When India had its AI moment: A Story from 2005
Even today, while Indian engineers lead AI developments globally, we struggle to back homegrown innovations that could compete with global tech giants
Ayyappa Nagubandi
The recent success of China’s DeepSeek, built at just $5.6 million, has sparked debates about India’s position in the global AI race. As someone who’s been in the technology space for over two decades, this discussion takes me back to 2005 — a time when we were quietly building something revolutionary in a small office in Hyderabad.
At TrulyIntelligent Technologies, we weren’t just dreaming about AI — we were building it. Our system, which we fondly named ‘Lilo’ (after the character in The Fifth Element), could understand and process natural language in ways that seem remarkably similar to today’s ChatGPT. The difference? We did it on desktop computers, without the massive GPU farms that everyone considers essential today.
Let me share a practical example we demonstrated back then. Imagine asking a computer, “How many airlines from the US travel to Kerala per week?” Today’s search engines would still do what they did back then — blindly list any webpage containing “Kerala” and “airlines.” Our system could understand the question’s intent, process the context, and provide specific, relevant answers. If this sounds familiar, it’s because this is exactly what we celebrate about ChatGPT today.
The technology caught attention. It was widely featured in newspapers and media. We were processing natural language queries, understanding context, and generating relevant responses — all in 2005, when most tech giants hadn’t even begun thinking about conversational AI.
But here’s where the story takes a turn that reveals a deeper truth about Indian innovation. When it came to scaling and competing with global giants like Google, our investors became hesitant. It wasn’t about the technology — it was about the courage to back something that seemed too ambitious for its time.
Even today, while Indian engineers lead AI developments globally, we struggle to back homegrown innovations that could compete with global tech giants.
The DeepSeek success proves what I’ve always believed — you don’t need billions of dollars or thousands of GPUs to innovate in AI. What you need is creative problem-solving, smart algorithms, and most importantly, a hungry mindset that doesn’t accept limitations.
Looking at old files, I found our project documentation. The technical architecture seems almost prophetic now — contextual processing, natural language understanding, automated learning — concepts that are at the heart of today’s AI revolution. We built these systems when India’s tech infrastructure was far more limited than it is today. We didn’t have AWS or cloud computing. We worked with what we had, and we made it work.
This isn’t about claiming “we did it first.” It’s about showing that India has always had the capability to innovate at a global level. The next DeepSeek can come from India. We have the talent, the technical expertise, and the innovative spirit. What we need is an ecosystem that believes in taking bold bets on revolutionary ideas.
The simple truth is, innovation isn’t about resources, it’s about resourcefulness. When people tell me we need massive computing power for AI innovation, I think back to our desktop computers from 2005. When they say we need billions in funding, I remember what we achieved with limited resources but unlimited imagination.
The question isn’t whether India can innovate in AI. The question is whether we’re ready to believe in our own potential and back our innovators when they dream big.
(The author is an inventor, entrepreneur who has been building innovative technologies for over 25 years)
