The AI Cyberwar Has Already Started. India Is in the Middle of It.
India processed 18.3 billion digital payment transactions in March 2025 alone. 265 million cyberattacks recorded that same year. Nearly 60% originated from the China-Pakistan axis.
Let that sink in.
India built one of the most ambitious digital public infrastructures in human history. UPI. Aadhaar. ONDC. DigiYatra. 971 million internet users. A digital economy worth $402 billion. State systems touching every citizen.
The attack surface has grown at exactly the same pace. And the nature of the attack has fundamentally changed.
For most of the internet era, a cyberattack required humans. Skilled ones. Weeks of probing, testing, building exploits. Expensive. Slow. Limited in scale. AI has broken that model entirely.
Anthropic’s newest model, Claude Mythos Preview, can autonomously discover, chain together, and exploit software vulnerabilities more effectively than almost every human researcher alive. At unprecedented scale. Cybersecurity experts are calling it a watershed moment in the history of the field. It is so capable at offensive cyber operations that Anthropic decided not to release it publicly at all. Instead it is being deployed only to select US technology companies to shore up American cyber defences.
That decision tells you everything you need to know about where this technology has arrived.
Now ask yourself what this means for India.
What India has built and what that means for its exposure.
UPI is not just a payments application. It is the connective tissue of the Indian economy. Every street vendor, every small business, every government transfer flows through it. Aadhaar underpins identity for over 1.3 billion residents. ONDC, DigiYatra, e-governance platforms touch nearly every citizen in the country.
A sustained AI-enabled attack on that single shared rail is not a corporate incident. It is a sovereign event that could freeze the economic activity of an entire nation. The security stakes scale with adoption. And adoption in India is near-total.
Who is on the other side.
India is not the frontier power in this race. It is a high-value target caught between the two parties who are. Pakistan-linked APT groups specifically target Indian defence networks and diplomatic systems. Chinese state actors probe India’s power grids, financial systems, and government platforms on a systematic basis.
India’s cyber defensive capabilities are assessed as superior to Pakistan’s. And inferior to China’s.
That sentence deserves a moment.
China’s AI models today lag behind America’s frontier models by roughly seven months. Within that window they will close further. The tools already capable enough that American labs refused to release them publicly will soon be in the hands of actors with demonstrated intent to target India. One AI-driven espionage campaign in 2025 already saw an autonomous tool execute 90% of malicious actions without a single human instruction.
Indian organisations face an average of 2,011 cyberattacks every week. Significantly above the global average. Less than 9% of sensitive cloud data in India is encrypted. India faces a shortage of over 80,000 cybersecurity professionals and trains only 2,000 new ones annually. The gap does not close. It compounds.
The strategic choice India faces right now.
Buying American or Israeli cybersecurity tools solves the immediate problem. It creates a different one. A country whose digital spine is defended by foreign technology, trained on foreign threat data, optimised for foreign adversaries, is not truly sovereign in its digital infrastructure even if it owns the infrastructure itself.
India has navigated this tension before. It was entirely dependent on foreign molecules and foreign IP to supply medicines to its own population. It made a deliberate choice to build domestic capability at scale. The result was the pharmacy of the world. The question is whether India applies the same logic to cybersecurity.
Indian founders are already building the answer.
Safe Security has raised $170 million total including a $70 million Series C backed by Cisco and John Chambers. They translate technical cyber risk into financial exposure language that boards actually act on. Fortune 500 clients. Global scale. Indian IP.
CloudSEK provides AI-powered dark web monitoring, attack surface mapping, and brand threat detection. $30 million raised. 250+ enterprise clients across banking, healthcare, and government. Sequretek’s AI cuts security alert volumes by 95%, making enterprise-grade operations accessible to hospitals and mid-market banks that cannot afford 24-hour security teams. QNu Labs is building quantum-safe cryptography for the post-quantum threat every BFSI and defence institution will face this decade.
All Indian companies. All trained on Indian threat data. Optimised against the specific adversaries that target India. No American company has spent years mapping Pakistan-linked APT behaviour against Indian defence networks. No Israeli firm has built models around UPI-scale payment fraud. Indian companies have no choice but to solve these problems. That constraint, as it so often does, produces advantage.
The window to act is measured in months. Every month below 9% encryption is another month of unacceptable exposure. Every month Indian capital fails to back Indian cybersecurity companies is another month global capital fills that gap on global terms.
The AI cyberwar is not coming. It is already here. And India’s digital infrastructure is squarely in the middle of it.
If you are a founder building in this space, give us a shout. Let’s talk: https://seafund.in/
Table of Content
- 1. What India has built and what that means for its exposure.
- 2. Who is on the other side.
- 3. The strategic choice India faces right now.
- 4. Indian founders are already building the answer.