Exploration of AI integration with MinIO's cloud-native object storage takes center stage in this episode of "Mixture of Experts" hosted at the New York Stock Exchange and facilitated by Dave Vellante. Garima Kapoor, co-founder and co-CEO of MinIO, offers in-depth insights on navigating the AI landscape, particularly addressing critical data infrastructure challenges faced by enterprises shifting towards next-generation, data-intensive workloads.
In this engaging session, Kapoor of MinIO discusses the company's sophisticated approach to data challenges associated with AI. The conversation highlights MinIO's strategic penetration into Fortune 500 companies and the way AI demands are reshaping data architectures. Kapoor emphasizes the crucial role of open formats and public cloud infrastructure in resolving traditional data silos.
Key insights include the need to consolidate data systems and standardize on industry protocols to ensure seamless AI scalability and governance. Kapoor underscores the efficacy of ObjectStore technology in addressing the unique demands of AI and discusses the significant impact of sovereign AI on national security. According to Kapoor, adopting industry standards such as the S3 API is essential for enterprises aiming to future-proof their architectures.
Forgot Password
Almost there!
We just sent you a verification email. Please verify your account to gain access to
theCUBE + NYSE Wired: Mixture of Experts Series. If you don’t think you received an email check your
spam folder.
Sign in to theCUBE + NYSE Wired: Mixture of Experts Series.
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 theCUBE + NYSE Wired: Mixture of Experts Series
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 theCUBE + NYSE Wired: Mixture of Experts Series.
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
theCUBE + NYSE Wired: Mixture of Experts Series. If you don’t think you received an email check your
spam folder.
Sign in to theCUBE + NYSE Wired: Mixture of Experts Series.
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 theCUBE + NYSE Wired: Mixture of Experts Series
Please sign in with LinkedIn to continue to theCUBE + NYSE Wired: Mixture of Experts Series. 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
Garima Kapoor, MinIO
In this theCUBE + NYSE Wired: Mixture of Experts segment from the New York Stock Exchange, theCUBE’s John Furrier sits down with Raj Verma, CEO of SingleStore, to unpack how the intersection of technology and finance is shaping enterprise strategy. Verma shares why SingleStore is “on course” for the public markets, reflects on brand-building through the company’s partnership with golf Hall of Famer Padraig Harrington and connects that ethos to how SingleStore helps organizations fix struggling data “swings.” The discussion zeroes in on what’s next as Wall Street watches the AI infrastructure buildout: after chips and systems, the software and data layers set the pace for value creation.
Verma outlines why enterprises must modernize “brown” data estates into “green” ones to safely bring corporate context, governance and compliance into LLM workflows via RAG – and why commoditized data-at-rest puts the advantage at the query layer that unifies data in motion with data at rest. He predicts agentic AI will gain reasoning capabilities in roughly 18 months, cites industry indicators like Google reporting ~25% of its software now built by AI and argues that high switching costs will give way to disruption as buyers reassess legacy vendors. The conversation closes with concrete momentum: ~33% YoY growth, ARR in the ~$135M range, gross dollar retention ~98%, cloud NDR ~130, ~50% of business now in the cloud, landing ~3 new customers per day, a path to cash-flow breakeven in the next two quarters and a teaser for AI-related announcements in the next two months. Listeners will find notable stats, real-world use cases and forward-looking views on how databases power reliable AI at enterprise scale.
play_circle_outlineMinIO's Rise in Fortune 500: Transforming Storage Technologies and Workloads Through AI Innovation
replyShare Clip
play_circle_outlineOvercoming Challenges in Accelerated Computing: Data Consolidation and Standardization for Successful AI Integration in Organizations
replyShare Clip
play_circle_outlineBenefits of industry standards for portability and minimizing complexities across environments.
replyShare Clip
play_circle_outlinePatterns of AI investment across various industries, particularly manufacturing and finance.
replyShare Clip
play_circle_outlineUpdate on MinIO’s funding and investment strategies moving forward.
>> Hi, everybody. Welcome to the New York Stock Exchange. My name is Dave Vellante and we're here as part of our mixture of expert series, NYSE Wired plus theCUBE. And Garima Kapoor is the co-founder and co-CEO of MinIO. Garima, good to see you again. Thanks for coming in remotely to our studio here at the NYSE.
Garima Kapoor
>> Likewise. No, thank you for having me. It's always a pleasure to speak with you.
Dave Vellante
>> Yeah, a lot's going on. Of course, re:Invent was last week. We heard a lot about storage. You guys are doing well. You've penetrated, I guess, more than half of the Fortune 500, which is pretty good. You got a good start there. But what's driving that penetration? What does that tell us about what's going on here in the AI era?
Garima Kapoor
>> I think a lot is going on. And specifically with respect to AI, that's the one that is driving and pushing the momentum forward for technologies like MinIO, which are built grounds up to address these modern workloads as compared to traditionally how enterprise has been more entrenched towards legacy systems like whether it is SAN or NAS-like systems, which have not been built to address this kind of performance, this kind of scale that AI requires and the simplicity to operate at massive, massive scale. So I think everything combined together is driving the momentum forward for us.
Dave Vellante
>> Nice. Thank you for that. So we talk all the time about organizations, what they have to do to adopt AI. It seems like everybody forgets about the data. Where do you see organizations struggling the most as they shift from sort of the traditional general purpose computing model to the accelerated computing model?
Garima Kapoor
>> If you see there are three pillars to anyone who wants to be successful in this AI space. Of course, firstly, it starts with hardware. Do you have enough compute to drive certain models? Secondly, it's data. And data is where you bring all the data structures together. From enterprise standpoint, traditionally where they have been stuck is that there have been different silos within the organization to address different workloads, but AI practice to be successful within any organization needs to bring the data together at scale to feed these AI models to draw inferencing and to draw the accuracy of inferencing. At certain point of time, the accuracy of the model largely depends on the scale of the right amount of data that you can feed in. So that becomes extremely important. And from storage standpoint, it's important that we break down these silos and make it easier for organizations to start bringing in the data in a single system in open formats that agents can start consuming that kind of data right away to build the intelligent applications on top. So I think applications from enterprise standpoint, the journey is still very much early on just because of so many things are evolving from large language models, from hardware innovation that is happening. So things are still at an early stage, but if enterprises have that view in terms of the first step to being successful in any AI practice is to consolidate the data and put it in open formats, that's a huge big step that can enable them or set them up for success in long-term.
Dave Vellante
>> Okay. Let's unpack that a little bit because you're setting the premise that it's, yes, we'd love to talk about GPUs, but you're saying the data is something that you have to get your house in order before you can actually take advantage of it. And you're talking about consolidating, even though I guess that it's virtually consolidating, and doing so in open table format. So that brings some interesting challenges. And I'm curious as to how you see your clients addressing that because even if you put all the data in a single lakehouse, you've still got silos in that the star schema of the sales data is different from the logistics data, that's different from the supply chain data. So there's more work that has to be done there. You've got to govern those open table formats. So it's a complicated situation. And then you've got to balance performance, you've got object size, you've got latency, you've got the parallelism of GPU. So it's a complicated situation for a lot of organizations. How are they dealing with that and how are you helping them?
Garima Kapoor
>> The good thing is that this problem, all the variables that you just listed, whether it is from the governance standpoint or whether it is consolidating different data formats into open formats, that problem is too much extent solved thanks to the public cloud architecture. That's how AWS has built this architecture. That's how Azure has built. So for enterprises, this is not unknown problems. So if they standardize from iceberg format itself, and that's where the industry is heading towards when it comes to open table formats, that itself solves a lot of that issues when data is written in proprietary format and how do you now bring different tools to understand that data? So that's why it's extremely important to bring the right practices in early on in a journey of any enterprise than to start dealing with the complexity at a much later stage when the scale of the data becomes big, then it becomes a problem that needs to be solved at multiple levels. So it's important to lay down those foundations correctly. And ObjectStore by itself, by the way the technology or how AWS promoted it, it's also meant for immutable data, the data that doesn't change. A lot of the governance and the factors come on per-object level. So even from data retention to access of even particular object, those granularity of the policies can be passed down on per-object level. So there is a lot more control that is completely built in the system itself. So I think standardizing on industry standards is extremely important, whether it is standardizing on iceberg formats from data lake perspective, whether it is standardizing on S3 API when it comes to ObjectStore or your storage infrastructure perspective, because that will free you up from a lot of the complexities because even for future proofing yourself, you can bring in lot more applications directly on your system than in terms of having to deal with different systems having to talk to each other.
Dave Vellante
>> Okay. So thank you for that. So there's consolidating the data and then there's sort of harmonizing it. That's not your job. That's upstream. The software guys are challenged to do that and bring in process data. But I'm still curious as to how you consolidate or unify data when it's like John Furrier says, "AI, everywhere all at once, it's spread out, it's at the edge, it's in the cloud, it's across clouds." So what do you specifically do to help consolidate or unify that data?
Garima Kapoor
>> So I think you're absolutely right in terms of all data will be AI data, or all data is AI data, and it is only right for the organizations to help them bring together in one format or open format. And from MinIO's perspective, what we tell is that any application that works on whether it is on cloud, which is AWS Azure, we do have plugins to bring in that data towards MinIO, which is very seamless without even... If your data is in AWS, then even without a single line of code change, you would be able to bring it to MinIO system. It's that compatible, even compatibilities to the level of error messages. When it comes to Azure or Google, there are connectors in place to bring in the data. When it comes to legacy systems, file systems, there are, again, connectors to make sure that the data is brought into MinIO in a format that is able to understand for the application. So, of course, we have done a lot of plumbing work on our side to make sure that the data transfer happens at a great pace and we are able to saturate the network even if you're doing it remotely from public cloud to MinIO or from legacy systems to MinIO. All of that is completely built into the product itself.
Dave Vellante
>> It was interesting at re:Invent last week to hear Matt Garman even talking about multi-cloud and he said to the analysts, "Hey, look, we evolve. We change." But so a lot of enterprises, and of course we saw this with the outage recently with AWS, they wanted to affect multi-cloud strategy. Some didn't see any impact. Others who maybe didn't have that cross-cloud architecture maybe struggle a little bit more. Now you've got sovereign cloud coming into the whole equation, which is kind of this air-gapped cloud, if you will, or quasi-air gap cloud. So how does MinIO address that? How do you think about it architecturally and what does your product do to help that problem?
Garima Kapoor
>> Again, I think going back to my previous point, if an enterprise standardizes on industry standards, they cannot go wrong. And that is where because AWS was the first one to get started in the cloud space and they are the leader in public cloud environment followed by Azure and Google. But if you stick to industry standards, your life will become extremely easy because then the portability of the applications or the portability of the data even becomes extremely, extremely easy. So if you see AWS, Azure, and Google inherently are incompatible with each other. MinIO software defined that you can run it on prem, that you can run it within public cloud environments and get that S3 compatibility for your storage environment. So if you've deployed MinIO in Azure, if you've deployed MinIO on prem within AWS, you can get that environment in terms of seamless failover of applications if one region is down or if applications are stretching across different cloud environments. So all that is very much what customers do. But again, it comes to setting the architecture correctly because if someone has not thought about a lot of these things early on from architecture level, then it becomes a problem that you have to solve in hindsight and certain decisions get made to just address certain small areas without thinking about the broader picture. So that's why, again, going back to industry standards, making sure that if you are sticking to open standards, industry standards, you just cannot go wrong with that.
Dave Vellante
>> Garima, what are you seeing in terms of sovereign AI? I mean, how real is it? You hear a lot of talk about it. What do you see developing there?
Garima Kapoor
>> It is quite real. It is unbelievable the scale that we are seeing for some of these sovereign clouds and rightfully so. And AI is very unique in a way that you have seen waves in technology earlier as well. There's cloud, there is virtualization, but nothing like AI has happened. And I think AI is going to change everyone's life very meaningfully. And for countries, it's very important to have complete control on their future. And that's where sovereign AI becomes extremely important. It's like your banking infrastructure, it's like your telco infrastructure, and that's where AI really comes into play, the importance of AI really comes into play. So sovereign clouds are very much a reality. Sovereign clouds, every country is right now investing in their own, building their massive data centers. I think India is investing in 10-gig data center, which is unbelievable. So that is something that we are going to see a lot more traction because if you have control on AI, you will have control on your own destiny. And for every country to be able to hold that, that's extremely, extremely important. So this is something that I do believe we are going to see a lot more of it.
Dave Vellante
>> So with the shift to cloud over the last 10 years, a lot of organizations have lost their muscle memory in terms of running data centers. Are there sort of out of-scope expectations for CIOs? Are there misunderstandings or misconceptions about sovereign cloud and sovereign AI that people should be aware of from your perspective?
Garima Kapoor
>> Yeah, I think a lot of those things are outsourced. I think nobody wants to manage their own data centers. That's just the reality of it all. I think there are things that there are players like Equinix, there are Digital Realty. There's, of course, from AI data centers perspective, CoreWeave is investing, neo clouds is investing in that environment. So it is not that a lot of the burden of managing the data centers will fall on the customers itself, but I think the way of operating their applications closer to their data, I think that is going to become a lot more meaningful than just running everything in cloud. If you see 10 years back, there was this huge wave, all the CIOs were like, "Okay, which is your preferred cloud? And let's just move everything there." That's not the case anymore. Now, enterprises are giving it a lot more thought in terms of how their AI stack looks like. And the important thing is they want to have complete control on their software and hardware stack that they deployed on because that is what is going to give them the freedom, freedom from vendor lock-in, freedom to have complete control on their data, and so on and so forth. So I think that is the important part of it. But in terms of day-to-day management of things, a lot of things, those things are outsourced already.
Dave Vellante
>> Are you seeing any particular industry patterns or do you have any customers that you can share within certain industries? Obviously, here in financial services, you're seeing a lot of AI investment there, but seems like it's across the board. Where are you guys having success? Anything you can share?
Garima Kapoor
>> Yeah, it's across the different verticals that we are seeing. Like I said, every enterprise will be an AI enterprise five years from now, and if they're not an AI enterprise, then they don't exist anymore. So it is that meaningful. So across the board, in terms of AI applications from the general readiness, what we see, there is a lot of activity that we see on the manufacturing side, especially as the data gets generated, there's computer vision applications that leverage MinIO in terms to drive the efficiency and the cost at the end of it for defective items and so on, so forth. So that's one use case in manufacturing that we see a lot of repeatability on. Then government is another vertical that we see a lot of repeatability from them deploying MinIO at mission-critical sites to bringing the data to a centralized location, which is again, MinIO, and being able to completely connect both that infrastructure together and be able to run high-performance AI and analytics on top of it. So that's another use case that we see. In finance, of course, the range is very broad because you have fraud detection, analytics that get run on MinIO. Then of course there are analytics and AI stack that gets run on MinIO from generating even the financial reports, and so on. So the use cases are very diverse that we see, but I think AI's nature is that it's going to be much more horizontal in nature. Then of course there are bigger LLM models that use MinIO underneath them just to train their models day in and out. So the use cases are very diverse, but we do see a repeatability from training inferencing computer vision applications and running mission-critical workloads on MinIO.
Dave Vellante
>> When I think about ObjectStore, the GET/PUT paradigm, really simple, obviously S3 popularized it. How does that fit with AI? I mean, obviously from a scale standpoint, it's probably a good thing. You got all this data, a lot of stuff still in files. How do you see the simpatico between object and AI?
Garima Kapoor
>> So AI is built on object. If you see the models, OpenAI, GPT models are built on ObjectStore or Anthropic, they're all built on ObjectStore natively. And the reason is that because ObjectStore gives them the scale, the object score. Also, the kind of data that they work on, it's all unstructured data. The data is immutable in nature. The data is best suited to run on the object storage workloads because legacy file systems, if you see those protocols are very chatty protocols. And rightfully so because of the requirement back in the day was to be able to operate the file, open the file, operated light, hold the locks and the other environment. And then there is lot of things that needs to get managed from a file system standpoint, which object storage lets go of that legacy because it's not needed anymore in this new way of doing training and inferencing world. So if you see most of the LLM models, they are built natively on ObjectStore by default. Cloud is built on ObjectStore. So those are the practices that are now coming back to enterprise as they're looking to modernize their environment, how they should invest in their storage stack and what their cloud-native architecture needs to look like.
Dave Vellante
>> So let's talk about your business a little bit. You've got some big backers, Intel, Masa Son is in there, Dell Tech, Capital, General Catalyst, and a number of others. I think you've raised them right about $120 million, which is, I mean, it's not an enormous amount of money considering you started mid last decade, so you're pretty capital-efficient. So how's it going? How are the investors doing? Are they happy? You're raising more money, you're going public. What's the plan?
Garima Kapoor
>> It's going well. I mean, we couldn't have asked for a better timing and we raised our Series B, I think we closed it in Jan '22, and we still haven't touched a single dollar off it just because the customer traction and the customer money is the best money and validation that any organization can ask for. In terms of fundraising, the next round that we are planning is going to be more strategic in nature to help the company elevate to another level as we are starting to invest ourselves in the AI stack and how that is going to look like. So that is going to shift in terms of how MinIO is going to be consumed in the market overall. So I think that is where we are looking to get the strategic investment, but right now we are not in a fundraising mode or anything.
Dave Vellante
>> So you have a pretty interesting career. I'm checking out your LinkedIn here. You got your PhD, you were doing angel investing, you have friends at Treasure Data. I remember them from the big data world, Sri Ambati's company, H2O, you invested there, and a very active angel investor. Then you came to MinIO as the CFO, and then you're heading operations, COO, and then now you're co-CEO and a co-founder. So that's a unique path. I wonder if you could describe how you got here, and would you advise somebody taking a similar path?
Garima Kapoor
>> I think for me, being founder is the best title that anyone can ask for. And that's the best thing that I did for myself to start my own startup. And in a journey of the startup, as a founder, you get to do multiple things and so on. So other titles are not that exciting other than the founder one because you are always a builder, you're creating value for your shareholders, you're creating value for the overall industry, and so on, and creating a world that did not exist before. So being a founder is extremely gratifying. And I think for anyone I would say, and especially I get asked this quite a bit as with AI being the centerfold of a lot of things that we do, how should anyone think about their career? I always tell them that think about more from a leadership level and how you want things to evolve because the roles are going to merge completely, whether it is leadership and IC roles are going to get merged pretty aggressively. We are already seeing this in the software space. So for anyone, I would just say that be a creator, be a builder. I think there is nothing as gratifying as that. And if you are a doer, you just cannot go wrong. I think that's the way to do things and rest, everything will fall in place.
Dave Vellante
>> I love it. Well, congratulations on building and best of luck in the future. Thanks so much for coming into our remote studio here at the NYSC Wired and theCUBE, and hopefully see you face to face in 2026.
Garima Kapoor
>> Absolutely. Thanks. Thanks, David.
Dave Vellante
>> Thank you. Garima Kapoor at MinIO. Thank you very much for watching. My name is Dave Vellante. John Furrier is here. Keep it right there for more action from our mixture of expert series from the NYSE right here on theCUBE.
>> Hi, everybody. Welcome to the New York Stock Exchange. My name is Dave Vellante and we're here as part of our mixture of expert series, NYSE Wired plus theCUBE. And Garima Kapoor is the co-founder and co-CEO of MinIO. Garima, good to see you again. Thanks for coming in remotely to our studio here at the NYSE.
Garima Kapoor
>> Likewise. No, thank you for having me. It's always a pleasure to speak with you.
Dave Vellante
>> Yeah, a lot's going on. Of course, re:Invent was last week. We heard a lot about storage. You guys are doing well. You've penetrated, I guess, more than half of the Fortune 500, which is pretty good. You got a good start there. But what's driving that penetration? What does that tell us about what's going on here in the AI era?
Garima Kapoor
>> I think a lot is going on. And specifically with respect to AI, that's the one that is driving and pushing the momentum forward for technologies like MinIO, which are built grounds up to address these modern workloads as compared to traditionally how enterprise has been more entrenched towards legacy systems like whether it is SAN or NAS-like systems, which have not been built to address this kind of performance, this kind of scale that AI requires and the simplicity to operate at massive, massive scale. So I think everything combined together is driving the momentum forward for us.
Dave Vellante
>> Nice. Thank you for that. So we talk all the time about organizations, what they have to do to adopt AI. It seems like everybody forgets about the data. Where do you see organizations struggling the most as they shift from sort of the traditional general purpose computing model to the accelerated computing model?
Garima Kapoor
>> If you see there are three pillars to anyone who wants to be successful in this AI space. Of course, firstly, it starts with hardware. Do you have enough compute to drive certain models? Secondly, it's data. And data is where you bring all the data structures together. From enterprise standpoint, traditionally where they have been stuck is that there have been different silos within the organization to address different workloads, but AI practice to be successful within any organization needs to bring the data together at scale to feed these AI models to draw inferencing and to draw the accuracy of inferencing. At certain point of time, the accuracy of the model largely depends on the scale of the right amount of data that you can feed in. So that becomes extremely important. And from storage standpoint, it's important that we break down these silos and make it easier for organizations to start bringing in the data in a single system in open formats that agents can start consuming that kind of data right away to build the intelligent applications on top. So I think applications from enterprise standpoint, the journey is still very much early on just because of so many things are evolving from large language models, from hardware innovation that is happening. So things are still at an early stage, but if enterprises have that view in terms of the first step to being successful in any AI practice is to consolidate the data and put it in open formats, that's a huge big step that can enable them or set them up for success in long-term.
Dave Vellante
>> Okay. Let's unpack that a little bit because you're setting the premise that it's, yes, we'd love to talk about GPUs, but you're saying the data is something that you have to get your house in order before you can actually take advantage of it. And you're talking about consolidating, even though I guess that it's virtually consolidating, and doing so in open table format. So that brings some interesting challenges. And I'm curious as to how you see your clients addressing that because even if you put all the data in a single lakehouse, you've still got silos in that the star schema of the sales data is different from the logistics data, that's different from the supply chain data. So there's more work that has to be done there. You've got to govern those open table formats. So it's a complicated situation. And then you've got to balance performance, you've got object size, you've got latency, you've got the parallelism of GPU. So it's a complicated situation for a lot of organizations. How are they dealing with that and how are you helping them?
Garima Kapoor
>> The good thing is that this problem, all the variables that you just listed, whether it is from the governance standpoint or whether it is consolidating different data formats into open formats, that problem is too much extent solved thanks to the public cloud architecture. That's how AWS has built this architecture. That's how Azure has built. So for enterprises, this is not unknown problems. So if they standardize from iceberg format itself, and that's where the industry is heading towards when it comes to open table formats, that itself solves a lot of that issues when data is written in proprietary format and how do you now bring different tools to understand that data? So that's why it's extremely important to bring the right practices in early on in a journey of any enterprise than to start dealing with the complexity at a much later stage when the scale of the data becomes big, then it becomes a problem that needs to be solved at multiple levels. So it's important to lay down those foundations correctly. And ObjectStore by itself, by the way the technology or how AWS promoted it, it's also meant for immutable data, the data that doesn't change. A lot of the governance and the factors come on per-object level. So even from data retention to access of even particular object, those granularity of the policies can be passed down on per-object level. So there is a lot more control that is completely built in the system itself. So I think standardizing on industry standards is extremely important, whether it is standardizing on iceberg formats from data lake perspective, whether it is standardizing on S3 API when it comes to ObjectStore or your storage infrastructure perspective, because that will free you up from a lot of the complexities because even for future proofing yourself, you can bring in lot more applications directly on your system than in terms of having to deal with different systems having to talk to each other.
Dave Vellante
>> Okay. So thank you for that. So there's consolidating the data and then there's sort of harmonizing it. That's not your job. That's upstream. The software guys are challenged to do that and bring in process data. But I'm still curious as to how you consolidate or unify data when it's like John Furrier says, "AI, everywhere all at once, it's spread out, it's at the edge, it's in the cloud, it's across clouds." So what do you specifically do to help consolidate or unify that data?
Garima Kapoor
>> So I think you're absolutely right in terms of all data will be AI data, or all data is AI data, and it is only right for the organizations to help them bring together in one format or open format. And from MinIO's perspective, what we tell is that any application that works on whether it is on cloud, which is AWS Azure, we do have plugins to bring in that data towards MinIO, which is very seamless without even... If your data is in AWS, then even without a single line of code change, you would be able to bring it to MinIO system. It's that compatible, even compatibilities to the level of error messages. When it comes to Azure or Google, there are connectors in place to bring in the data. When it comes to legacy systems, file systems, there are, again, connectors to make sure that the data is brought into MinIO in a format that is able to understand for the application. So, of course, we have done a lot of plumbing work on our side to make sure that the data transfer happens at a great pace and we are able to saturate the network even if you're doing it remotely from public cloud to MinIO or from legacy systems to MinIO. All of that is completely built into the product itself.
Dave Vellante
>> It was interesting at re:Invent last week to hear Matt Garman even talking about multi-cloud and he said to the analysts, "Hey, look, we evolve. We change." But so a lot of enterprises, and of course we saw this with the outage recently with AWS, they wanted to affect multi-cloud strategy. Some didn't see any impact. Others who maybe didn't have that cross-cloud architecture maybe struggle a little bit more. Now you've got sovereign cloud coming into the whole equation, which is kind of this air-gapped cloud, if you will, or quasi-air gap cloud. So how does MinIO address that? How do you think about it architecturally and what does your product do to help that problem?
Garima Kapoor
>> Again, I think going back to my previous point, if an enterprise standardizes on industry standards, they cannot go wrong. And that is where because AWS was the first one to get started in the cloud space and they are the leader in public cloud environment followed by Azure and Google. But if you stick to industry standards, your life will become extremely easy because then the portability of the applications or the portability of the data even becomes extremely, extremely easy. So if you see AWS, Azure, and Google inherently are incompatible with each other. MinIO software defined that you can run it on prem, that you can run it within public cloud environments and get that S3 compatibility for your storage environment. So if you've deployed MinIO in Azure, if you've deployed MinIO on prem within AWS, you can get that environment in terms of seamless failover of applications if one region is down or if applications are stretching across different cloud environments. So all that is very much what customers do. But again, it comes to setting the architecture correctly because if someone has not thought about a lot of these things early on from architecture level, then it becomes a problem that you have to solve in hindsight and certain decisions get made to just address certain small areas without thinking about the broader picture. So that's why, again, going back to industry standards, making sure that if you are sticking to open standards, industry standards, you just cannot go wrong with that.
Dave Vellante
>> Garima, what are you seeing in terms of sovereign AI? I mean, how real is it? You hear a lot of talk about it. What do you see developing there?
Garima Kapoor
>> It is quite real. It is unbelievable the scale that we are seeing for some of these sovereign clouds and rightfully so. And AI is very unique in a way that you have seen waves in technology earlier as well. There's cloud, there is virtualization, but nothing like AI has happened. And I think AI is going to change everyone's life very meaningfully. And for countries, it's very important to have complete control on their future. And that's where sovereign AI becomes extremely important. It's like your banking infrastructure, it's like your telco infrastructure, and that's where AI really comes into play, the importance of AI really comes into play. So sovereign clouds are very much a reality. Sovereign clouds, every country is right now investing in their own, building their massive data centers. I think India is investing in 10-gig data center, which is unbelievable. So that is something that we are going to see a lot more traction because if you have control on AI, you will have control on your own destiny. And for every country to be able to hold that, that's extremely, extremely important. So this is something that I do believe we are going to see a lot more of it.
Dave Vellante
>> So with the shift to cloud over the last 10 years, a lot of organizations have lost their muscle memory in terms of running data centers. Are there sort of out of-scope expectations for CIOs? Are there misunderstandings or misconceptions about sovereign cloud and sovereign AI that people should be aware of from your perspective?
Garima Kapoor
>> Yeah, I think a lot of those things are outsourced. I think nobody wants to manage their own data centers. That's just the reality of it all. I think there are things that there are players like Equinix, there are Digital Realty. There's, of course, from AI data centers perspective, CoreWeave is investing, neo clouds is investing in that environment. So it is not that a lot of the burden of managing the data centers will fall on the customers itself, but I think the way of operating their applications closer to their data, I think that is going to become a lot more meaningful than just running everything in cloud. If you see 10 years back, there was this huge wave, all the CIOs were like, "Okay, which is your preferred cloud? And let's just move everything there." That's not the case anymore. Now, enterprises are giving it a lot more thought in terms of how their AI stack looks like. And the important thing is they want to have complete control on their software and hardware stack that they deployed on because that is what is going to give them the freedom, freedom from vendor lock-in, freedom to have complete control on their data, and so on and so forth. So I think that is the important part of it. But in terms of day-to-day management of things, a lot of things, those things are outsourced already.
Dave Vellante
>> Are you seeing any particular industry patterns or do you have any customers that you can share within certain industries? Obviously, here in financial services, you're seeing a lot of AI investment there, but seems like it's across the board. Where are you guys having success? Anything you can share?
Garima Kapoor
>> Yeah, it's across the different verticals that we are seeing. Like I said, every enterprise will be an AI enterprise five years from now, and if they're not an AI enterprise, then they don't exist anymore. So it is that meaningful. So across the board, in terms of AI applications from the general readiness, what we see, there is a lot of activity that we see on the manufacturing side, especially as the data gets generated, there's computer vision applications that leverage MinIO in terms to drive the efficiency and the cost at the end of it for defective items and so on, so forth. So that's one use case in manufacturing that we see a lot of repeatability on. Then government is another vertical that we see a lot of repeatability from them deploying MinIO at mission-critical sites to bringing the data to a centralized location, which is again, MinIO, and being able to completely connect both that infrastructure together and be able to run high-performance AI and analytics on top of it. So that's another use case that we see. In finance, of course, the range is very broad because you have fraud detection, analytics that get run on MinIO. Then of course there are analytics and AI stack that gets run on MinIO from generating even the financial reports, and so on. So the use cases are very diverse that we see, but I think AI's nature is that it's going to be much more horizontal in nature. Then of course there are bigger LLM models that use MinIO underneath them just to train their models day in and out. So the use cases are very diverse, but we do see a repeatability from training inferencing computer vision applications and running mission-critical workloads on MinIO.
Dave Vellante
>> When I think about ObjectStore, the GET/PUT paradigm, really simple, obviously S3 popularized it. How does that fit with AI? I mean, obviously from a scale standpoint, it's probably a good thing. You got all this data, a lot of stuff still in files. How do you see the simpatico between object and AI?
Garima Kapoor
>> So AI is built on object. If you see the models, OpenAI, GPT models are built on ObjectStore or Anthropic, they're all built on ObjectStore natively. And the reason is that because ObjectStore gives them the scale, the object score. Also, the kind of data that they work on, it's all unstructured data. The data is immutable in nature. The data is best suited to run on the object storage workloads because legacy file systems, if you see those protocols are very chatty protocols. And rightfully so because of the requirement back in the day was to be able to operate the file, open the file, operated light, hold the locks and the other environment. And then there is lot of things that needs to get managed from a file system standpoint, which object storage lets go of that legacy because it's not needed anymore in this new way of doing training and inferencing world. So if you see most of the LLM models, they are built natively on ObjectStore by default. Cloud is built on ObjectStore. So those are the practices that are now coming back to enterprise as they're looking to modernize their environment, how they should invest in their storage stack and what their cloud-native architecture needs to look like.
Dave Vellante
>> So let's talk about your business a little bit. You've got some big backers, Intel, Masa Son is in there, Dell Tech, Capital, General Catalyst, and a number of others. I think you've raised them right about $120 million, which is, I mean, it's not an enormous amount of money considering you started mid last decade, so you're pretty capital-efficient. So how's it going? How are the investors doing? Are they happy? You're raising more money, you're going public. What's the plan?
Garima Kapoor
>> It's going well. I mean, we couldn't have asked for a better timing and we raised our Series B, I think we closed it in Jan '22, and we still haven't touched a single dollar off it just because the customer traction and the customer money is the best money and validation that any organization can ask for. In terms of fundraising, the next round that we are planning is going to be more strategic in nature to help the company elevate to another level as we are starting to invest ourselves in the AI stack and how that is going to look like. So that is going to shift in terms of how MinIO is going to be consumed in the market overall. So I think that is where we are looking to get the strategic investment, but right now we are not in a fundraising mode or anything.
Dave Vellante
>> So you have a pretty interesting career. I'm checking out your LinkedIn here. You got your PhD, you were doing angel investing, you have friends at Treasure Data. I remember them from the big data world, Sri Ambati's company, H2O, you invested there, and a very active angel investor. Then you came to MinIO as the CFO, and then you're heading operations, COO, and then now you're co-CEO and a co-founder. So that's a unique path. I wonder if you could describe how you got here, and would you advise somebody taking a similar path?
Garima Kapoor
>> I think for me, being founder is the best title that anyone can ask for. And that's the best thing that I did for myself to start my own startup. And in a journey of the startup, as a founder, you get to do multiple things and so on. So other titles are not that exciting other than the founder one because you are always a builder, you're creating value for your shareholders, you're creating value for the overall industry, and so on, and creating a world that did not exist before. So being a founder is extremely gratifying. And I think for anyone I would say, and especially I get asked this quite a bit as with AI being the centerfold of a lot of things that we do, how should anyone think about their career? I always tell them that think about more from a leadership level and how you want things to evolve because the roles are going to merge completely, whether it is leadership and IC roles are going to get merged pretty aggressively. We are already seeing this in the software space. So for anyone, I would just say that be a creator, be a builder. I think there is nothing as gratifying as that. And if you are a doer, you just cannot go wrong. I think that's the way to do things and rest, everything will fall in place.
Dave Vellante
>> I love it. Well, congratulations on building and best of luck in the future. Thanks so much for coming into our remote studio here at the NYSC Wired and theCUBE, and hopefully see you face to face in 2026.
Garima Kapoor
>> Absolutely. Thanks. Thanks, David.
Dave Vellante
>> Thank you. Garima Kapoor at MinIO. Thank you very much for watching. My name is Dave Vellante. John Furrier is here. Keep it right there for more action from our mixture of expert series from the NYSE right here on theCUBE.
Dave Vellante