In this interview from the theCUBE + NYSE Wired: Mixture of Experts series, James White, CTO at CalypsoAI, joins theCUBE’s John Furrier to unpack CalypsoAI’s newly launched Security Index – the first comprehensive safety ranking of major generative AI models. White explains how the weekly updated leaderboard and the CASI (CalypsoAI Security Index) score enable apples-to-apples comparisons that blend quality and security, helping enterprises move beyond POC purgatory and toward ROI. The discussion connects model selection and risk posture to enterprise strategy at the intersection of tech and finance – where governance, vendor constraints and performance/latency considerations shape deployment choices at scale.
White details CalypsoAI’s Red-Team product and three attack lenses: signature attacks, operational attacks (e.g., overwhelming outputs that mimic denial-of-service) and “agentic warfare,” which uses autonomous agents to probe for jailbreaks and prompt-injection gaps. He breaks down CASI’s inputs across severity, complexity, decay of older tactics (like DAN variants) and defensive breaking points, alongside an average performance column so teams can weigh capability vs. security. Highlights include Anthropic models leading the safety pack (with Microsoft among the leaders), Claude 3.5 scoring 96.25, Claude 3.7 trending into the #2 slot with different security trade-offs, DeepSeek-R1 landing mid-table and GPT-3.5 Turbo dropping from the top 12. White also previews a human-in-the-loop Purple-Team approach, and shares guidance for continuous testing in CI/CD, model family choices across cloud stacks and real-world implications for POCs, benchmarks and production hardening.
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
Umesh Sachdev, Uniphore
In this interview from the theCUBE + NYSE Wired: Mixture of Experts series, James White, CTO at CalypsoAI, joins theCUBE’s John Furrier to unpack CalypsoAI’s newly launched Security Index – the first comprehensive safety ranking of major generative AI models. White explains how the weekly updated leaderboard and the CASI (CalypsoAI Security Index) score enable apples-to-apples comparisons that blend quality and security, helping enterprises move beyond POC purgatory and toward ROI. The discussion connects model selection and risk posture to enterprise strategy at the intersection of tech and finance – where governance, vendor constraints and performance/latency considerations shape deployment choices at scale.
White details CalypsoAI’s Red-Team product and three attack lenses: signature attacks, operational attacks (e.g., overwhelming outputs that mimic denial-of-service) and “agentic warfare,” which uses autonomous agents to probe for jailbreaks and prompt-injection gaps. He breaks down CASI’s inputs across severity, complexity, decay of older tactics (like DAN variants) and defensive breaking points, alongside an average performance column so teams can weigh capability vs. security. Highlights include Anthropic models leading the safety pack (with Microsoft among the leaders), Claude 3.5 scoring 96.25, Claude 3.7 trending into the #2 slot with different security trade-offs, DeepSeek-R1 landing mid-table and GPT-3.5 Turbo dropping from the top 12. White also previews a human-in-the-loop Purple-Team approach, and shares guidance for continuous testing in CI/CD, model family choices across cloud stacks and real-world implications for POCs, benchmarks and production hardening.
play_circle_outlineUniphore's $260 Million Funding Round and Strategic Partnerships with NVIDIA, AMD, Snowflake, and Databricks: Shaping the Future of AI
replyShare Clip
play_circle_outlineNavigating AI Adoption: Uniphore's Customer-Centric Innovation and Challenges for Fortune 500 Companies in Data Readiness
replyShare Clip
play_circle_outlineExplanation of the four-layer architecture of the Business AI Cloud model.
replyShare Clip
play_circle_outlineRevolutionizing Agentic AI: Strategic Go-to-Market Approaches, Partnership Expansion, and Pathways to Exponential Growth
>> Hello, I'm John Furrier with theCUBE here at our New York Stock Exchange CUBE Studios on the East Coast. Of course, we have our CUBE Studios in Palo Alto, connecting Silicon Valley and Wall Street. We've got a great exciting CUBE alumni in the news today doing the rounds. I saw them on CNBC, now here inside the CUBE at the NYSE. Again, this is part of our Wired programming in the CUBE partnership. Uniphore scores $260 million on a Series F from big names that we know and cover, NVIDIA, Snowflake, Databricks, AMD, Umesh in the house, CEO, co-founder. Great to see you. Thanks for coming in. Congratulations.
Umesh Sachdev
>> Well, always great to be back with you, John. And what a location on, we are looking on top of one of the last standing trading floors. This is great to be here.>> It's a lot of energy. The open outcry, the options. I can see the boards. I'm quite the expert in the option floor here, probably do a column on it, but I'm super excited. When I saw the news come out on your funding, it's a Series F, which means you're in the metrization and the wave and our previous coverage where you guys came from, the trajectory is built on data. And the last conversation we had with you, we were talking data, data layers, how to leverage all that data, but now the Business AI Cloud, the new growth vector for you has been smashing success. Tell us about the funding round. Give us the quick stats on the players involved, the use of funds, what are you going to use it for? Because you have a new growth vector on the Business AI Cloud, which is the business transformation we're seeing.
Umesh Sachdev
>> Well, clearly we are very excited today at Uniphore, but I think it's not only just a big day for us at Uniphore. It's a huge day for the entire business AI industry, bringing together the ecosystem, some of the biggest names in AI and data. What it means is a validation of two things, a validation of our strategy. Couple of years ago we put in place this strategy of a sovereign and open data and AI platform because we saw the largest businesses in the world, Fortune 500, Fortune 1000, as they adopt AI, data sovereignty and AI sovereignty was going to be huge for them. And we positioned Uniphore to lead that space, and the fact that the adoption of business AI itself is coming of age. We are seeing agent AI not getting a wide enough adoption yet, but going very deep. We are seeing some of the big four consulting firms adopted, some of the biggest banks adopt agentic AI. And so this funding round and with the players like NVIDIA and AMD and Snowflake and Databricks choosing to invest in our funding round along with other financial investors, I think it's a huge validation of the entire strategy that we put in place two years ago.>> I was on the phone this morning with NVIDIA on preparation for GTC in DC, which is next week. It's almost as big as the main event. The demand is so high. I want to talk about NVIDIA and some of the names you mentioned because we cover them, but NVIDIA specifically, because I think what you're doing with NVIDIA is a great parallel to what I call a new next generation ecosystem around AI, the tsunami of AI native applications that's emerging, and two, companies that have been in the data game, kind of in the AI arena. You guys can see the opportunities. It's not like you pivoted. You were already kind of AI-ready with the data play. You see the growth vector, you jump on it, you can deliver it. So one, this business AI kind of ecosystem, it's different. You have to have with agents trusted relationships. It's not just a Barney deal, like hey, let's do some go-to-market together. There's integration involved. Can you share your thoughts on reaction to that and am I getting it right? And if so, or what is the market dynamics around that? Because everyone that's successful has "an ecosystem play", but that's kind of like a cliche, but I see something bigger. What is that all about? Can you share your vision on that ecosystem in this era of AI?
Umesh Sachdev
>> Well, so this takes me back to two years ago. At Uniphore, we've always practiced a customer-centric innovation philosophy. And when gen AI and LLM was about to take off, we realized that our customers had four big impediments in their adoption. They were all doing POCs and pilots, but there were four things I was repeatedly hearing from my CEO or CIO clients that we were meeting. The first was data readiness. Every large Fortune 500 that we met or still continue to meet don't feel confident that their data is ready. They know they have enough data. It was just not prepared or kept in a way that would suit the AI avalanche that we are seeing now. The second was sovereignty. Every boardroom were starting to realize that the biggest intellectual property in the era of AI is going to be data. And if you're a bank, if you're an insurance company, if you're a telecom service provider, what is your biggest moat in this AI world? It's your data. And so sovereignty of that, not having to send it to a cloud and hope and pray that nobody misuses it was big. The third was the pace change. Every week, LLMs are changing, new GPUs are coming out, new platforms are coming out, and business leaders are finding it hard to make decisions. And finally, adoption. So these were four issues that I kept hearing, and we took a strategy of creating a platform that up until then, our own engineers were using the same platform. But when we saw this was going to be a problem for our clients, we opened up our platform, which we call the ->> It's kind of like a Slack strategy.
Umesh Sachdev
>> Exactly right.>> Built for the engineers and then it becomes the hottest product.
Umesh Sachdev
>> Yeah, Slack, AWS, GCP, all great products in the enterprise arena are built exactly like this, solving a real problem first for yourself and turns out it's a problem for everybody. So the Business AI Cloud was born with this philosophy. As it was born, we now have seen an inflection in our business in the last two years. In the last financial year, Uniphore grew about a hundred percent. We are on target to repeat that kind of growth, and we are not a tiny company. We're a Series F. Our revenues are in the hundreds of millions. And so seeing this kind of inflection point, over 2,000 large businesses using our Business AI Cloud and our platforms, that's a real validation that those problems that we saw and the solutions we put in place are working. Now let's talk about the partnerships that we are striking. Uniphore is a very large user of the software stack of NVIDIA, the NVIDIA AI for enterprise. That entire stack of software has a big role to play in our platform architecture. So are components from AMD, so are our integrations with Snowflake and Databricks. Uniphore is at the tip of the spear, driving enterprise AI adoption. And each time Uniphore drives adoption at agentic AI or fine-tuning of models or the data layer, we drive consumption for GPUs. We drive consumption for data platform providers, and that's why this ecosystem is coming together.>> I really appreciate that answer and I'm glad you brought up NVIDIA, and I'll bring this back to the ecosystem. If you look at what those guys are doing, they're nailing all the infrastructure side. So they do a good job on that. Now when you look at the business model transformation, it's almost a transformation of the transformation, meaning go back a decade, we talked about digital transformation, IT. Now this business model transformation, so if you look at NVIDIA, you look at distributed computing, edge is right around the corner. You can do inference at the edge right now, training's going to be coming out soon with the Thor chip that they have and other smaller devices. The sovereignty is distributed, so you have countries, and that's, or geographies. So you got a distributed kind of horizontal play, which I heard you mention on the John Ford interview, but the cloud is vertically integrated. So you start to see the flip script. So in the SaaS era, the cloud was a horizontally scalable resource for SaaS apps, and I heard you talk about SaaS transitioning to agents. So in the agentic, it seems to me, and I want to get your reaction to this, that the network is distributed. That's the horizontal scale, and you could go vertical to the cloud for the app, which means the business model logic becomes the critical asset if you're taking advantage of the architecture in this new AI. Do you agree with that? And what's your thoughts on this new agentic architecture? Because this doesn't work to me any other way.
Umesh Sachdev
>> Let me make this real, the point you're making. Let me make this real for our viewers with an example. We work with a telecom service provider out of Europe. Their business in their home country runs out of their Google Cloud partnership. Then they have their business in other European countries runs on a version of the Microsoft cloud. And finally, this telecom service writer has a business in Africa where they're yet to find a data center which has 24/7 power supply. So they're going on-prem in Africa. The chief AI officer of this telecom company has a challenge now to say, "Between my home country and Google, some other countries and Microsoft, and the African nations on-prem, how do I build a common AI platform delivering consistent employee and customer experience when the underlying infrastructure is so different in different countries?" This is the challenge that business AI has to grapple with. This is why when people say, "Let's just slap a ChatGPT into enterprise and it'll work," I go, "You have no idea." Consumer AI and business AI cannot be more different from one another.>> And I think, that's multi-cloud in principle, but I think what I heard you mention earlier on the CNBC interview was the data's critical that it's horizontally available. So in order to get the data for AI, you have to take the disparate vendors, in this case, cloud vendors and on-prem, and integrate it before you can do anything.
Umesh Sachdev
>> That's exactly right.>> And you sit on top with the Business AI Cloud, right? Is that how it works?
Umesh Sachdev
>> We are the Business AI Cloud. And our Business AI Cloud connects to hundreds of fragmented data sources. We have a customer who has a Databricks, a Snowflake, a Palantir, a Teradata, three catalogs, three ETLs, and their chief data officer said, "Before we do agentic AI, give me three years." And Business AI Cloud, we showed up and said, "You don't have to wait three years.">> Give us a month.
Umesh Sachdev
>> We can get you ready by connecting to wherever your data sits.>> How long does that take, time-wise?
Umesh Sachdev
>> Well, today we are at a time to value it in six to eight weeks, we can get a customer up and running in a real use case. And that's why we are beginning to see this kind of a hockey stick type adoption that the customer goes, "Too good to be true, let me try out in one area," six to eight weeks, live in production, and then the flood gates open.>> Define Business AI Cloud, because I really think that you hit this right on the money because the conversations we're having on theCUBE here and in Palo Alto and at the events is the same theme. The CFOs are involved. You're seeing different personas involved in decision-making. In the old IT world, you're smiling because I know you probably agree, the old IT world is the CISO, the CIO, let's buy that rack, the new servers and the switches. Now the decision-making is coming from CFOs because they have to instrument the business model. So we've been seeing a business model transformation. So talk about that trend, how you see it, and two, what is, for the folks watching, what is a Business AI Cloud?
Umesh Sachdev
>> So let's start with the Business AI Cloud. It's an end-to-end data and AI architecture for the entire enterprise. What does that mean? Think of it as a four-layer cake. The bottom-most layer is the data layer. Surprise, surprise, the layer that can now connect to hundreds of fragmented disparate data sources, bring it all together and prepare the data to serve up to the AI layer. That's the data layer. We then move to the knowledge layer. The knowledge layer is nothing but a factory, John, a factory for fine-tuned domain-specific small language models. People don't realize, LLMs are great for consumer use cases like internet search. For enterprises, they want models to be small and focused, like a telecom company wants a billing model. An insurance company wants a churn model or a claims model. So our knowledge layer is a factory to very rapidly create these domains.>> So you're saying vertical models are real? Small-
Umesh Sachdev
>> Vertical models are real. Then we put them on the model layer, all these models, it's like a parking slot, and we give them an iron dome, which is an AI security layer, and protect all the models. And finally, we give it the agentic layer where now we say, "You can build these AI agents end-to-end, take a process where you had four people who had to sign off and then the manager had to approve. Now there's four AI agents that can sign off and then a human agent, human in the loop can approve." But this agentic process will be governed by vertical models, those domains SLMs. Those SLMs in turn would be learning from all the disparate data coming together, and this architecture coming in the enterprise is the Business AI Cloud. Now, the beauty of this architecture as we've done it is that it's sovereign, which means what? It can run on cloud but also on premise. So the enterprise doesn't have to have anxiety of where they're sending data.>> And that, by the way, kicks in for the folks watching, just to call that out, massive in-country compliance, a lot of the stuff that's part of the AI agents, which is-
Umesh Sachdev
>> Absolutely.... >> first-party designed in, compliance has to be designed in. That seems to be the key thing for agents. Is that, do you agree?
Umesh Sachdev
>> That's not only a key thing for agents. The other key thing for agents is the openness. You wouldn't find in today's world a company which goes, "All my AI use cases will be on only one LLM, only one GPU, only one type of data lake," because everything is becoming fit for purpose. We have some open-wage models coming out of different countries. We have some very potent American proprietary models, but for different use cases, think about it this way. If you give me a use case to automate or identify customer service calls, I would say that's a low latency use case. You want a less than one second response time from the LLM.
I want to run the highest capacity GPU and the most efficient LLM. Now you say, "By the way, my next use case, Umesh, is in the finance department. For my CFO, source to cash, let's say gentrify that." I say, "How important is latency here? Can it deliver the result tomorrow to you?" And you go, "That's better than a month," but that's what the human process is. So now because it's high latency, I can go to a lower cost GPU, which is of a different provider, a different type of LLM, which means the optionality and openness of the architecture->> So you can flex on costs and also performance because-
Umesh Sachdev
>> Exactly right.... >> with the latency as a policy, you can say, "I want better reasoning," for instance. Or-
Umesh Sachdev
>> You want better, so it's latency, cost accuracy.>> Accuracy. That's the word.
Umesh Sachdev
>> There are some mission-critical use cases, and there are others where 80% is good enough. And so because enterprise is not a monolith, it is fragmented. You need an architecture that is not only sovereign, but open, and that's when AI agents can really proliferate into the enterprise.>> Yeah. Umesh, I love the story. I think you nailed the positioning and the timing of the Business AI Cloud, and I can see why Snowflake and NVIDIA and Databricks and all the folks that put money in want to work with you, because you help them, and you help them sell more and do more. So take me through the use case of I'm a, you sold me on the Business AI Cloud. Say I'm a customer. Okay, I'm sold. I love it. I want my business model to be transformed in all use cases, every value chain, every value activity I want optimized. I want to reimagine my business. You sold me. What do I do? Am I buying a cloud? Am I buying a framework? How do I deploy this Business AI Cloud?
Umesh Sachdev
>> That's a great segue. So let's use your question to also answer the business transformation issue that you're heading upon. Take for example, you are a global system integrator or a big four consulting firm. Up until yesterday, you used to go to your clients and say, "You pay me a premium because I have a handful of people in my company who are experts, experts in telecom. So pay me a premium for my time and money. Experts in insurance, pay me a premium for the time my people will spend with you." Well, the problem is in agentic AI, the stuff that professional services teams would do, getting automated. So my proposition to our customers who are either consulting firms or system integrators is that you have a unique opportunity now to codify your company's IP, which used to be in the minds of human beings, into these verticalized small models. So a big four firm takes oil and gas as an industry, creates a small model, takes telecom as an industry, creates a small model, takes banking as an industry, creates a small model. Against each small model, you have to build agents. Now, this big four firm goes to its clients and says, "I'm still bringing you my expertise, but instead of staffing your project with 1,400 people, I can do with four people and 800 AI agents. Instead of paying me time and material for the hours my people used to spend, how about we do gain share? If I deliver the savings to you, if I deliver the growth to you, I'll share some of that outcome." So that's an example of a whole business model.>> So new pricing mechanisms are emerging?
Umesh Sachdev
>> A hundred percent.>> Okay. I was looking up on LinkedIn while you were talking, because last week at Salesforce Dreamforce, I interviewed Todd Lohr. He's the national managing principal partner at KPMG. He's on the management committee. He told me basically the same thing, but he said it a little bit differently. He said, "With agents, we're going to change how we do our delivery model." He says, "So we want to reimagine, bring all the resource of KPMG," and McKinsey already started doing this. So when they go on a deployment, they're bringing all the intellectual capital to the table on all engagements without the labor. So they're using-
Umesh Sachdev
>> That's exactly right.... >> the AI to the front lines, which you just said. So the delivery mechanism and the efficiency is off the charts. So that's the dream scenario. So there's already things happening in there. How does an average customer do that? What's the playbook? Take me through, okay, sold me on that execution. I see that as reality. That's a preferred future. No doubt about it in my mind. What do I do? Do I have to get my data in line first? Is there a sequence of events? What are the requirements to deploy this?
Umesh Sachdev
>> Well, the first thing is in the absence of something like the Business AI Cloud, you would first go to the GPU provider, buy up some GPUs, then you'll go to a data provider, buy up a data cloud, then you'll go to a middleware provider. The Business AI Cloud shows up and says, "Listen, this is how you get started. I'm bringing all those AI infrastructure as a service to you. The number one thing you should focus on is do you have clean, ready data?" And nine out of 10 times, the answer is no, we don't. Okay, so then let's get->> Get a data lake, unify.
Umesh Sachdev
>> Get experts in your company, even before you do the data lake, because that could take months. You don't want to wait months for your agentic AI. Synthetic data, synthetic data. Okay? If there are subject matter experts in your company, let's start to simulate what good would look like, at least to get started. Because these models are good at learning from something small. So synthetic data usually turns out to be the first step to start training these verticalized domain-specific models, and then build AI agents to start to replicate what was a manual process with an agentic process. Now, put this production, and as you put it to production, it's either about growth or efficiency.>> Yeah, and context becomes a important point. Umesh, this is a great story. Again, love to do a follow-up. We can talk an hour for it. Love agents. I'm just intrigued by the positioning of the Business AI Cloud because it makes so much sense. We used to call it Super Cloud when we were talking about multi-cloud. Once you get things connected, then you got your full hybrid distributed computing environment, number one, and then two, figure out where to leverage that domain expertise or domain data or workflows. So workflows and data moats seem to be the advantage. With the funding, what's your go-to-market plan? What's the use of funds going to be, partnerships, ecosystem, ramp up? What's the plan? What are you optimizing for?
Umesh Sachdev
>> Yeah, so look, the biggest utilization of this funding is continuing to invest in R&D. That's the number one area for us. We are innovating at a furious pace on behalf of our customers. But I get back to this is a topic that you and I will be talking hopefully for years to come. This is evolving at a fast pace. What's really important today, I bring it back, our Series F, It's a $260 million-dollar raise, but more importantly, it's bringing together the whole industry. And again, what am I most excited about is that we are at the cusp of that hockey stick curve of agentic AI adoption. And coming together of NVIDIA, AMD, Snowflake and Databricks to drive Uniphore's growth validates our strategy, our position in the market, our early revenue indicators are very strong, and this seems like for years to come, we can enjoy this inflection.>> Yeah, you're a platform. You're a cloud platform. Umesh, thank you so much for coming in. Congratulations.
Umesh Sachdev
>> Always a pleasure, John.>> Big funding news here on theCUBE, breaking it down. Agents are really the beginning of the business model transformation, where the value creation and extraction is going to be really on the business side. CFOs and the whole company's involved. Of course, this is technology under the covers, and we're doing our best to bring you that value here on theCUBE. I'm John Furrier, your host. Thanks for watching.
>> Hello, I'm John Furrier with theCUBE here at our New York Stock Exchange CUBE Studios on the East Coast. Of course, we have our CUBE Studios in Palo Alto, connecting Silicon Valley and Wall Street. We've got a great exciting CUBE alumni in the news today doing the rounds. I saw them on CNBC, now here inside the CUBE at the NYSE. Again, this is part of our Wired programming in the CUBE partnership. Uniphore scores $260 million on a Series F from big names that we know and cover, NVIDIA, Snowflake, Databricks, AMD, Umesh in the house, CEO, co-founder. Great to see you. Thanks for coming in. Congratulations.
Umesh Sachdev
>> Well, always great to be back with you, John. And what a location on, we are looking on top of one of the last standing trading floors. This is great to be here.>> It's a lot of energy. The open outcry, the options. I can see the boards. I'm quite the expert in the option floor here, probably do a column on it, but I'm super excited. When I saw the news come out on your funding, it's a Series F, which means you're in the metrization and the wave and our previous coverage where you guys came from, the trajectory is built on data. And the last conversation we had with you, we were talking data, data layers, how to leverage all that data, but now the Business AI Cloud, the new growth vector for you has been smashing success. Tell us about the funding round. Give us the quick stats on the players involved, the use of funds, what are you going to use it for? Because you have a new growth vector on the Business AI Cloud, which is the business transformation we're seeing.
Umesh Sachdev
>> Well, clearly we are very excited today at Uniphore, but I think it's not only just a big day for us at Uniphore. It's a huge day for the entire business AI industry, bringing together the ecosystem, some of the biggest names in AI and data. What it means is a validation of two things, a validation of our strategy. Couple of years ago we put in place this strategy of a sovereign and open data and AI platform because we saw the largest businesses in the world, Fortune 500, Fortune 1000, as they adopt AI, data sovereignty and AI sovereignty was going to be huge for them. And we positioned Uniphore to lead that space, and the fact that the adoption of business AI itself is coming of age. We are seeing agent AI not getting a wide enough adoption yet, but going very deep. We are seeing some of the big four consulting firms adopted, some of the biggest banks adopt agentic AI. And so this funding round and with the players like NVIDIA and AMD and Snowflake and Databricks choosing to invest in our funding round along with other financial investors, I think it's a huge validation of the entire strategy that we put in place two years ago.>> I was on the phone this morning with NVIDIA on preparation for GTC in DC, which is next week. It's almost as big as the main event. The demand is so high. I want to talk about NVIDIA and some of the names you mentioned because we cover them, but NVIDIA specifically, because I think what you're doing with NVIDIA is a great parallel to what I call a new next generation ecosystem around AI, the tsunami of AI native applications that's emerging, and two, companies that have been in the data game, kind of in the AI arena. You guys can see the opportunities. It's not like you pivoted. You were already kind of AI-ready with the data play. You see the growth vector, you jump on it, you can deliver it. So one, this business AI kind of ecosystem, it's different. You have to have with agents trusted relationships. It's not just a Barney deal, like hey, let's do some go-to-market together. There's integration involved. Can you share your thoughts on reaction to that and am I getting it right? And if so, or what is the market dynamics around that? Because everyone that's successful has "an ecosystem play", but that's kind of like a cliche, but I see something bigger. What is that all about? Can you share your vision on that ecosystem in this era of AI?
Umesh Sachdev
>> Well, so this takes me back to two years ago. At Uniphore, we've always practiced a customer-centric innovation philosophy. And when gen AI and LLM was about to take off, we realized that our customers had four big impediments in their adoption. They were all doing POCs and pilots, but there were four things I was repeatedly hearing from my CEO or CIO clients that we were meeting. The first was data readiness. Every large Fortune 500 that we met or still continue to meet don't feel confident that their data is ready. They know they have enough data. It was just not prepared or kept in a way that would suit the AI avalanche that we are seeing now. The second was sovereignty. Every boardroom were starting to realize that the biggest intellectual property in the era of AI is going to be data. And if you're a bank, if you're an insurance company, if you're a telecom service provider, what is your biggest moat in this AI world? It's your data. And so sovereignty of that, not having to send it to a cloud and hope and pray that nobody misuses it was big. The third was the pace change. Every week, LLMs are changing, new GPUs are coming out, new platforms are coming out, and business leaders are finding it hard to make decisions. And finally, adoption. So these were four issues that I kept hearing, and we took a strategy of creating a platform that up until then, our own engineers were using the same platform. But when we saw this was going to be a problem for our clients, we opened up our platform, which we call the ->> It's kind of like a Slack strategy.
Umesh Sachdev
>> Exactly right.>> Built for the engineers and then it becomes the hottest product.
Umesh Sachdev
>> Yeah, Slack, AWS, GCP, all great products in the enterprise arena are built exactly like this, solving a real problem first for yourself and turns out it's a problem for everybody. So the Business AI Cloud was born with this philosophy. As it was born, we now have seen an inflection in our business in the last two years. In the last financial year, Uniphore grew about a hundred percent. We are on target to repeat that kind of growth, and we are not a tiny company. We're a Series F. Our revenues are in the hundreds of millions. And so seeing this kind of inflection point, over 2,000 large businesses using our Business AI Cloud and our platforms, that's a real validation that those problems that we saw and the solutions we put in place are working. Now let's talk about the partnerships that we are striking. Uniphore is a very large user of the software stack of NVIDIA, the NVIDIA AI for enterprise. That entire stack of software has a big role to play in our platform architecture. So are components from AMD, so are our integrations with Snowflake and Databricks. Uniphore is at the tip of the spear, driving enterprise AI adoption. And each time Uniphore drives adoption at agentic AI or fine-tuning of models or the data layer, we drive consumption for GPUs. We drive consumption for data platform providers, and that's why this ecosystem is coming together.>> I really appreciate that answer and I'm glad you brought up NVIDIA, and I'll bring this back to the ecosystem. If you look at what those guys are doing, they're nailing all the infrastructure side. So they do a good job on that. Now when you look at the business model transformation, it's almost a transformation of the transformation, meaning go back a decade, we talked about digital transformation, IT. Now this business model transformation, so if you look at NVIDIA, you look at distributed computing, edge is right around the corner. You can do inference at the edge right now, training's going to be coming out soon with the Thor chip that they have and other smaller devices. The sovereignty is distributed, so you have countries, and that's, or geographies. So you got a distributed kind of horizontal play, which I heard you mention on the John Ford interview, but the cloud is vertically integrated. So you start to see the flip script. So in the SaaS era, the cloud was a horizontally scalable resource for SaaS apps, and I heard you talk about SaaS transitioning to agents. So in the agentic, it seems to me, and I want to get your reaction to this, that the network is distributed. That's the horizontal scale, and you could go vertical to the cloud for the app, which means the business model logic becomes the critical asset if you're taking advantage of the architecture in this new AI. Do you agree with that? And what's your thoughts on this new agentic architecture? Because this doesn't work to me any other way.
Umesh Sachdev
>> Let me make this real, the point you're making. Let me make this real for our viewers with an example. We work with a telecom service provider out of Europe. Their business in their home country runs out of their Google Cloud partnership. Then they have their business in other European countries runs on a version of the Microsoft cloud. And finally, this telecom service writer has a business in Africa where they're yet to find a data center which has 24/7 power supply. So they're going on-prem in Africa. The chief AI officer of this telecom company has a challenge now to say, "Between my home country and Google, some other countries and Microsoft, and the African nations on-prem, how do I build a common AI platform delivering consistent employee and customer experience when the underlying infrastructure is so different in different countries?" This is the challenge that business AI has to grapple with. This is why when people say, "Let's just slap a ChatGPT into enterprise and it'll work," I go, "You have no idea." Consumer AI and business AI cannot be more different from one another.>> And I think, that's multi-cloud in principle, but I think what I heard you mention earlier on the CNBC interview was the data's critical that it's horizontally available. So in order to get the data for AI, you have to take the disparate vendors, in this case, cloud vendors and on-prem, and integrate it before you can do anything.
Umesh Sachdev
>> That's exactly right.>> And you sit on top with the Business AI Cloud, right? Is that how it works?
Umesh Sachdev
>> We are the Business AI Cloud. And our Business AI Cloud connects to hundreds of fragmented data sources. We have a customer who has a Databricks, a Snowflake, a Palantir, a Teradata, three catalogs, three ETLs, and their chief data officer said, "Before we do agentic AI, give me three years." And Business AI Cloud, we showed up and said, "You don't have to wait three years.">> Give us a month.
Umesh Sachdev
>> We can get you ready by connecting to wherever your data sits.>> How long does that take, time-wise?
Umesh Sachdev
>> Well, today we are at a time to value it in six to eight weeks, we can get a customer up and running in a real use case. And that's why we are beginning to see this kind of a hockey stick type adoption that the customer goes, "Too good to be true, let me try out in one area," six to eight weeks, live in production, and then the flood gates open.>> Define Business AI Cloud, because I really think that you hit this right on the money because the conversations we're having on theCUBE here and in Palo Alto and at the events is the same theme. The CFOs are involved. You're seeing different personas involved in decision-making. In the old IT world, you're smiling because I know you probably agree, the old IT world is the CISO, the CIO, let's buy that rack, the new servers and the switches. Now the decision-making is coming from CFOs because they have to instrument the business model. So we've been seeing a business model transformation. So talk about that trend, how you see it, and two, what is, for the folks watching, what is a Business AI Cloud?
Umesh Sachdev
>> So let's start with the Business AI Cloud. It's an end-to-end data and AI architecture for the entire enterprise. What does that mean? Think of it as a four-layer cake. The bottom-most layer is the data layer. Surprise, surprise, the layer that can now connect to hundreds of fragmented disparate data sources, bring it all together and prepare the data to serve up to the AI layer. That's the data layer. We then move to the knowledge layer. The knowledge layer is nothing but a factory, John, a factory for fine-tuned domain-specific small language models. People don't realize, LLMs are great for consumer use cases like internet search. For enterprises, they want models to be small and focused, like a telecom company wants a billing model. An insurance company wants a churn model or a claims model. So our knowledge layer is a factory to very rapidly create these domains.>> So you're saying vertical models are real? Small-
Umesh Sachdev
>> Vertical models are real. Then we put them on the model layer, all these models, it's like a parking slot, and we give them an iron dome, which is an AI security layer, and protect all the models. And finally, we give it the agentic layer where now we say, "You can build these AI agents end-to-end, take a process where you had four people who had to sign off and then the manager had to approve. Now there's four AI agents that can sign off and then a human agent, human in the loop can approve." But this agentic process will be governed by vertical models, those domains SLMs. Those SLMs in turn would be learning from all the disparate data coming together, and this architecture coming in the enterprise is the Business AI Cloud. Now, the beauty of this architecture as we've done it is that it's sovereign, which means what? It can run on cloud but also on premise. So the enterprise doesn't have to have anxiety of where they're sending data.>> And that, by the way, kicks in for the folks watching, just to call that out, massive in-country compliance, a lot of the stuff that's part of the AI agents, which is-
Umesh Sachdev
>> Absolutely.... >> first-party designed in, compliance has to be designed in. That seems to be the key thing for agents. Is that, do you agree?
Umesh Sachdev
>> That's not only a key thing for agents. The other key thing for agents is the openness. You wouldn't find in today's world a company which goes, "All my AI use cases will be on only one LLM, only one GPU, only one type of data lake," because everything is becoming fit for purpose. We have some open-wage models coming out of different countries. We have some very potent American proprietary models, but for different use cases, think about it this way. If you give me a use case to automate or identify customer service calls, I would say that's a low latency use case. You want a less than one second response time from the LLM.
I want to run the highest capacity GPU and the most efficient LLM. Now you say, "By the way, my next use case, Umesh, is in the finance department. For my CFO, source to cash, let's say gentrify that." I say, "How important is latency here? Can it deliver the result tomorrow to you?" And you go, "That's better than a month," but that's what the human process is. So now because it's high latency, I can go to a lower cost GPU, which is of a different provider, a different type of LLM, which means the optionality and openness of the architecture->> So you can flex on costs and also performance because-
Umesh Sachdev
>> Exactly right.... >> with the latency as a policy, you can say, "I want better reasoning," for instance. Or-
Umesh Sachdev
>> You want better, so it's latency, cost accuracy.>> Accuracy. That's the word.
Umesh Sachdev
>> There are some mission-critical use cases, and there are others where 80% is good enough. And so because enterprise is not a monolith, it is fragmented. You need an architecture that is not only sovereign, but open, and that's when AI agents can really proliferate into the enterprise.>> Yeah. Umesh, I love the story. I think you nailed the positioning and the timing of the Business AI Cloud, and I can see why Snowflake and NVIDIA and Databricks and all the folks that put money in want to work with you, because you help them, and you help them sell more and do more. So take me through the use case of I'm a, you sold me on the Business AI Cloud. Say I'm a customer. Okay, I'm sold. I love it. I want my business model to be transformed in all use cases, every value chain, every value activity I want optimized. I want to reimagine my business. You sold me. What do I do? Am I buying a cloud? Am I buying a framework? How do I deploy this Business AI Cloud?
Umesh Sachdev
>> That's a great segue. So let's use your question to also answer the business transformation issue that you're heading upon. Take for example, you are a global system integrator or a big four consulting firm. Up until yesterday, you used to go to your clients and say, "You pay me a premium because I have a handful of people in my company who are experts, experts in telecom. So pay me a premium for my time and money. Experts in insurance, pay me a premium for the time my people will spend with you." Well, the problem is in agentic AI, the stuff that professional services teams would do, getting automated. So my proposition to our customers who are either consulting firms or system integrators is that you have a unique opportunity now to codify your company's IP, which used to be in the minds of human beings, into these verticalized small models. So a big four firm takes oil and gas as an industry, creates a small model, takes telecom as an industry, creates a small model, takes banking as an industry, creates a small model. Against each small model, you have to build agents. Now, this big four firm goes to its clients and says, "I'm still bringing you my expertise, but instead of staffing your project with 1,400 people, I can do with four people and 800 AI agents. Instead of paying me time and material for the hours my people used to spend, how about we do gain share? If I deliver the savings to you, if I deliver the growth to you, I'll share some of that outcome." So that's an example of a whole business model.>> So new pricing mechanisms are emerging?
Umesh Sachdev
>> A hundred percent.>> Okay. I was looking up on LinkedIn while you were talking, because last week at Salesforce Dreamforce, I interviewed Todd Lohr. He's the national managing principal partner at KPMG. He's on the management committee. He told me basically the same thing, but he said it a little bit differently. He said, "With agents, we're going to change how we do our delivery model." He says, "So we want to reimagine, bring all the resource of KPMG," and McKinsey already started doing this. So when they go on a deployment, they're bringing all the intellectual capital to the table on all engagements without the labor. So they're using-
Umesh Sachdev
>> That's exactly right.... >> the AI to the front lines, which you just said. So the delivery mechanism and the efficiency is off the charts. So that's the dream scenario. So there's already things happening in there. How does an average customer do that? What's the playbook? Take me through, okay, sold me on that execution. I see that as reality. That's a preferred future. No doubt about it in my mind. What do I do? Do I have to get my data in line first? Is there a sequence of events? What are the requirements to deploy this?
Umesh Sachdev
>> Well, the first thing is in the absence of something like the Business AI Cloud, you would first go to the GPU provider, buy up some GPUs, then you'll go to a data provider, buy up a data cloud, then you'll go to a middleware provider. The Business AI Cloud shows up and says, "Listen, this is how you get started. I'm bringing all those AI infrastructure as a service to you. The number one thing you should focus on is do you have clean, ready data?" And nine out of 10 times, the answer is no, we don't. Okay, so then let's get->> Get a data lake, unify.
Umesh Sachdev
>> Get experts in your company, even before you do the data lake, because that could take months. You don't want to wait months for your agentic AI. Synthetic data, synthetic data. Okay? If there are subject matter experts in your company, let's start to simulate what good would look like, at least to get started. Because these models are good at learning from something small. So synthetic data usually turns out to be the first step to start training these verticalized domain-specific models, and then build AI agents to start to replicate what was a manual process with an agentic process. Now, put this production, and as you put it to production, it's either about growth or efficiency.>> Yeah, and context becomes a important point. Umesh, this is a great story. Again, love to do a follow-up. We can talk an hour for it. Love agents. I'm just intrigued by the positioning of the Business AI Cloud because it makes so much sense. We used to call it Super Cloud when we were talking about multi-cloud. Once you get things connected, then you got your full hybrid distributed computing environment, number one, and then two, figure out where to leverage that domain expertise or domain data or workflows. So workflows and data moats seem to be the advantage. With the funding, what's your go-to-market plan? What's the use of funds going to be, partnerships, ecosystem, ramp up? What's the plan? What are you optimizing for?
Umesh Sachdev
>> Yeah, so look, the biggest utilization of this funding is continuing to invest in R&D. That's the number one area for us. We are innovating at a furious pace on behalf of our customers. But I get back to this is a topic that you and I will be talking hopefully for years to come. This is evolving at a fast pace. What's really important today, I bring it back, our Series F, It's a $260 million-dollar raise, but more importantly, it's bringing together the whole industry. And again, what am I most excited about is that we are at the cusp of that hockey stick curve of agentic AI adoption. And coming together of NVIDIA, AMD, Snowflake and Databricks to drive Uniphore's growth validates our strategy, our position in the market, our early revenue indicators are very strong, and this seems like for years to come, we can enjoy this inflection.>> Yeah, you're a platform. You're a cloud platform. Umesh, thank you so much for coming in. Congratulations.
Umesh Sachdev
>> Always a pleasure, John.>> Big funding news here on theCUBE, breaking it down. Agents are really the beginning of the business model transformation, where the value creation and extraction is going to be really on the business side. CFOs and the whole company's involved. Of course, this is technology under the covers, and we're doing our best to bring you that value here on theCUBE. I'm John Furrier, your host. Thanks for watching.
