🔥 Money & Machines |
In today’s Money & Machines edition, we bring you our observations from the India AI Impact Summit held in New Delhi this week. On the fintech side, we write about fresh partnerships and early experiments; and on the AI side we write about foundational model launches, global big-tech presence and billion-dollar infrastructure bets. |
AI everywhere. Fintech…somewhere |
Since the AI Impact Summit is the talk of the town – with every second tweet on my social media feed dissecting everything from global "AI stars" like Sam Altman, Demis Hassabis and Dario Amodei to innovation hype to some venting about Day One mismanagement. I'm quietly glad I skipped the chaos and showed up on Day Two. |
By the time I walked into Bharat Mandapam on Day 2, the "war room" the government promised seemed to have worked its magic. Despite some 2.5 lakh registrations, the day felt like a breeze. The expo halls were massive – almost intimidating – and despite the crowds, things seemed under control. Or perhaps scale, I guess, fixes everything. |
I spent most of my time wandering through 5-6 buzzing halls from the likes of Sarvam and sprawling government pavilions like MeitY and many more. Robots everywhere. Robots that waved. Robots that blinked. Robots that looked like they might replace us by 2028. Spacecraft models… |
But what stayed with me wasn't the tech. It was the crowd. |
Students. Young founders. Curious faces. There's something genuinely moving about seeing rows of students with "glittery eyes," staring at robots and spacecraft, trying to grasp the sheer scale of what AI actually is. The energy was palpable. The excitement almost childlike. You could actually feel a generation is trying to decode its future. |
And then, instinctively, I began looking for something familiar. |
Fintech. |
Surprisingly, the sector that digitized money, rode UPI to scale, and reimagined lending – felt shy. |
Mastercard had a dominant footprint, and NPCI made a splash by launching FiMI – a domain-specific Small Language Model (SLM) designed to handle the messy world of UPI disputes and customer service. Even Razorpay held its own, announcing a "vibe coding" partnership with Replit to help developers monetize AI-built apps instantly. |
But the most interesting fintech of the day was – none other than Paytm. |
|
The country's largest consumer fintech was present, almost hidden in a corner with an unusually small stall. It felt like a deliberate choice – or may be they didn't want to disturb the robots :) |
Naturally, I stood on the sidelines for a bit and watched 5-6 young peeps crowding the stall and asked the representative with the ultimate existential question, "What is Paytm actually doing at an AI Summit?" |
The young company representative, to his credit, handled the probing well… and from what I gathered, they were collecting email IDs, trying to understand/work how to leverage AI for merchants. |
Seeing a fintech giant sitting in a corner, silently taking notes, felt different. A decade ago, fintech – especially Paytm – walked into banking conferences as a bold, impatient challenger. This week, at an AI Summit, fintech felt almost…traditional.…almost like the incumbents once looked when fintech first arrived. |
Meanwhile, deep-tech startups were confidently demoing foundational models, autonomous systems, AI agents that promise to automate workflows, write code, research and maybe even write investigative stories and newsletters (calm down). |
I'm no AI expert, but it was fascinating to see the role reversal. Usually, fintech is the one disrupting the room; here, they were the ones trying to find a seat. |
This scene made me wonder – will three or four years from now, will the tables turn? Will these data and cash-rich fintechs make bold AI bets and capture premium, glittery stall spots at future summits? Or will AI-native companies quietly eat into financial services from the edges? |
Because disruption rarely waits for anyone to feel ready. |
I'm happy to close this piece with a line – money finally met machines this week! |
-by Arti Singh |
|
India hosted AI amid pomp. Now comes the harder part |
No country can host a global AI summit of this scale without ensuring its own innovation takes centre stage. That is why the unveiling of India’s foundational large language model (LLM), built from scratch by startup Sarvam AI, carried symbolic and strategic weight. It signalled that India is not merely consuming artificial intelligence, but building it. |
Ideally, the summit’s narrative would have remained anchored around this milestone and the broader domestic push to strengthen India’s AI ecosystem. With the CEOs of Google, OpenAI and Anthropic in attendance, alongside world leaders such as French President Emmanuel Macron, the message was clear: India intends not just to participate in the AI revolution, but to help shape it. |
|
But in the initial days long queues and access bottlenecks tested patience while an episode involving a university exhibit briefly drew online scrutiny. |
Corrective steps were taken, apologies were issued, and credible innovations continued to attract attention. While execution was not flawless, the ambition and long-term direction behind India’s AI push remained evident. |
Notwithstanding moments of disruption, the summit managed to hold its ground. |
Sarvam AI, the first startup selected under the government’s IndiaAI Mission, launched two large language models -- Sarvam-30B and Sarvam-105B -- which it said were fully developed in India. Startups such as Gnani.ai and BharatGen also introduced indigenous AI models. |
Pratyush Kumar, co-founder of Sarvam, said the company’s Sarvam-105B outperformed the 600-billion-parameter DeepSeek R1, as well as Google’s Gemini 2.5 Flash, on an Indian-language technical benchmark. |
Yet even as the launch generated excitement, Sarvam’s leadership acknowledged that significant work lies ahead. In a media interview, co-founder Vivek Raghavan described the models as an important first step, noting that it may not operate at the scale of frontier systems such as Gemini, Anthropic’s models or ChatGPT. |
This aligns with the government’s broader AI vision. The Economic Survey 2025–26 advocates a “bottom-up” strategy focused on application-led innovation, local language needs and frugal AI, rather than immediately replicating capital-intensive frontier foundation models. |
Raghavan also pointed to capital constraints in building larger-scale systems. Training and sustaining frontier models demand enormous compute and sustained funding, something Indian startups are still navigating. Globally, AI firms such as OpenAI and Anthropic have raised billions to build and scale their models. Earlier this year, Anthropic announced a $30 billion Series G round at a $380 billion post-money valuation, a single round that dwarfs India’s cumulative private AI investment of roughly $11 billion between 2013 and 2024. |
The funding asymmetry underscores a structural challenge: building foundational capability is one milestone, competing at the frontier requires sustained capital and infrastructure depth. |
It is also worth noting that the government in the Budget has proposed to scale back spending under the IndiaAI Mission for FY27, from Rs 2,000 crore to Rs 1,000 crore. While startup funding alone is not a measure of efficiency, capital intensity remains a defining factor in the AI race. |
Encouragingly, startups such as Sarvam, backed by Peak XV Partners and Lightspeed India, have rolled out foundational models despite these constraints. Global AI leaders acknowledged India’s potential. Sam Altman noted that “India has all the ingredients: homegrown tech talent, a national strategy, and an infectious optimism about what AI can do for the country,” while Google’s Sundar Pichai remarked that India “is going to be a full-stack player in AI.” |
Another positive outcome was the renewed commitment from Indian conglomerates including Reliance Industries, the Tata Group and the Adani Group to deepen AI investments. Mukesh Ambani said Jio and Reliance Industries would invest $110 billion over the next seven years in AI and digital infrastructure, framing it as nation-building capital rather than speculative spending. |
Innovation was visible beyond the sprawling pavilions of Google, Qualcomm and Nvidia. In the Bihar pavilion, startup Neogentech displayed a humanoid robot, raw and occasionally glitchy due to bandwidth constraints, but ambitious. Students from across India were also active participants, building connections and seeking internships and freelance opportunities. |
Not all commentary was celebratory. Vinod Khosla warned that India’s IT services and BPO industries could “almost completely disappear” in the coming years due to AI-driven automation. Vineet Nayar, former CEO of HCL Technologies, cautioned that IT firms are likely to prioritise profitability over employment as automation reshapes traditional service models. |
One striking moment came during a photo opportunity featuring global AI leaders. It drew attention when Sam Altman and Anthropic CEO Dario Amodei appeared to avoid holding hands during the group shot. While awkward, it was also symbolic. The leaders may share stages, but they represent companies locked in a high-stakes race for talent, capital, compute and market share. Strategic alignment in AI is not sentimental; it is shaped by incentives and scale. |
For India, that distinction matters. Global AI firms will engage deeply with the country as a market, talent base and policy partner so long as strategic interests align. Collaboration exists, but it is pragmatic. In the race to build frontier systems, national ambition and corporate ambition intersect but they do not always coincide. |
-by Joseph Rai |
