Also in the letter:
■ India’s startups layoffs
■ TCS’s AI diktat to staff
■ InfoEdge’s growth-stage fund
Building for consumers is hard because of free products: Sarvam AI cofounder Vivek Raghavan
Building AI for consumers is tougher because it has to compete with free tools like ChatGPT and Gemini, said Sarvam AI cofounder Vivek Raghavan in a fireside chat with ET’s Samidha Sharma.
He added that Indian large language models will only see wider use if they offer real value.
Sarvam’s monetisation plan: Raghavan said the company is mainly focused on enterprises. “Our approach is largely B2B. We’re also working with governments and different sectors. These are the primary channels we are focused on right now,” he said.
On frontier models: Raghavan explained why India should develop its own foundational models, drawing a comparison to countries that have nuclear power and those that do not.
“Many countries are perfectly fine using ChatGPT, Gemini, and other frontier models, and then building applications on top of them. That’s a reconciliation they’ve made. India is somewhere in between. We have to decide which side we are on,” he said.
Also Read: AI Impact Summit: Beyond the optics, where is the ‘foundational’ capital?
Neysa targets IPO in the next few years: CEO Sharad Sanghi
Cloud infrastructure startup Neysa is aiming to go public in a few years, founder and chief executive Sharad Sanghi said on Thursday at the event.
Quote, unquote: “We’ve got a partner like Blackstone, and the aim is to go public in a few years. Once we have that growth engine going and know that we can start growing at the pace the market expects, we can hopefully go public,” he said.
Why sale over IPO: Sanghi said a majority stake in his first startup Netmagic was sold to let investors and employees exit, as data centres were still a new market at the time.
“That meant we didn’t have to raise capital, because data centres are a very capital-intensive business, and NTT had a lot of money that we could use to scale,” he said.
Also Read: India’s Rs 10,000 crore AI mission fund akin to weekend expense for OpenAI: Fractal CEO
India is nowhere close to building frontier AI models, and everyone knows it: Wingify’s Paras Chopra
At the event, Wingify founder Paras Chopra said India is far from building frontier AI models and must invest in deep, basic research if it wants to catch up in AI.
On research: Chopra noted that today’s AI models are the result of years of research that once seemed pointless, showing how OpenAI kept working for years before finally making a breakthrough.
India’s biggest constraint: Chopra stressed that India’s biggest constraint is the mindset. “We need to shift from a business mindset to a builder’s mindset — from chasing near-term profit to building the future,” he said.
Also Read: ET AI Conclave & Awards 2025: “If it works in India, it can work in 40–50% of the world”
Don’t rush to deploy AI build foundations first: PhonePe CTO Rahul Chari
Rahul Chari, founder and chief technology officer of IPO-bound PhonePe, said companies should move quickly and embrace AI’s full potential — but first put the right organisational scaffolding in place before rolling it out widely.
On AI use: This may not be the most popular opinion, but discretion is often the better part of valour. AI is here to stay. Its potential is immense, said the CTO. “The real challenge is using it the right way, not just the fastest way,” he said
PhonePe’s AI strategy: Chari said, “At PhonePe, we think about this in three ways. First, we want to increase the efficiency of a large engineering team significantly. Second, we want to take AI-first features to production as fast as possible for our users. Third, we want to increase the efficiency of operations across the entire organisation.”
Apple in talks with banks to start payment service in India: Report
Apple is in advanced discussions with leading Indian lenders — ICICI Bank, HDFC Bank and Axis Bank — alongside global networks Mastercard and Visa to launch Apple Pay in India by mid-2026.
Tell me more:
- Apple Pay is expected to support both NFC-based tap-to-pay transactions at point-of-sale (POS) machines and India’s UPI QR-code payments, run by the National Payments Corporation of India.
- News of the potential entry weighed on One97 Communications (Paytm) shares, which slid amid competition from Google Pay, PhonePe and Amazon.
Yes, but:
Zoom out: The move fits CEO Tim Cook’s broader India strategy, which spans retail expansion, local iPhone manufacturing, and services growth. Apple is also opening its sixth retail store in India this week, in Mumbai’s Borivali.
Also Read: SBI Mutual Fund buys 2.4 lakh shares in Pine Labs, crosses 5% stake
Over 4.5K startup employees lose jobs since July
India’s startup ecosystem has seen more than 4,500 layoffs since July last year, TOI reported, citing Longhouse data, as funding patterns shift and profitability takes centre stage.
Job cuts:
Redesigned workflows: Founders are designing companies differently from day one, limiting large-scale hiring, prioritising senior and revenue-critical roles, and building “lean by design” teams, Anshuman Das, CEO and founder at Longhouse, told TOI.
Hiring picture:
- Entry-level roles remain cautious.
- Demand is rising for professionals with 4–10 years of experience, especially in AI and green-tech, said Viswanath PS, MD & CEO at Randstad India.
- More rolesare shifting to tier-II cities, alongside selective global remote hiring to manage costs.
TCS is asking staff to use AI even if it hits revenues: CEO K Krithivasan
Tata Consultancy Services (TCS) is urging employees to deploy AI-led efficiencies even if it dents near-term revenue, CEO K Krithivasan said.
Verbatim: “If you find that you can do something faster, better, cheaper with AI, you should probably go and tell your customers, even if it cannibalises your revenue,” Krithivasan said, outlining the message to TCS’s 600,000 employees as the firm adapts to the AI cycle.
Yes, and: Change must start at the top, Krithivasan said. “We insist that all the senior management build something of their own, so they understand…it’s not about asking questions, it’s how you build,” he said.
Also Read: DeepSeek withholds latest AI model from US chipmakers including Nvidia
Amazon’s $50 billion OpenAI investment may depend on IPO or AGI milestone: Report
Tell me more:
- The structure under discussion includes $15 billion upfront.
- A further $35 billion would be tied to strategic triggers such as AGI progress or public listing.
Tracking investments: Talks follow Reuters’ January report on a potential partnership and OpenAI’s seven-year, $38 billion cloud deal with Amazon, aimed at scaling its compute backbone.
What lies ahead:
Capex push: The investment race unfolds amid soaring AI capital expenditure.
Info Edge launches first growth-stage investment vehicle with Rs 250-crore corpus
Info Edge has set up B8 Fund-I, a Rs 250-crore growth stage fund, marking its first dedicated vehicle for later-stage investments, according to an exchange filing.
Tell me more: The Naukri.com parent already runs four early-stage investment platforms:
- Info Edge Ventures, which has raised Rs 2,300 crore across three funds
- Capital 2B with a Rs 280-crore corpus
- Redstart Labs, focused on AI-led startups
- Direct balance sheet investments
Info Edge’s early bets include Gnani AI, Ixigo, Shiprocket, Zingbus and Truemeds.
Goal setting:
- B8 Fund-I will back growth-stage, tech-enabled companies in India or businesses primarily focused on India.
- The fund is registered with Sebi as a category-II alternative investment fund and will run for eight years from the first close.
Lending firm Prayaan Capital raises Rs 110 crore from Peak XV Partners
Non-banking finance company Prayaan Capital has raised Rs 110 crore in a round led by Peak XV Partners.
Former Five Star Business Finance CEO Rangarajan Krishnan recently acquired a controlling stake in the 2018-founded lender.
Fund usage:
- Scale the lending platform
- Expand the team
- Deepen presence across key MSME markets in India.
Also Read: Brnd. me completes Singapore-to-India merger, eyes IPO in 12-18 months
Source link