Content Detail

• Content META

  Title	For AI, CIOs Should Concentrate On Data Value, Not Volume
  ID	27738683
  External ID	f46577cc9f1130098795a1aa0511c28b
  Author	Megan Poinski
  Published Date	Nov. 13, 2025, 12:19 p.m.
  Summary	Also in the Forbes CIO newsletter: Data chiefs start the path to better AI use, Google to test data centers in space, AMD predicts skyrocketing revenues.
  Tags
  URL	https://www.forbes.com/sites/cio/2025/11/13/for-ai-cios-should-concentrate-on-data-value-not-volume
  Sentiment	Negative: 0.02121    Neutral: 0.86125    Positive: 0.11754
  Relevance	0.00778

• Content Details

  Content ID	35421446
  Key	952797aa-2379-4c7d-8657-d91325f289da
  Content	AI is a potential game changer for how work gets done in any industry. But in order to work, your organization's data house needs to be in order. In a study shared exclusively with Forbes , IBM surveyed more than 1,700 chief data officers worldwide and found that most of them say their data isn't ready to take full advantage of AI. Just 26% of them feel that the unstructured data at their organizations is good enough to power AI-enabled revenue streams. Most companies are on the right path to be taking full advantage of data. More than four out of five said their data policies were integrated with their company's technology roadmap and infrastructure investments. But that's the beginning of the journey—nowhere near the end. Only a third said they can clearly convey how data leads to business results, and even less have clear measures to determine the value that data-driven outcomes bring to their businesses. Chief data officers are all trying to accelerate AI at their companies, mostly for competitive reasons. A total of 81% are investing in things that allow them to do more with AI, but issues with the data mean that the same proportion are creating more ad hoc solutions that bring AI to the data is—not the other way around. These leaders are also trying to promote an office culture of using data, with nearly three-quarters actively pushing data stewardship among employees. But that culture has proven to be challenging, with nearly half saying that attracting, developing and retaining advanced data skills is a top challenge. What's a data manager to do? The study found that 83% of respondents believe that the potential benefits of AI and AI agents outweigh the risks in their organization. So move forward. Figure out what you want your data and AI to do and prioritize business strategies. Keep giving employees access to data, and continue training them in how to take advantage of it. And look at your proprietary data, which could hold the key to breakthroughs that outrun the competition. Another very important task: Figure out what data you have and get it organized in a way that makes sense to use it for AI. I talked to Maggie Laird, president at data integration and optimization platform Pentaho, about why this is so important and how to tackle the challenge. An excerpt from our conversation is later in this newsletter. We're always looking to improve the stories and format of this newsletter. We'd appreciate your feedback in this brief reader survey. This is the published version of Forbes' CIO newsletter, which offers the latest news for chief innovation officers and other technology-focused leaders. Click here to get it delivered to your inbox every Thursday. NOTABLE NEWS While AMD and Nvidia are making the chips for AI data centers, the data centers themselves present a whole host of other issues. They take up a lot of space, generate a lot of noise, need a lot of water for cooling, and use a lot of energy. Some estimates indicate that by 2030, all of the hyperscale AI data centers now planned across the U.S. could use an additional 81 gigawatts of power by 2030—roughly equivalent to all energy used in Texas today. Google floated a radical solution to the problem: putting them in space. Forbes contributor Anisha Sircar writes about Project Suncatcher, the company's newest moonshot program. Here's how it would work: Clusters of solar-powered satellites equipped with Google's custom AI chips would be launched in a sun-synchronous, low-Earth orbit and linked by ultra-high-bandwidth optical connections to function as a single data center. These satellites would take advantage of the sun, which, according to Senior Director of Google's Paradigms of Intelligence Travis Beals, is the ultimate energy source. It's an inspired idea, but a daunting effort. Satellite launches are expensive, and it will be a challenge to keep a cluster of satellites close enough to function as a single data center. It's also not clear how well Google's AI chips will be able to hold up over the long run in outer space. However, challenges have not stopped Google before. The company said in a blog post it plans to start testing prototypes with satellite data and analytics company Planet in 2027. FROM THE HEADLINES AMD CEO Lisa Su displays an ADM Instinct M1300 chip at CES 2023.Semiconductor stocks have been sky-high, and AMD CEO Lisa Su gave her company's shares more firepower this week. On Tuesday at a financial analyst day event, Su said the chipmaker is entering a 'new era of growth,' and said she thinks AMD could be a leader in what she predicts will be a $1 trillion AI compute market. Su anticipated AMD's revenues could grow 35% annually over the next five years due to 'insatiable' AI chip demand. Data center revenue, she said, could increase 60% over the next three to five years, Yahoo Finance reported. AMD's stock is up more than 10% in the last five days. TECHNOLOGY + INNOVATION An IBM researcher holds a 300mm IBM Quantum Nighthawk wafer. IBM Quantum is coming, and IBM announced this week that big leaps are likely in the very near future. At IBM's annual Quantum Developer Conference in New York, Forbes senior contributor John Koetsier writes the company said it will confirm quantum advantage—verifiable proof that a quantum computer can solve a practical, real-world problem more efficiently, accurately or cost-effectively than a classical computer—by the end of 2026. It will get to fault-tolerant quantum computing—correct quantum operations even when there are errors or faults—by 2029. IBM has been focusing on quantum development in the recent past, announcing at this week's conference that it doubled the speed of chip development, created a more advanced and connective quantum processor, and has made vast improvements in the speed and cost of error correction and mitigation. But IBM isn't the only one at the front of the pack in the quantum race. Other companies ranging in size and capitalization are working on quantum; Koetsier estimates there are about 80 major players worldwide now. There's been a lot of activity on the quantum computing front lately. In Forbes' The Prototype newsletter, my colleague Alex Knapp reports that Google and IonQ both claimed quantum computing milestones last month, while Nvidia recently entered the space with new architecture that allows AI supercomputers to be integrated with quantum processors. BITS + BYTES How To Transform Your Data Management Principles From Preserving Volume To Focusing On Value Pentaho President Maggie Laird. Pentaho, Getty Despite what any excited executive or board member says, your company probably can't just 'do AI.' AI is only as good as the data you've got, and with data stretching back sometimes decades—as well as unstructured data suddenly becoming much more useful—you've got to get your data in order. I talked to Maggie Laird, president at data integration and optimization platform Pentaho, about how to begin this process and keep it going. This conversation has been edited for length, clarity and continuity. How do companies get started taking inventory of what they've got, where it's at, what they need, and how to use it? Laird: The larger companies have data that's really spread out all over the place, so herein lies the challenge. You might go after a certain set of data in your ERP and understand that really well, but chances are, what you really should do is zoom up and look at the whole data state. What do you have in the cloud? What do you have in your systems? You want to take that holistic look because of the hidden cost of multiple pieces of data. The first step is: What do you have? Making sure you get a look across this state, what's in here? That typically requires some automation to scan to see what you have. There's a term in the industry called ROT, and it is redundant, obsolete or trivial. Most companies aren't super excited to do this work, to be honest. [If] you've got a regulatory commitment around some of these policies, you're probably looking at it. Most companies haven't really had to do this because they could just store everything. But this is where going out, understanding what you've got, and the metadata around it, is going to be able to let you know what your optimization opportunities are, and what are the pieces that you want to make your source of truth and build your insights and decisions on—because it's not everything. You then need to classify it: what's good, what's bad, what can you get rid of? And then it's about driving out your insights from quality pieces that you have identified, and having a mechanism to make sure that's continuously governed. As you're setting it up, there are some pretty fantastic stats about how much data is created on a day-to-day-to-day basis, so you understand it. Once you get it organized, you're going to have to continually manage it to keep that data ecosystem fresh, connected and quality. It's a big job, and I think if people are serious about the opportunities that AI is bringing, you're going to have to address this to get further faster. You've got to do the hard work. I know that what's useful is different from company to company, but what is the process for determining what is useful and what is not? And what should companies do with the data that is not useful? We've been in a data volume game, and we've got to get into the data value game. Data value is going to come from getting really clear on what is driving your value. The companies that are more mature have data governance policies around how long do we need to hold the data. If you are in a financial institution, you may have seven or 10 years that it needs to be held onto. In an e-commerce situation where you do the transaction, you may need to hold it for a year and that's it. There's some business contexts that govern how long you need to hold it, but then you can start to be more efficient around where you hold it. This is where it gets kind of interesting because if you do need to hold on to it for a long time, but you're not using it frequently—it just needs to be there to access if there's a legal suit or something happens—then you start to have some choices around where you put it. Do you put it in offline storage? Do you keep it in a cloud? Do you put it in a lower tier of storage? There's cold storage and hot storage, and this really depends upon the frequency that you're going to need to use it, and the performance of what you need. I think this is where companies will be waking up to this: the cost of where they decided to put it—cloud cost, on-prem cost, or archive cost. This is going to be the work of the next decade: Getting those policies in place so that you can more efficiently keep the data that you need in the right kind of tier so that you can use it as the business needs to. That's going to be very dependent upon the business model and the business use case, but there's a lot of optimization I think that will be coming as this moves from just being a budget issue. Every byte of data that you choose to save is going to cost you money. It's going to cost risk. It's also going to cost you carbon. That changes the equation: from this is a budget problem, to this is a brand and a board problem. After you get everything in order, you're still going to produce lots more data. How do you recommend that companies keep track of what comes after this process and make sure that everything stays orderly and useful? This is really important. You don't want to do all this work and then have it be one-and-done. You need a consistent, automated way to maintain and monitor moving forward. What we recommend to companies is capabilities around data cataloging, data governance. This sometimes isn't the most fun, but it's so foundational. The automatic processing that goes through and monitors the quality, it will follow the processes that you've outlined for archiving. You've told it what to do when new data comes in. It's going to be too much data to do this in a manual way. The next layer to that will be as the agentic workforce comes online, they'll be helping in the equation, but you also have to have governance on the agentic automation workflows, too. This just comes down to the automated system that you're going to put in to drive your trust, your ongoing data operations and AI operations. It's becoming more of a requirement for anybody that is investing in AI and wants to take that forward. COMINGS + GOINGS Digital infrastructure company Equinix appointed Douglas Merrill as chief information security officer, effective November 6. Merrill had been working as Equinix's interim CISO for the past six months, and he previously worked as a partner at McKinsey. appointed as chief information security officer, effective November 6. Merrill had been working as Equinix's interim CISO for the past six months, and he previously worked as a partner at McKinsey. Occupational health services provider Concentra selected Jason Cooper as chief data, analytics, and artificial intelligence officer. Cooper most recently held leadership roles at HMS and Paradigm Corp. selected as chief data, analytics, and artificial intelligence officer. Cooper most recently held leadership roles at HMS and Paradigm Corp. Automotive maintenance brand Valvoline tapped Hitesh Patel to be its senior vice president, chief technology and cybersecurity officer, effective November 3. Patel most recently worked as senior vice president, chief information officer of Sleep Number. STRATEGIES + ADVICE As technology rapidly brings major changes to the way organizations do business, leaders have to become experts in change management. Employees often resist change—here's a look behind the curtain on what leads to those feelings—and how you can build trust to guide them through it. It's a good idea to read books, take courses and look at frameworks about leadership, but you need to understand one solid truth: Someone else's 'proven' solution might not work for you. Learn about what others have done, but always be tuned into your own instincts. QUIZ Nvidia's shares slumped this week after Japan's SoftBank disclosed it sold its entire stake in the company. What is SoftBank investing that money in instead? A. The White House ballroom B. SpaceX C. OpenAI D. Databricks See if you got the answer right here.
  Content Reference URL	None

• Publisher

  Publisher Name	Forbes (#1659)
  Publisher URL	https://www.forbes.com
  Description	None
  Trust Flow Score	85
  Citation Flow Score	69

• Data Source

  Data Source Name	NewsCatcher (#303)
  Data Source Type	news
  Data Source URL	https://www.newscatcherapi.com/
  Description
  Weight	1