India’s Defence AI Reality Check

There’s little denying that India’s defence tech moment is here. Thanks to VCs loosening their purse strings, a focus on building up indigenous drones and UAVs, greater policy support, and more startups entering the space, defence tech is no longer confined to the sidelines. 

Nearly 90% ($68 Mn) of the total $78 Mn raised by defence tech startups over the last decade came in 2025 alone.

But if there’s one thing that the recent conflicts in the Middle East have made clear, it’s that the future of defence tech is AI and software-driven systems. In this regard, key gaps remain in India. 

India does not yet have large defence AI companies like Palantir, Anduril, Skydio and others, which are increasingly playing a key role in global defence tech adoption, primarily in the US. 

As global conflicts become more focused on AI, autonomous systems, and software-defined hardware, the rules of the game are being rewritten with each passing day. For India, which has long relied on imports for critical defence technology, this shift presents both a challenge and a rare opportunity to build at home.

But, are we ready for the defence AI future? Let’s find out in this week’s edition of The AI Shift.

India’s Defence AI Dream Isn’t There Yet

While India’s defence tech ecosystem is moving in the right direction, the progress is still uneven. 

From the spectacle of a defence tech startup founder, Amardeep Singh of drone defence company Armory, “This time will be seen as the birth of India’s self-reliance, because whoever your friends are today, [may become the enemy tomorrow].”

According to Singh, India’s software pedigree is well established, but manufacturing remains the harder problem. Building at the scale, quality, and precision that defence and aerospace demand is where the country still needs to close the gap.

India’s defence budget has seen a 15.19% YoY increase to ₹7.85 Lakh Cr in FY27, which Santosh Mishra, cofounder of IG Defence, sees as a step in the right direction.

This has helped build a solid scaffolding around indigenisation. For instance, the Innovations for Defence Excellence (iDEX) scheme funds startups up to ₹25 Cr under the ADITI scheme. Then, the Army Design Bureau, along with equivalent bodies in the Air Force and Navy, publishes annual compendiums of problem statements that startups can bid against. This, Singh said, is a structured procurement path that barely existed five years ago.

He, however, identified talent attraction as a major bottleneck. “Defence startups often struggle to compete with compensation levels in mainstream technology sectors, even though mission-driven work can be a strong motivator for engineers.”

The Hardware Gap In The Defence Stack

When we drill down into the ecosystem, it’s clear that India’s defence stack is also being shaped by a software-first, AI-led approach. Hardware realities, however, continue to hold it back. 

While startups are building smarter, upgradeable systems, much of the core electronics and IP still depend on foreign sources. This is precisely the gap between surface-level indigenisation and true self-reliance.

“At the core, it will definitely be a software-led AI approach, while hardware will be an integral part,” said Mishra of IG Defence.

Armory’s CEO Singh envisions a shift toward software-defined hardware, allowing drones to be upgraded post-delivery. This ensures detection and neutralisation algorithms evolve alongside rapidly changing threats, providing long-term value and security without replacing physical components.

Current priority areas span computer vision and automated threat classification, logistics optimisation under battlefield conditions, and edge-deployed AI in air-gapped environments. This is because defence systems cannot run on public cloud infrastructure. 

Highlighting the importance of AI integration, Preet Sandhu, the founder and MD of drone maker AVPL International, said, “If you are not using AI in drones, then you are just picking a stone and throwing it at somebody. But if you are putting AI in it, that stone has all kinds of senses — where to go, what to do, and how to do it.”

However, India’s indigenisation story needs to be more nuanced. “While frames, motors, and batteries are increasingly being made domestically, the core IP, particularly for electronic components like drone controllers and motherboards, are still being made abroad or through technology transfers from China, Israel, or Russia,” Manish Gupta, managing partner at growX Ventures. 

Gupta added that it may look like the manufacturing is happening in India, but the IPs originate elsewhere. “Companies that control a full stack can swap components when supply chains fracture,” he added about how companies can build moats.

Sai Pattabiram, founder and MD at Zuppa, seconds Gupta’s sentiment. He said wireless communication systems and camera payloads continue to be imported because the skill sets needed to close that gap are nonexistent. “Government, corporate, and academic R&D are unable to keep pace with the evolution of defence technology due to structural limitations and a lack of skin in the game.”

Where Is The VC Money Flowing?

All said and done, defence tech has quickly become a serious area of interest for VCs, as they are now actively backing startups across drones, autonomous systems, and digital warfare. 

Abhishek Prasad, the managing partner at Cornerstone Ventures, which has backed aerospace & defence R&D company NewSpace Research and Technologies (NRT), lays out a three-layer stack for evaluating opportunities in this space.

• Physical Defence Infrastructure: This covers the likes of tanks, ballistic missiles, rockets and other ammunition, which is dominated by large defence contractors, and therefore less attractive for VCs.
• Autonomous Defence Systems: The most compelling near-term bet for startup investors, where hardware IP is compounded by AI and algorithmic capabilities. As seen in the case of drones, UAVs and other aerospace solutions.
• Cyber Defence: This segment offers the highest commercial upside, extending beyond government into financial and corporate security markets. Digital infrastructure protection and cyber defence tech will form a key part of next-gen solutions.  

Vipul Patel, partner at IIMA Ventures, which has backed ideaForge, Agnikul, and Sagar Defence, among others, sees the most credible bets emerging at the intelligence stack layer, specifically perception, sensor fusion, and autonomous decision-making. 

Value, he argues, is visibly shifting from hardware to system intelligence. His thesis on viability is to validate with Indian forces first, then scale exports to Southeast Asia, Africa, and allied markets. “The strongest moats will belong to startups that serve defence and adjacent commercial markets simultaneously.”

Prasad of Cornerstone invests in founders who balance domain depth with commercial speed, avoiding over-engineering to ensure they don’t miss market windows. 

Gupta from growX Ventures echoes this, adding that his firm backed Armory pre-revenue, pre-product, and pre-technology — purely on the strength of the founding team.

Ready For The World?

India’s demanding topography and geography is a competitive advantage for defence tech startups. Systems tailored for these on-ground conditions are battle-hardened and can typically stand up to global procurement standards demand.

IG Defence’s Mishra said, “Built in India, built for India, but later for the world. Any technology we are building for the extreme terrains of India can be used by any other country as well.”

Top Stories From India & Around The World

• VerbaFlo Raises $7 Mn Seed Round: Real estate AI startup VerbaFlo has raised $7 Mn led by Pi Labs, with participation from Haatch, Navigate Ventures, and Old College Capital. Its conversational AI platform automates leasing, operations, and tenant engagement for property developers across email, WhatsApp, and webchat in over 200 languages.
• Razorpay’s AI Avatar: Besides partnering with Sarvam this week for voice commerce, IPO-bound Razorpay is repositioning itself from just a payments provider to an AI-powered business operating system for small merchants through its Agent Studio launch. Here’s our deep dive into this turn
• Google AI Studio Gets Full-Stack Vibe Coding: Google upgraded AI Studio’s vibe coding experience with its Antigravity coding agent, enabling production-ready apps from prompts. New capabilities include multiplayer support, Firebase-backed databases and authentication, Next.js support, Secrets Manager for API keys, and persistent cross-session memory.
• Meesho Launches Voice Shopping Assistant: Meesho launched Vaani, a GenAI-powered conversational voice shopping assistant designed for tier II+ users who shop through natural language rather than filters and search

The Weekly Buzz: Engineers May Soon Get AI Budget On Top Of Their Pay Cheque

“How many tokens come along with my job?” Well, this may become the defining question of Silicon Valley recruiting if NVIDIA CEO Jensen Huang’s vision of the near future comes to pass.

At his GTC 2026 keynote, the NVIDIA CEO laid out a striking proposition — every engineer at NVIDIA would eventually receive an annual token budget worth roughly half their base salary. For someone earning a few hundred thousand dollars a year, that means an equivalent sum in AI compute credits — designed, in Huang’s words, to ‘amplify’ their output tenfold.

The logic is straightforward. Engineers with access to tokens are more productive. Companies that offer generous token budgets attract better talent. And as AI factories scale up production, the cost of those tokens will be underwritten by the very infrastructure partnerships NVIDIA is building with enterprises worldwide.

But the token budget idea is only one piece of a larger structural shift Huang described. He argued that software companies are now transitioning into what he called “Agentic as a Service” or AaaS companies. In this new model, enterprises don’t just provide tools for humans to use; they deploy autonomous agents that act on behalf of users. The companies enabling this transition become simultaneously token users for their own employees and token manufacturers for their customers.

This is not a distant forecast. Huang framed it as a renaissance already underway — one where the measure of an engineer’s leverage is no longer just their skills or their team size, but the compute budget they command. Whether the rest of the industry follows NVIDIA’s lead on formalising token budgets remains to be seen, but the direction of travel seems hard to argue with.

Startup In The Spotlight: Entri

India has over 300 Mn vernacular language speakers across Malayalam, Tamil and Telugu, yet most edtech platforms are built for English-first, white-collar aspirants. For the vast majority targeting government jobs, police exams, or entry-level private employment, there is little quality infrastructure for exam prep and upskilling in their native language.

Kochi-based Entri, founded by Mohammed Hisam and Rahul Ramesh in 2017, was built around this gap. What began as a mock exam platform for Kerala Public Service Commission aspirants has since grown into a full-stack vernacular learning platform spanning exam prep, spoken English, coding, digital marketing, and job-linked upskilling programmes.

The platform today operates across three product tiers — crash courses with recorded content and mock exams, cohort-based mentorship programmes, and intensive placement-linked upskilling tracks with certifications from partners, including PwC and the Institute of Technology, Chicago.

AI is increasingly central to Entri’s product strategy. The platform uses AI to track student attention in live classrooms, automate teacher follow-ups, and improve course completion rates. The next phase involves AI avatars of popular teachers that can answer student queries around the clock.

Backed by Google, Neon Fund, Udemy, and Omidyar Network, among others, the startup has raised around $17 Mn to date. The startup is now focused on going deeper into its core three-language markets rather than expanding geographically, targeting 35,000 new paid users per month as it scales.

Prompt Of The Week

What prompts and hacks are CTOs, CEOs and cofounders using these days to streamline their work? 

Here’s Arjun Nagulapally, CTO of AIONOS, on using AI to streamline product decision-making:

“Act as my VP of product and head of customer insights.

Input: [paste product analytics snapshots, user research notes, support tickets, feature requests, roadmap, or call transcripts].

• Synthesise the top 5 user problems and usage patterns you see, grouped by segment/persona.

• Identify which parts of the product are delivering outsized value versus underperforming, and explain why in plain language.

• Propose 3-5 high‑leverage product bets for the next 1-2 quarters, with: target user, expected impact, effort level, and main risks.

• Translate this into an exec‑ready ‘Product Decision Brief’ with:

• Problem statement

• Evidence (data + voice of customer)

• Options and trade‑offs

Clear recommendation and success metrics.

Keep the output concise enough to drop straight into a product/EXCO review deck.”

Editor’s Note: Some prompts may need to be adjusted by users for best results or may not work as intended for certain users.

[Edited by: Shishir Parasher]
[Creatives by: Varshita Srivastava]

The post India’s Defence AI Reality Check appeared first on Inc42 Media.