, backup scheduling, and anomaly detection without human intervention. This ensures backups are performed consistently and accurately, freeing IT staff to focus on more strategic initiatives. Predictive analytics and proactive recovery One significant advantage of AI in backup and recovery is its predictive capabilities. Predictive analytics allows systems to anticipate hardware failures, optimize storage management, and identify potential threats before they cause damage. By learning from historical data, AI systems can predict when a system might fail and automatically initiate preventative measures. This enhances system reliability and ensures data recovery processes are initiated before a failure is fully realized. Improved incident response AI-driven backup and recovery systems significantly improve incident response times. In the event of a system failure or cyberattack, AI can quickly diagnose the issue and execute a predefined recovery plan, minimizing downtime and ensuring business continuity. For example, AI systems can monitor for signs of ransomware attacks by analyzing patterns and detecting unusual data access behaviors, triggering automatic responses to isolate and neutralize threats. Integration with IT operations Modern AI-powered backup solutions integrate seamlessly with broader IT operations. This integration facilitates better visibility and management of the entire IT environment, making it easier for organizations to maintain compliance and ensure data integrity. Tools like Veeam’s AI-driven solutions dynamically adjust backup processes based on current system performance and workload demands, optimizing resource utilization and ensuring that service level agreements (SLAs) are consistently met. Continuous improvement and learning AI and ML systems continuously learn and improve from the data they process. This means backup and recovery systems become more efficient over time, adapting to new threats and operational changes without requiring manual updates. This self-optimization ensures backup processes are always aligned with the latest best practices and technological advancements. Veeam’s approach to addressing integration challenges Veeam provides comprehensive training programs that equip IT professionals with the necessary skills to manage and optimize AI-driven data protection systems. These programs cover various aspects of AI and ML, from basic concepts to advanced implementation techniques. By offering these educational resources, Veeam helps bridge the skills gap, empowering organizations to build the required expertise in-house. Future trends in AI-driven data protection As AI and ML technologies continue to advance, their impact on data protection strategies will only grow. Emerging trends include AI-driven data privacy and compliance, self-healing systems, AI in threat intelligence, decentralized data protection, adaptive backup strategies, and sustainable data protection. Conclusion Integrating AI and ML into data backup and recovery processes enhances how organizations protect their vital information. These technologies streamline data protection, improve recovery times, and fortify defenses against new threats. Veeam’s AI-powered solutions provide a reliable framework, ensuring businesses maintain continuity and resilience. As AI and ML advance, adopting these technologies will equip organizations to handle evolving data challenges and enhance their security measures. With Veeam’s innovative solutions you can leverage the tools necessary to stay ahead in a dynamic digital environment. Learn more about how Veeam is bringing backup into the future with AI. About the author Veeam Dave Russell is Vice President, Enterprise Strategy at Veeam Software. Dave has 33 years of experience in the backup/recovery and storage management industry as a developer (IBM), industry analyst (Gartner) and strategist (IBM and Veeam). At Veeam, Dave is responsible for driving strategic product and go-to-market programs, spearheading industry engagement, and evangelizing Veeam’s vision for Modern Data Protection and Veeam in the Enterprise at key events across the globe. Follow Dave on Twitter @BackupDave, or LinkedIn www.linkedin.com/in/backupdave/. source

Enhancing data backup and recovery with AI and ML Read More »

Taxation With Representation: Simpson Thacher, Covington

By Zak Kostro ( January 17, 2025, 2:24 PM EST) — In this week’s Taxation With Representation, Eli Lilly and Co. buys a precision breast cancer program, Applied Digital Corp. enters a financing agreement for its high-performance computing business, Clearwater Analytics buys Enfusion, and Lantheus Holdings Inc. buys Life Molecular Imaging Ltd…. Law360 is on it, so you are, too. A Law360 subscription puts you at the center of fast-moving legal issues, trends and developments so you can act with speed and confidence. Over 200 articles are published daily across more than 60 topics, industries, practice areas and jurisdictions. A Law360 subscription includes features such as Daily newsletters Expert analysis Mobile app Advanced search Judge information Real-time alerts 450K+ searchable archived articles And more! Experience Law360 today with a free 7-day trial. source

Taxation With Representation: Simpson Thacher, Covington Read More »