In this Future of Data Platforms Summit conversation, Bill Waid, chief product and technology officer at FICO, sits down with theCUBE’s Rob Strechay to unpack how FICO is evolving its data platform to meet the demands of enterprise-scale AI and real-time decision intelligence. Waid details how data has become the foundation of the FICO platform, spanning ingestion, cataloging, linkage, privacy controls and KPI-driven monitoring, to enable transparency and measurable outcomes.
The discussion explores the challenges of building resilient SaaS and PaaS platforms that operate at massive scale while still ensuring consistency, governance and open integration with third-party data sources. Waid emphasizes how FICO’s approach to data management balances resilience, privacy, speed and scale, highlighting the company’s proprietary techniques for unifying disparate data and ensuring consistent business context.
Looking ahead, Waid shares how FICO is investing in R&D to help customers move beyond simply collecting data toward extracting relevant insights tied directly to business outcomes. He explains why features derived from data combinations are the real drivers of predictive power and how linking those features back to key performance indicators enables organizations to continuously refine decisions. From federated architectures and metadata-rich governance to preparing for agentic AI, this conversation offers an inside look at how FICO is shaping the next era of decisioning platforms.
Forgot Password
Almost there!
We just sent you a verification email. Please verify your account to gain access to
Future of Data Platforms Summit. If you don’t think you received an email check your
spam folder.
In order to sign in, enter the email address you used to registered for the event. Once completed, you will receive an email with a verification link. Open this link to automatically sign into the site.
Register For Future of Data Platforms Summit
Please fill out the information below. You will recieve an email with a verification link confirming your registration. Click the link to automatically sign into the site.
You’re almost there!
We just sent you a verification email. Please click the verification button in the email. Once your email address is verified, you will have full access to all event content for Future of Data Platforms Summit.
I want my badge and interests to be visible to all attendees.
Checking this box will display your presense on the attendees list, view your profile and allow other attendees to contact you via 1-1 chat. Read the Privacy Policy. At any time, you can choose to disable this preference.
Select your Interests!
add
Upload your photo
Uploading..
OR
Connect via Twitter
Connect via Linkedin
EDIT PASSWORD
Share
Forgot Password
Almost there!
We just sent you a verification email. Please verify your account to gain access to
Future of Data Platforms Summit. If you don’t think you received an email check your
spam folder.
In order to sign in, enter the email address you used to registered for the event. Once completed, you will receive an email with a verification link. Open this link to automatically sign into the site.
Sign in to gain access to Future of Data Platforms Summit
Please sign in with LinkedIn to continue to Future of Data Platforms Summit. Signing in with LinkedIn ensures a professional environment.
Are you sure you want to remove access rights for this user?
Details
Manage Access
email address
Community Invitation
Molly Presley, Hammerspace
In this Future of Data Platforms Summit conversation, Bill Waid, chief product and technology officer at FICO, sits down with theCUBE’s Rob Strechay to unpack how FICO is evolving its data platform to meet the demands of enterprise-scale AI and real-time decision intelligence. Waid details how data has become the foundation of the FICO platform, spanning ingestion, cataloging, linkage, privacy controls and KPI-driven monitoring, to enable transparency and measurable outcomes.
The discussion explores the challenges of building resilient SaaS and PaaS platforms that operate at massive scale while still ensuring consistency, governance and open integration with third-party data sources. Waid emphasizes how FICO’s approach to data management balances resilience, privacy, speed and scale, highlighting the company’s proprietary techniques for unifying disparate data and ensuring consistent business context.
Looking ahead, Waid shares how FICO is investing in R&D to help customers move beyond simply collecting data toward extracting relevant insights tied directly to business outcomes. He explains why features derived from data combinations are the real drivers of predictive power and how linking those features back to key performance indicators enables organizations to continuously refine decisions. From federated architectures and metadata-rich governance to preparing for agentic AI, this conversation offers an inside look at how FICO is shaping the next era of decisioning platforms.
play_circle_outlineTransforming Infrastructure Strategies: Preparing Organizations for AI Readiness and Addressing Data Capacity Challenges in Today's Landscape
replyShare Clip
play_circle_outlineTransforming AI: Hammerspace's Evolving Metadata Solutions and Data Orchestration Progress as Discussed by Molly Presley
replyShare Clip
play_circle_outline67% of data stored in cloud or hybrid environments, impacting AI deployment strategies.
replyShare Clip
play_circle_outlineFlexibility in data movement is critical for enterprises utilizing GPUs for inference jobs.
In this interview from the Future of Data Platforms Summit, Molly Presley, senior vice president of global marketing at Hammerspace, joins theCUBE’s Rob Strechay to discuss the critical infrastructure challenges facing enterprises as they scale for AI. Presley details how organizations are shifting focus from general storage strategies to specific "data readiness" for AI inference. She explains how Hammerspace is building intelligence and global metadata into the data platform, allowing disparate data across PDFs, large file stores and object stores to be eas...Read more
exploreKeep Exploring
What challenges are organizations facing as they adapt to rapid changes in technology, particularly in relation to AI and data management?add
What is the focus of the data platform in relation to enterprise and government data management?add
What percentage of data is stored in cloud or hybrid environments compared to on-premise, and what is the significance for organizations regarding AI and hybrid cloud inference?add
What are the challenges that language model builders face when transitioning their infrastructure and how do companies like Hammerspace assist in addressing those challenges?add
>> Hello and welcome back to the Future of Data Platform Summit: Update Edition. In this episode, I'm joined by Molly Presley, who's the SVP of Global Marketing for Hammerspace. Welcome onboard, Molly.
Molly Presley
>> Great to be here, Rob.
Rob Strechay
>> I'm so excited. I think, again, there's been so much going on over the last six months since we talked. Actually, not even six months, last five months since we talked last. You guys have been making great strides at Hammerspace in the data platform space. But Molly, organizations are really struggling to make data ready for AI. In fact, scaling for AI is identified as a primary challenge by 65.4% of respondents to the Future of Data Platform Summit survey that we did. What has Hammerspace been working on since we spoke last in August to help address this?
Molly Presley
>> It's interesting that as things are evolving at such a breakneck speed. Organizations are trying to figure out, "Do we continue to do things the old way? Do we experiment with new ideas? How do we accomplish the goals that we have as things are changing so quickly?" And in this idea of being ready for AI, a lot of what organizations are trying to do is figure out, "What is my infrastructure strategy?" That's one thing. "How do I have enough capacity, enough bandwidth, enough GPUs?" That type of thing. But then, "How do I actually have my data in a state where I can do stuff with it?" And when I say do stuff, most enterprises are thinking inference in the enterprise that they know they need data, that data probably needs to be used somewhere other than where it was originally created to do the training to use the GPUs. So, where we're really focused right now, and it's different than storage companies who are trying to build capacity and a place to put the data, we're trying to build the intelligence, the metadata, the information about which data exists in PDFs and large file stores and object stores, have all the enterprise data and have that available through one MCP server that you can then know what data you have. And then, separate from that, activate it by moving it or orchestrating it, which is the current catchphrase people are using, to where you want to use it. So, our data platform focus with enterprises, governments, not so much the language model builders, but the folks who are trying to do inference, is around building that metadata and making it simple to move the appropriate data to where you need it for AI.
Rob Strechay
>> Yeah, I totally agree. I think that when you start to look at it and your comment about, "Hey, it was created somewhere and then it needs to be used somewhere else," is dead on. Like I was saying, in that study that we conducted, really we found that 67% of the data was stored either in cloud or hybrid environments, with 33% still on-premise and a lot of that is the high-ROI data that people want to use as part of their AI. How does this impact organizations that Hammerspace is talking to and particularly around hybrid cloud inference and enterprise AI?
Molly Presley
>> Yeah, absolutely. And you think about organizations over the last 10 or 20 years have been making decisions around, "What is my cloud strategy? What is my data center strategy?" And that's been a little bit more about infrastructure. Do I want my people managing the infrastructure? What is the cost of my infrastructure? But those hybrid decisions now when you're talking about AI are more about where are my GPUs, where are the models? Where is the data I want to run? And that's the change that I think we're seeing as an organization is because of cloud, data is distributed in a lot of different places. Certainly edge computing drives that too. And when we think about what organizations are trying to do with inference, they don't want to have to rework their entire hybrid cloud strategy. Are they cloud first or their data center first strategy? They just want to go do AI. Those infrastructure decisions are a different group, a different conversation in a lot of cases. And that's where Hammerspace gets really interesting. Even with some of the big language model builders, like Meta, when they're shifting from their strategies from Llama 2 to Llama 3 to Llama 4, where their infrastructure was, AKA, where their GPUs were, were different. Some were in their data centers, some were in clouds, and they needed to have flexibility to train where they had power and GPUs, essentially. And Hammerspace, making the data nimble, made that possible. Now, think about enterprises who have their data distributed, as you found in the research, Rob. And what they're really trying to do is, again, generate that global metadata about which data they have. And then, when they run an inference job, generally speaking, most enterprises don't have enough GPUs that they own themselves, so they're going to use a neocloud or traditional large-scale cloud for the GPUs. And what they tend to do, and this is how our customers are using us for inference today, is move the data that's going to be used for an inference job, so orchestrate it up to the cloud GPUs. And then, once it's there, they tend to just place it down in object storage in the cloud. So, the dynamics are shifting that global access to the data, and then very dynamic mobility and automated mobility to GPUs is great. And then, the decision can be made, which infrastructure makes sense to keep the data in once that move has been done. And leaving it in cloud object makes sense, we automate that. If just deleting it and leaving it where originally was built or stored, we can make that possible too. So, it gives a lot of flexibility without having to rework your enterprise's infrastructure decisions.
Rob Strechay
>> Yeah, I agree with that totally. And I think the focus on inference is really where the ROI is going to be in 2026. One of my predictions was that really AI needs to become real and get to inference where ROI can be really attained by organizations. We're hearing that throughout. And, like you said, the bigger models and the model builders are important, but they're not the end-all, be-all of all of this. But I think one of the things that we're also seeing is that we saw that standards are a big deal with 82% of organizations saying that open storage standards are critical to their future strategy for data platforms. What are you seeing in the realm of evolving standards for AI data in particular?
Molly Presley
>> I think the main thing is that in AI, the application, the data user, the model that's using the data most likely isn't what generated the data. So, think of Life Sciences Genomics Institute. You have instruments, like sequencers creating data, you have researchers creating data, and those are not the ones necessarily who are running the AI. And this is where standards are so incredibly important in AI is that unlike the past where the application and the application owners who are creating data know about the data, know about the infrastructure as being written into, the formats it's written to, AI does not. So, we have this many-to-one relationship of users, both machines and humans of data instead of that one-to-one that you had in the past. And so, having a standard format, having it able to be read in a standard way on standards-based machines, instead building out proprietary infrastructure and proprietary client software to read it is really important. And Rob, I know you know the supercomputing space fairly well and probably a fair number of the listeners do too, but in the past you would think about these high-performance compute environments. You've got a supercomputer, dedicated networking, dedicated infrastructure. And if you had some proprietary software and clients and formats in there, it was okay because only the supercomputer was using the data. Now, we have this large scale computing with GPUs where a lot of users and different models are using that data. And so, taking those ideas of anything proprietary being embedded into the infrastructure or the data access makes it impossible to use. So, standards are important and standards can occur in a lot of layers. It can occur with an operating system, within the file system, within the clients. Now, we're seeing a data description, like MCP, going into the standards bodies, but I think we all agree that for AI to scale, it's critical to not have anything proprietary in there.
Rob Strechay
>> Yeah. I love that the whole MCP thing where, again, it becomes just another API into different infrastructure, like you said, and making it simpler way to communicate in a more standard way. I think I say everything that's old is new again kind of thing. This reminds me of when we said API first 10 years ago, and we were really going down this path in a big way because standards are really key. And we also saw in there that people were really leaning into open standards on their data description as well when they're looking at thing like open table formats was a big push that people were looking for on the structured side of things as well, which again, is coming together with the unstructured side. But one of the things I'd be remiss if we didn't talk about cost, it's always a top of mind around data platforms and infrastructure in general. And around 49% of organizations said it was one of their top-three challenges for their data platform was cost. So, it came in forth out of all of the rankings, but still almost half of the 400 plus organizations. One of the things that's going on now, and given the SSD shortage, how can organizations keep strategic initiatives on track despite procurement challenges that they're seeing? Because you may not even be able to get the SSDs that you think you need for the servers or for your arrays.
Molly Presley
>> Yeah. This is such a evolving topic and it seems to be getting worse every day as far as not just what will the cost of SSDs or hard drives, because hard drives are going up too. B, so not just the cost, but will you be able to procure them? And what we're seeing with our customers right now is first they're having to stop and take a look at their architectures and just rationalize, does going and buying more capacity make sense if it's going to cost 3X as much, or whatever the number is at the moment, when they do their purchasing? And a lot of organizations' procurement teams are pushing back and saying, "Hang on a second, that's too much. We have available capacity in these other systems. Why aren't you using that?"
Vanderbilt just recently was on my podcast and is a Hammerspace customer that brought this up. And they're a great case study of procurement just saying, "Hey, until you use up all the capacity you've already bought, you don't get to buy anymore." And that's an architectural challenge because different applications are attached to and designed for the different storage systems. And procurement's like, "I don't care, you have capacity. Figure it out. " And that's a great place where Hammerspace came in and why they selected Hammerspace was we were able to aggregate all that existing capacity off of NAS filers, object stores and aggregate it into available capacity that they could make available to a new application or to the capacity-starved users. And that satisfied procurement's requirement of, "Okay, you've now used everything you have, you can go ahead and buy more even if it's a bit more expensive or a lot more expensive."
So, that's one of the things that we're seeing on this SSD shortage is relook at your architecture. And instead of having that one-to-one ratio of storage systems to applications, maybe you should make that into a global data platform, so you can leverage what you already own. And that goes further than just your filers, or your object stores, or where you have capacity in that area, but also to your tier zero is what we call it, different vendors have different names for it, but the SSDs that are local to your GPU servers or to your compute nodes. A lot of cases, those aren't being used very strategically either. And you may have hundreds of terabytes for a really big environment, petabytes of SSDs that are being unused within your compute nodes, and there's some really great solutions to aggregate those into your globally-available capacity. So, opening up capacity, but that happens to be incredibly high performance, low latency too. And so, those architecture decisions are things that organizations absolutely are looking at before they go pull the trigger on just buying more capacity.
Rob Strechay
>> Yeah. I think that to me is one of the big things is that when you look at your architecture is how do you use what you have smarter? I said, one of my predictions for this year was optimization was going to definitely be one of the buzzwords. It would seem that that's part of that whole tier zero and being able to cash on NVMe, for instance, if you're doing models and things like that, because you're really getting more usage out of it and you're not, per se, having to persist it there. Isn't that a big piece of it?
Molly Presley
>> It is. So, use the capacity when you need it, and that could be because you're doing checkpointing, it could be because you're doing an analytics job, whatever it might be that you're doing that requires performance, but you can also open that up as just an initial landing point of capacity. One of the other language model builders that we worked with had a similar situation. It wasn't because of the shortage, it's just they couldn't rack and stack more capacity fast enough to train the language model they were working on. It was one of the foundation models. And so, they used this capacity as a way to just integrate off of their, I think they had about 1,000-GPU servers, all the capacity in those and make that available as part of their shared storage, just so they could keep working faster. But that analog certainly works when you're worried about cost as well. And so, for sure, just using what you already own, what's the cheapest storage you can have? And we're seeing a lot of kitschy terms like that by different vendors out there right now, yes, the storage you already own. Let's make sure you're using that efficiently and with the applications who need it. And then, keeping in mind that the cloud, I believe, has available capacity now and has their allocation for the year more or less set aside. Being able to really seamlessly have your workloads move between data center and cloud, while in the past six months we would've said that was for access to GPUs, now it may be for access to capacity too, making that possible where your infrastructure, again, isn't making the decisions or holding you back from your AI initiatives. That's really what we're focused on.
Rob Strechay
>> Yeah. No, I think that makes total sense. I really want to thank you for coming onboard, Molly, and joining us on this Future of Data Platforms update. I think this has been great. You guys have been doing some great stuff since we talked last in August, so thanks for coming onboard.
Molly Presley
>> It's great to have a conversation with you, Rob.
Rob Strechay
>> And thank you for joining us to see what's new in data platforms on theCUBE. Stay tuned for more.
Rob Strechay