Startup Street | Latest Developments From The Startup Space | Business News | CNBC TV18

Added: 2026-06-11

[00:00:16]Hello and welcome to Startup Street. I'm Shuti Mishra and with me is Arunati Raman. These are top headlines we're tracking today.

[00:00:24]My gate raises 225 cr rupes in a funding round led by Dharna Capital. The housing society management platform is now valued at 2,00 to 2,200 cr rupes following the mix of primary and secondary investment.

[00:00:39]Bazar now bagged 72 crore rupes in a funding round led by peak 15 partners and others. The fresh capital will be used to expand into new cities, strengthen supply chains and enhance its tech platform.

[00:00:51]Data Breaks is in talks to raise fresh capital in a funding round that could value the company at nearly $175 billion, says reports. The data analytics and AI software firm is looking to kick off the fund raise as early as next month.

[00:01:06]Pediatric healthcare startup Hoola Health gets $5 million in a funding round led by Peak 15 surge program. The firm currently operates five clinics and plans to open 30 more over the next two years.

[00:01:17]Salesforce lays off people across its agent force AI Mulesoft and Marketing Cloud teams. Reports say the affected US employees could receive up to 30 weeks of severance pay.

[00:01:31]Well, those were the headlines we've been tracking for you today on what's brewing. As governments and enterprises race to build AI capabilities, a new battleground has emerged around sovereign intelligence, data security, and indigenous technology. My guest is building AI systems for some of India's most sensitive national security and intelligence operations. NFU Labs, which develops AI powered platforms for defense, law enforcement, and intelligence agencies, has raised $30 million in series B funding from Panther Growth Partners to accelerate its sovereign AI ambitions and global expansion. Joining me now is Abishek Sharma, the co-founder and CTO of Infu Labs. Abhishek, welcome to Startup Street. Now, Infu's technologies are deployed across defense, intelligence, law enforcement and revenue intelligence agencies. Now, without getting into sensitive details, Abhishek, can you help us understand the kinds of problems AI is solving today in national security and how much of this is about automation versus decision-m support.

[00:02:32]>> Firstly, thank you Shi, thank you for having me on this show and thank you for that decorated introduction to us also.

[00:02:39]Uh, now coming to the question. So at the simplest level you have to understand the problem we are trying to solve is that the intel and the security agency framework that we look at today is drowning in data right and we are talking about billions of signals which are generated on on a monthly basis we are talking about phone records signal intelligence financial data um imagery in intelligence satellite imagery open source intelligence and somewhere within this all this noise the real threat is hidden somewhere the real pattern is hidden somewhere and no human team no matter you know how many hands or how many brains you put together can really do it at you know the kind of speed that we really require. it.

[00:03:11]>> Sure.

[00:03:11]>> And that is the gap that AI is trying to fill, right?

[00:03:14]>> Okay.

[00:03:14]>> Uh so in practice, you can think of it that you know it's kind of divided into three layers. First is kind of fusing all this data together.

[00:03:22]>> No matter what kind of var veracity, variety or speed it is coming at. The second is more about you know surfacing the pattern the the signal the actual intel out of this you know terabytes or pabytes of data. And the third of course is that you have to do it at speed of light right? Right? I mean you have to do it at a speed which is much faster at the threat that is coming in. Right?

[00:03:41]>> And I think I think this next part of the question was in terms of autonomous versus decision making are and we will always firmly be toward the side of decision making.

[00:03:52]>> Um our job is to do all the heavy lifting give the necessary signals give the necessary trace behind it and our reasoning behind it but any automation system should actually take that call.

[00:04:02]The final trigger should always be in the hand of the of the analyst or the decision maker system should not take that.

[00:04:09]>> All right. You know Vishik, historically many critical security technologies have come from global defense contractors.

[00:04:14]What has changed now? Our government agencies becoming more willing to adopt indigenous AI platforms and how significant is the opportunity for Indian companies in this space?

[00:04:24]>> I mean certainly I mean if you look at the recent geopolitical turmoil that is happening all around us and I mean let's get some basics right. See defense and security agency and intelligence cannot be run on borrowed technology. It cannot be run on rented technology. It has to be on-prem. It has to be sovereign. It has to be completely in your control. Uh if you look at India when you're talking about a one and a half billion lies, you're talking about guarding borders which are 10,000 km a coastline which is 15,000 km. You cannot really do it where you know somebody else is pulling the threats and even a single line of code or a bit of bite here or there can basically have catastrophic u uh impediment. So the idea is to basically and I think this is where India is also moving in and if you look at historically fine we are not making our checks but we are asking for the source code and the reasoning behind that is actually to having sovereign control over whatever we do and all the countries across the world are moving towards and that's the gap that we are trying to fill here at by creating sovereign stack and sovereign data which the agencies can be using.

[00:05:18]>> Sure. You know let's get to the fund raise. You plan to use part of this capital to advance agentic AI and build a dedicated physical AI and robotics division. you know, these are two of the hottest themes in AI globally today. Uh, can you give us one specific use case uh, you know, that you're targeting and how close are we to seeing autonomous AI systems playing a meaningful role in security and defense operations?

[00:05:40]>> The second part is a very tricky question, but I'll try to ask it as answer it as best as possible. So, I mean we we are kind of edging our bets on two different fronts but with a common vision. So, first of course is the agentic AI piece. Now, the use cases that we are trying to solve. So agentic AI in a way I mean this is where the world is moving. It is to transform the traditional AI which is kind of you know just reading your data talking to you and converting it into something which can actually execute something for you and give you tools. So in our domain we are looking at use cases wherein you know you're not just conversing with the data you are actually enabling the AI to do the analytics to run statistical models to churn out numbers to understand satellite imagery and then based on whatever the users had asked them you know think of it like 10 different analysts which are at your disposal with all the data access that they have and all the tools that are available and give you one single coherent picture right so I think agentic AI is the future it is something which is happening right now when we talk about physical AI I mean I'll be very candid there's not a lot that we can dable as far as the physical AI is there. I mean for for for certain confidential and obvious reasons. Um but you know we we going to be starting at the place where we are the strongest which is the defense. That is where we're going to start at.

[00:06:50]>> Okay.

[00:06:50]>> And the idea is basically to you know move forward. So that is the near term.

[00:06:54]In the long term we are basically looking at solving more robust use cases and at some point in time getting into the concepts of exoskeletons getting into the concepts of uh um exos and humanoid robots also as we move forward.

[00:07:06]>> All right Abhishek very quickly if you could answer you know you've already made initial inroads into the Middle East and are now looking at global markets national security is an extremely sensitive space you know it's mostly dominated by local players and geopolitical considerations. How do you compete internationally and what gives an Indian AI company a credible edge in this market?

[00:07:26]See you you're absolutely right that you know we we are we are working in an environment wherein you know the sovereignty of solutions very important the kind of edge that we are looking at everybody is security conscious everybody is become becoming conscious in terms of the sovereignty of the solution that they are basically looking at so pretty much every nation has got a similar part of a concern right so for us I think what is important is that you know if you look at most of the established players which are currently there they come with their own geopolitical baggage right And although they are giving you the solution, they're not really giving you the control. Our job is to basically give you the control. So everything that we give you is onremise. It is available with you and the entire control basically is available with the end customer which is the you know uh your security agencies and they basically pull the strings and there is absolutely no security threat in that kind of a scenario and India being a more neutral sort of a country and you know which not having that much kind of a geopolitical baggage the actual input becomes easier for us and you know once anything is tried and tested in India trust me it is ready for the world. All right. You know, we'll have to wrap up our conversation on this note and uh I understand that you're also looking at uh preparing for an IPO. We wish you the very best and many thanks Abhishek for joining us on Startup Street today.

[00:08:39]>> Thank you so much. Thank you to the team. Thank you. It was a pleasure being here. Thank you so much for this time.

[00:08:44]>> All right. And moving on, India is one of the world's largest shrimp exporters, but much of the industry is still driven by small fragmented firms, farms with limited access to technology, market linkages, and global buyers. Aquapulse is looking to change that. The Odishia based agriculture startup has raised 45 cr rupes in a funding round co-led by NAV ventures and backed by IIA and alpha fund. The company is building a tech enabled farm-to-port platform that connects small shrimp farmers directly to domestic and global markets. So how does it plan to scale this model and what's next for India shrimp export story? Joining me now is um Abishek Vivei the co-founder of Aquapulse.

[00:09:21]Abishek thank you so much for taking time out and being here with us on Startup Street. you know you've raised this 45 cr rupes take us through how you plan to deploy this fresh capital especially across tech farmer onboarding and processing infrastructure and export expansion right and what outcomes are you sort of targeting over the next 12 to 24 months >> yeah thank you uh team for letting me and Aquapus coming into the por and sharing our thoughts into it so if you see aqua pulse has been spearheading middle east operations starting from the eastern coast of India and penetrating into the western in part of India. So this fund raise is not a testament to what we are doing.

[00:09:58]>> Who's who's rolling?

[00:10:02]Please continue.

[00:10:03]>> Yeah. So if you see the major context of the funding is that we are spearheading this trailblazing context. We are leading it in the technology aspect. We are leading it into the like physical aspects. So aquapulse is enabling the preh harvest side farming community to produce a product. If you see the seafood community or the market size is somewhere valued upon global seafood market is 49 billion USD. So it is growing at 7% compound annual growth rate which is poised to be something around 800 billion USD. So considering the protein requirement the global community has to consume we have to build sustainable practices that to take enabled and farm level technology is something that we are predicting and disease management systems to waste aquaculture practices is something that we are priing on a aggregation model based on market linkages solid market linkages both in domestic and both in also Middle East Asia and also the infrastructure ability to process a shrimp that is also of the world-class quality. The consumer centric product shrimp that is what we are developing and that too it is devoid of any sort of antibiotics and banned substances. We are on the verge of creating a tech enabled safe seafood platform which will not only be used by the community to consume but it will create a aura for safe seafood and safe protein consumption because the interest of red meat and poultry and other products >> absolutely and this global protein requirement has to be met and that pawn to port pawn to port concept and pawn to plate concept with integrating harvest processing and exports is something that we are closely looking at. Absolutely.

[00:11:49]And you know India is one of the world's largest shrimp exporters but the sector still remains highly fragmented. Right.

[00:11:54]We're running out of time. So if you could quickly tell us what are the biggest pain points for small shrimp farmers today and how are you sort of solving this for them?

[00:12:02]>> The farmers are facing indiscriminate post-h harvest losses. If if you ask me what is the post-h harvest loss because shrimp is a highly perishable commodity.

[00:12:11]We have our own products that same day delivery system, same day processing and same day the product is being proiled where we cut out the processing losses to the tune of from 15% to 2%. So we are working on this infrastructure ability.

[00:12:27]So that the processed shrimp or the frozen shrimp is not only organolytically high in quality but it is also tech enabled and it is also traceability enabled and it is also a labelled product that the product is integrated into processing and harvest.

[00:12:42]So if you ask me what is the biggest dilemma in shrimp farming it's market linkage. It's the right information coming from the right source from the right time from the right channel is the most important thing that aquapulse manages through its aquatic model and also in the back end deep tech model that we enable >> right uh I'll come to the tech part of it in just a bit but you currently export to markets like China Vietnam and Japan right how are sort of global demand trends evolving and where do you see the biggest export opportunity for Indian shrimp over the next few years are you looking at new markets as well >> yes Currently with a free trade agreement done by government of India with European Union. We are planning to embark on a European journey so that our product can be exported to Europe and coming and going further. We are also targeting to sell our products to USA because USA is a bigger consumer and with tariffs issue easing out I think the market is going to open up more because US has always been the evergreen market for shrimp industry. But diversification is the core of aquapulse and we always we always enjoy like playing in safe heavens and the farmers protocols is something that we have developed in such a way that the farmers are also getting diversified like sizes or developing different sizes of shrimps. Those size of shrimps are going to the different markets be it China, be it Japan, be it Vietnam and we are also targeting different countries like Hong Kong, Singapore, Mauritius which are tourist friendly destinations.

[00:14:10]So lots of new markets on the horizon there for Aqua Pulse. You know having expanded from Odisha into Andhra Pradesh and West Bengal. What does the next phase of growth look like for Aquapulse?

[00:14:20]Are you looking at new sea fat seafood categories or deeper integration across the value chain itself?

[00:14:26]>> Yeah, correct. So we are looking for new diversified products. Presently we are operating with Venami shrimp. We are planning to expand into the western part of India that is Gujarat where we would be operating with black tiger shrimp and we would also be willing to operate out of India in a context that we would be uh on more deeper on the export supply chain side and where we would be controlling the product and the entire traceability of the product. So the movement of the product and the quality of the product reaches the right consumer from the right point of time.

[00:14:57]So the hub and spoke operating model with the farmers and that two backend model with the buyers is something that we are trying trying to embark upon.

[00:15:05]>> All right. And where are your revenues currently at and what's the target if you can tell me in 10 seconds.

[00:15:10]>> Yes. So last year we clocked a revenue of 130 crores. So this year we are trying to clock a revenue of 300 crores with the new fresh capital from nav venture and in fund and we are trying with the support of mark investors both these investors. We are trying to also leverage on European market where we will be closing down on something around 250 to 300 odds of revenue seafood exports.

[00:15:32]>> All right. All right. Abishek, thank you so much for taking time out and explaining that to us. We wish you all the best on your journey going forward.

[00:15:37]>> You thank you Tim. Thank you. Thank >> Well, moving on. Meta has entered into an agreement with Reliance to lease the company's first AI enabled data center in Jamnagard. The facility's first phase will deliver 168 megawatt of capacity with an option to scale. Rajna Tendrajani joins us now with the details. Rajna, take us through the contours of this agreement.

[00:15:59]That's right. Meta and Reliance have announced this agreement whereby they would build an AI enabled data center in Jamnagar, Gujarat, which we know is Reliance's hub for uh and the for the ambition going forward. Now the facility will have a capacity of 168 megawatts in its first phase with options to scale further. Reliance will construct it. Met Meta leases it runs entirely on renewable energy and will cool these data centers using desalinated sea water. Now if you know anything about data centers the conversation around electricity, power and water is a consistent one. So they want to solve that by using renewable and desalinated water. Meta will also cover the full cost of energy and water supporting this facility. Now like I said Jamnagar is where Reliance is already developing one of the largest data center campuses in the world. Paired that with Meta's project waterorth which is the longest subcable system ever laid. This puts significant compute capacity on Indian soil connected directly to the global internet backbone. For meta India users which are about billion of us using WhatsApp, Instagram and Facebook. This could mean AI being served from local infrastructure. So maybe lower latency, faster performance and tools that are that are made for Indian users rather than adapted. For Reliance, it is the next step in a partnership that has moved through each layer of the digital stack. The $5.7 billion investment in Jio in 2020 was the first. Then came the Llama AI joint venture and now its physical infrastructure. On clean energy, Meta has contracted nearly 1 gawatt of new solar and wind capacity across five states, 837 megawatts with clean max and 88 megawatts with fourth partner energy. All in all, this partnership could indicate that India is moving from spectator to stakeholder in the global AI race.

[00:17:54]That was the standard disclaimer on your screen. But Anthropic has released a new AI model which the company had itself deemed too dangerous for the general public. Anthropic has rolled out a Mythos class model with certain guardrails. Anthropic has filed to go public on Wall Street this year. The new model named Claude Fable 5 comes with restrictions due to cyber security concerns. Anthropic is also offering an unrestricted version of Mythos to companies and organizations which includes nearly 200 organizations across 15 countries globally. Well, with that, it's time for us to head into a short break. But coming up next, Coralogix has raised $200 million in a series F round, taking its total funding to $550 million. We speak with CEO and co-founder of Coralogics, Ariel Assaraf, on the other side. Stay tuned.

[00:18:56]Well, Coral Logix has raised $200 million in a series frown, taking its total funding to $550 million. The company is positioning itself at the intersection of two key trends.

[00:19:07]Exploding volumes of machine data and the rise of AI agents that can increasingly monitor, investigate, and even operate technology systems.

[00:19:15]Coralogic says traditional observability tools were built for a dashboard world while the next phase is going to be driven by AI powered intelligence and automation. For more on the road ahead for the company, joining us now is Ariel Assuraf, the CEO and co-founder of Kora Logix. Ariel, thanks very much for joining us here. Uh, you know, investors have just committed another $200 million to Kora Logix in a market where capital is pretty selective now. What do you think investors are really betting on?

[00:19:45]Is it the observability opportunity, the AI opportunity, or a mix of both?

[00:19:50]>> That's a great question. I think the main concern investors have today is what is going to be cannibalized by AI and what is to what is actually going to benefit from AI? what's going to get some tailwind from AI. And what we're seeing right now is that companies that are holding significant business data have a real advantage in this world because whatever AI models you're running, they're going to have to feed on data to actually generate insights.

[00:20:16]and observability being the most highest volume, highest dimension, real-time information the organization has become extremely important these days because with AI you can actually infer business level insights in an instant.

[00:20:31]>> You know you made the statement that dashboards are no longer the starting point in intelligences. So for someone who's running engineering teams today for instance what does that shift actually look like in practice and how different could operations be say 3 years from now?

[00:20:46]operations uh used to be really heavily rellyant on dashboarding and alerting.

[00:20:51]People would open up the dashboard as their first stop whenever there was an incident or anything that they wanted to know. With the interfaces today and models today, you can just ask what you need to know. So if something goes wrong, instead of having to look at dashboards and sift through data, you can just ask why is this user failing?

[00:21:09]Or if you want to understand your business better, you can just ask what is causing my my customers to be frustrated? what is causing me to lose revenue? What issues do I have that I need to fix today? That ability makes dashboards, alerting, and saving for data obsolete. And we believe it'll happen within the next 2 to 3 years.

[00:21:27]Dashboards will remain important as a a sync tool, meaning actually a dashboard that's hang on the wall and the entire team looks at, but not as an investigation tool.

[00:21:38]>> Every software company today seems to be adding AI co-pilots and AI agents. What makes your approach different? I mean, why do you believe that your architecture gives you an advantage in this AI era rather than simply adding AI on top of a legacy observability track?

[00:21:54]>> The biggest difference is um AI is not something you can just apply to a legacy software. the ability that Core Logix has to build a complete data lake that actually stores all the data infinite retention with extremely powerful uh query engine that lets you uh model the data on the fly. Data prime or core engine actually models data on the fly based on the question that the model asked. When you have a defined schema that is strict, you're not really getting AI, you're just getting a plain language API. But with the ability to model to have infinite data to have full coverage agents can actually run on top of coral logics and infer those deep insights that people are seeking to >> you know AI applications are also generating unprecedented amounts of data and you've argued that this is exposing the limitations of traditional monitoring tools. How much has the scale of the problem changed in the last couple of years and what are customers telling you that they struggle with the most? I think customers are struggling with a variety of issues today that the legacy systems are running into. One, uh, models need context. Models unlike historically or not even historically, 5, six years ago, everyone were talking about how do I reduce the amount of data I have? How do I only store what matters? How do I only store what important? But models today actually seek for the entire data set. They need full context. models today need full cardality and actually understanding at the user level, not the machine level.

[00:23:26]And they also need longer retention. You think of the current tools out there in the market, they're too expensive for full coverage. They do not allow you to have thousands and tens of thousands of keys. They do not do not allow you to model data on the fly when you need to ask a question that was not already set in the data. And also, they do not have the retention you need to build the long-term trends that models are really good at. And so the the capabilities that we've built, streaming everything in a in a real-time way and then writing a data lakeink to the customer's own storage is the only solution moving forward.

[00:24:00]>> You know, one of the promises of Coral Logix is full fidelity observability without a runaway cost. Now, historically, companies have had to choose between keeping all their data or controlling spend. How are you solving for that trade-off?

[00:24:14]>> Exactly. So I think one of the biggest problem is that um the old observability was you send all your data the data is being stored with the vendor you got to pay for the retention the cost per gigabyte was enormous and with core logics because of our architecture is fundamentally different we analyze everything in stream and then we write it to the customer's own storage. First of all we use lowc cost storage and high compression. Secondly we don't need to mark up storage because it's the customer's storage and we don't control his retention. We don't charge for it.

[00:24:44]And thirdly, the streaming analytics is extremely extremely efficient. Uh, Coral Logic slashes cost by over 60 sometimes 80% for our customers using this technology, but also allowing them an open format that they own in their own storage, which is a big advantage these days because companies are looking at what can they build on top data that they're generating and they want to own that data in a format that allows them to actually create on top of it and not be dependent on the vendor. All right, Ariel, we'll leave it at that. Thanks very much for joining us here today on Startup Street.

[00:25:20]>> Thank you very much. I appreciate it.