Multi-Cloud Migration Moves From Months to Minutes
Multi-cloud migration has long been slow, costly and risky. Moving workloads between clouds, or even between accounts, can take the better part of a year. Demand for faster portability keeps growing.
FluidCloud Co-founder and CEO Sharad Kumar examines that challenge with Alan Shimel at KubeCon + CloudNativeCon in Salt Lake City.
A Problem Learned the Hard Way
Kumar’s previous startup, Accurics, was acquired by Tenable. Afterward, moving from one AWS account to another took eight to nine months. Customers were lost during the delay.
He resolved never to face that again. His team took on migration projects to learn where the real friction was.
Mapping Across Cloud APIs
Teams often don’t understand how resources work in the destination cloud. Containerized apps may move easily, but IAM, network policies and firewall rules get left behind.
FluidCloud built a mapping across cloud APIs down to individual attributes and values. Give it a source cloud, and it produces destination Terraform in seconds. Kumar calls it a hypervisor for the hyperscalers.
Customers then asked about data. FluidCloud now handles volumes, file storage, object storage and databases. Kumar keeps the scope narrow on purpose, so the company doesn’t become a catchall.
Hyperscaler Backlogs and Neoclouds
Partners such as AWS, Oracle Cloud Infrastructure, Microsoft Azure and Nutanix put FluidCloud in front of their customers. That is a powerful channel for a young company. Kumar notes the hyperscalers hold large booking backlogs. Faster migration turns those bookings into revenue sooner.
Neoclouds are also growing quickly and launching their own infrastructure services. FluidCloud partners with Vultr. Kumar sees customers seeking GPU and CPU capacity across several providers rather than one.
Sovereignty and Resiliency Converge
Geopolitical risk is pushing organizations to avoid dependence on a single provider. Kumar describes data, operational and digital sovereignty as distinct needs.
He expects sovereign cloud, resiliency and multi-cloud orchestration to merge into one platform. His takeaway from KubeCon is that multi-cloud is unstoppable. That makes fast, reliable multi-cloud migration essential.
Transcript
Hey everyone, welcome back here to the show floor at KubeCon in Salt Lake City 2026. It's a rocking crowd here today. A lot of interesting things.
There's always interesting people at KubeCon. Absolutely. Let me introduce you to one of those interesting people, my friend Sharad Kumar.
Actually, I just met Sharad, but he helped me get tickets to the Jazz game tonight, so I'm forever grateful. Thank you. Sharad, welcome.
Thank you. I appreciate it. Sharad is the CEO of Fluid Cloud.
That's correct. And we're going to hear all about that. But before we get to Fluid Cloud, Sharad, let's hear a little bit about you.
Oh, yeah. No, so this is my second startup. The first one was into cloud security posture management space, and we sold the company to Tenable.
We got acquired. What company was that? It was called Accurx.
Sure, I know Accurx. Yes. I was one of- Who was your co-founder there?
Piyush and Om. Yes. Yes, that's right.
So we worked very closely, and I think we did create a dent in the CSPM space once Shiv left. Yes, you did. And, I think I had a lot of learnings from that, including the cloud scanning and how to write the policies.
And based on that, once we got acquired, it took us eight to nine months to move from one cloud to another cloud and one account to another account within AWS. Mm-hmm. " So I took the problem to my head and then started solving it.
I've founded four or five venture-backed companies. Not this one. This one was no venture.
I bootstrapped it. Yeah. But I've done four or five.
Yeah. What you're telling me resonates- Yeah ... on a couple of levels.
Number one, I think you learn something, whether you have a successful exit or not. Sometimes you learn more- Yeah ... from your failures- That's correct ...
than you do your success. That's correct. But you learn something from everyone.
That's correct. And then I think when you're truly... I've met some serial entrepreneurs, friends of mine, who, they start companies to start companies.
Yeah. You know what I mean? But I think to be really successful, one of the lessons I've learned from 35 years of starting companies is if you're not passionate about what problem you're solving- That's correct ...
about how in whatever small way you're helping people- Yeah ... it's not successful. It's not successful.
Because when the going gets tough in every startup, you know this- That's correct ... there's those periods- That's correct ... where you question yourself.
You doubt yourself. Absolutely. What am I doing this for?
That passion is what sees you through. Thank you. And it's not just you, it's the team.
Thank you. Because when you're a startup, the team feeds off- And that's correct ... that passion.
Of the passion of the founders. Absolutely. It's like the milk of the thing.
That's correct. Absolutely. So I appreciate you.
Hundred percent. And appreciate what you've done there. So you saw this problem, though, and again, this is another common pattern I've heard.
I can't be the only one with this problem. No. Other people have to have this problem, too.
Everyone has the problem. Yeah. Now, so you saw the problem.
How do you fix it? So what we did was, we kind of started going into the weeds, like why this problem exists. " I couldn't get the answer.
Like, can I take a migration project myself and see what problems happens? And so we took a couple of migration projects. I think the biggest problem what we figured out, if it is cloud to cloud, people don't know the destination cloud and what people are looking for is, can somebody explain them in plain vanilla how the cloud resources on the other side work?
How can we map from one resource to another resource? And then if it's a Kubernetes containerized application, the movement is shift. But what happens with, even if you move it, it's an application.
Nobody takes care of the IM, the network policies, the firewall rules, and all those kind of things. Like, okay, that's the problem we are looking at. So when we are looking today, about moving from cloud to cloud, we have built the mapping across cloud APIs, not just resource to resource level.
What we have built the mapping is by resource. Within the resource, we have attributes. Within the attributes, we have values.
We have the mapping to the most detailed level of attributes. So if you give me a cloud on the source side, I can give you the destination Terraform in seconds. Right.
So that's what this was the biggest problem. And then as soon as we were giving this to the customer, the customer's like, "Hey, everything fine. " It's like there are thousands of tool for data.
You can do for it, go for it. And then we went into the data part of it, like, what's the problem? And then we figured out more problems on the data side of the world, and so far and so forth.
So we started building a whole stack. But our core at heart, what we do is mapping across cloud APIs. In a simplest way, I can tell you, we are the hypervisor on a hyperscalers.
A hypervisor for a hyperscaler. That's correct. Now, once you start getting into data, though, it's funny, especially in today's world, data lives everywhere.
That's correct. Right? It's not just in a hypervisor even.
It can live on the edge. It lives on-prem. That's correct.
Everywhere. That's correct. How do you normalize that?
How do you handle that raw data of a problem? It's- That's correct ... it's hard to get your head around that.
Hundred percent. That is correct. So for datas, we take care of four kind of data, which is volume storage, files- Mm-hmm ...
and object storage. So that's is what we are kind of restrict... And databases, sorry.
So we have restricted ourselves to these four. For now. Right.
What I have learned over the period of two companies is if you don't narrow down enough, you become a catchall thing. Right. You can't be everything to everyone.
You cannot be everything to everyone. It's either- There are people who are looking for your solution. Your job as a founder is to figure out who those people are.
I think I find myself very lucky to partner right now with AWS, with OCI, with Azure, as well as with Nutanix, and they're the one now who are putting us in front of their customers. And that's a big channel, man. So, I feel very blessed and to be very fortunate to be in this situation, that they understand the niche which we have created, and they're able to take it forward.
So when it comes to not everyone is looking for the solution, not all the solution can be built. I think they understand what we have built, and then they have exact use cases, exact problems, and then we go in front of the customer, this works. And it's right there.
I learned this lesson myself. One of the startups I did back in... I started in 2001, but we got very involved with the federal government as a customer, and more and more they were steering our product.
Yep. And it was great because we built what we called military grade, because it was DoD- That's correct ... for them.
But what we found is, unfortunately, we wound up building a product that if you weren't the US Department of Defense, you didn't really need that level. Okay. And it was a mistake, right?
And so the lesson I learned there was you got to stay true to your market, not just any one customer. That's absolutely right. Right?
That's almost the opposite of what you said, where you can't be everything to everyone, or you can't be just one thing to one customer. You got to pick your target personas, your target- Absolutely ... product match.
Absolutely. When I'm looking at this problem, who's my market primarily? Of course, it's the end customer.
Sure. But who is facing the biggest problem for this? These are hyperscalers.
When you look at all the hyperscalers, they have billions in dollars of backlogs. They do the booking, but the revenue is lagging. And that's the reason, when you see they have great announcements, great bookings, but when the revenue is not able to support those bookings, the problem is they're not able to move.
So what we are here to challenge is, challenge to the industry, can we... For example, if a hyperscaler has 100 billion worth of backlog, can I take 10 to 15 billion worth of backlog for you and deliver that to you in next three to four months? Absolutely.
Because it's the cash what we are delivering in hand. Yeah, absolutely. And that's the beauty of partnering with the hyperscalers, though, too.
This is something, again, when I was doing my companies, we'd look to partner, and we did. I did a lot of deals with Dell, HP, the usual IBM, the usual suspects. That's right.
But today it's the hypervisors where that kind of action... And I would say not just the hypervisors, but there's this whole neo cloud- That's going to- ... come under them, right?
The next benefit. They're not a hyper provider. Yeah.
But they're nevertheless huge consumers- Absolutely ... or middlemen is a better word, of this. So now, very interestingly, now they're also coming up with their own IAS services.
Yeah. Like when you're looking at cloud compare- Well, because they have to difference themselves from the hypervisor. Absolutely.
When you're looking at Nevius, you're looking at VULCAIN Cloud. We just met VULCAIN Cloud. Yeah.
Right? We are partner to VULCAIN Cloud. So when we see the migration is happening faster, people are trying to build a multi-cloud orchestration layer at this point in time.
No one wants to be locked in. No. And then that's not the use case, only with the lock-in in itself.
Now the multiple use case gets allocated, like sovereign cloud. Because of the geopolitical risk, which- Sure ... every country is facing right now, people don't want to be dependent, and they want the supply chain risk to be minimum at this point in time.
They don't want the rug pulled out from under them, number one. Number two, they want to control their own IP. That's right.
It may not be on a server sitting under my chair, but that's my data, that's my IP. That's correct. And that again is another, I think that's driving huge part of this.
And then there is three sides of sovereignty as well. It's like data sovereignty. You have operational sovereignty, you have digital sovereignty.
Sovereignty. You want to attain digital sovereignty by building the data center. But if you don't have operational sovereignty, there is nothing which is going to move.
There isn't anything, right. Because that is the multi-cloud orchestration layer. I agree.
Agreed. So that's one of the use cases. Resiliency becomes another use case.
And what I'm seeing from the market today is resiliency, sovereign cloud, and multi-cloud orchestration, everything is going to merge, and that's a singular tool, and a singular platform is needed for that. And that's what we are trying to do here. Agreed.
Excellent. Now, let's talk to the people. This camera right here, where could they go find out more information about Fluid Cloud?
com, and I'll be excited to respond back. One last question. So most of the day, first day here at KubeCon's done.
What are you hearing? What do you think is the buzz of the people here? It's a very good question.
The buzz of the people primarily is more towards the AI. I see now companies which were typical SaaS companies, and they're trying to sprinkle AI, and they're trying to get into this AI race, but if their fundamental architecture is not driven from AI, like it's not built for it, it's going to be challenged. So I see that often.
But one thing which I take away from this conference is eventually I think the multi-cloud is unstoppable. I'm looking at all the neo clouds as what we are talking about. They have the biggest booth at this point in time- Yeah ...
and the maximum crowd. So people are looking for like, "I need GPUs-" Alternatives. "...
" But I need to distribute. And I need to distribute. I can't put all my eggs in one basket.
Yeah. That's what I'm taking from the conference here. Love it.
Thank you so much. Thank you so much. We're live.
We got more coming to you on day one, KubeCon here in Salt Lake City. You're watching Techstrong TV. Thank you.