How language LLMs will lead India's AI leap
The next great power struggle in technology won't be about speed or scale, it'll be about whose language AI speaks. Because trust in technology begins with something deeply human: being understood.
You trust a doctor who speaks your language. You trust a banker who understands your context. So why would you trust an algorithm that doesn't know who you are, where you're from, or what your words mean?
This question is being asked by governments, developers, and communities across the Global South who have seen how powerful large language models (LLMs) can be—and how irrelevant they often are to people who don't speak English or live in Silicon Valley.
In India, the response until now has been BharatGPT. This is a collaboration between startups like CoRover.ai, government-backed platforms like Bhashini, and academic institutions such as the IITs. Its aim is not to chase ChatGPT on global benchmarks. Instead, it hopes to solve problems at home—helping citizens navigate government forms in Hindi, automating railway queries in Tamil, or enabling voice assistants in other regional languages. CoRover has already deployed multilingual chatbots in sectors like railways, insurance, and banking. The value here isn't just in automation. It's in comprehension.
This isn't unique to India. In South Africa, Lelapa AI is working on InkubaLM, a small language model trained in African languages. In Latin America, a consortium is building LatAm GPT, rooted in Spanish, Portuguese, and indigenous dialects. Each of these projects is a rebellion: against invisibility, against standardization, against a worldview where the technology speaks only in one accent.
What's driving this shift? 'Current large language models do not adequately represent the linguistic, cultural, or civic realities of many regions,' says Shrinath V, a Bengaluru-based product coach and Google for Startups mentor. 'As governments begin exploring AI-powered delivery of public services, from education and legal aid to citizen support, they recognize the need for models that reflect local languages, data, and social context. Regional LLMs are being positioned to fill that gap,' he explains.
Manoj Menon, founder of the Singapore-based research firm Twimbit, is on the same page as Shrinath: 'With AI there are several nuances that come into play — how we train them to be contextually relevant for our local, national needs.'
At the heart of it lies something more political: digital sovereignty. Shrinath breaks it down and says, 'Data sovereignty is no longer an abstract idea. Countries don't want to depend on models trained on data they don't control. Indigenous models are a way to retain that control.'
It boils down to geopolitical leverage. Nations that build their own models won't just protect cultural identity—they'll shape trade, diplomacy, and security doctrines in the AI era. 'This is a reasonable argument,' says Menon. 'How we interpret a particular subject or issue depends completely on the context. Hence geo-politics is a significant input. Also the ability to train based on local issues and context.'
Viewed through this lens, the shift underway towards frugal AI is more radical than most people realise. These are models that don't need massive GPUs or high-speed internet. They're lean, nimble, and context-rich. Think of it like this: if ChatGPT is a Tesla on a six-lane highway, BharatGPT is a motorbike designed for rough, narrow roads. Not as flashy. But it gets where it needs to go.
'Most countries will want a say in shaping how AI is adopted, governed, and deployed within a sovereign context,' points out Shrinath. This matters because AI is starting to mediate access to public services—healthcare, legal advice, welfare. And in that context, a model that doesn't understand a citizen's language isn't just ineffective. It's dangerous. It can mislead, it can exclude and it can fail silently.
So yes, Silicon Valley still leads the headlines. But away from the noise, something deeper is unfolding. A shift in who gets to define intelligence, in whose language it speaks and in whose image it is built. Regional AI, says Menon, 'won't go head-on with what is built in Silicon Valley. They will complement it and their opportunity will help AI be more relevant locally.'
These regional AI efforts don't seek applause, they seek agency. They aren't chasing scale, they're chasing significance instead. This revolution is not being televised, it's being trained.
Get 360° coverage—from daily headlines to 100 year archives.
Hashtags
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
NDTV
2 hours ago
- NDTV
Billionaire Sanjay Kapur's Net Worth Makes Him Richer Than Shah Rukh Khan. Report
New Delhi: Karisma Kapoor's ex-husband, Sunjay Kapur, passed away on June 12, 2025. He suffered a heart attack during a polo match in England. The industrialist was 53. As per reports, the business tycoon's net worth makes him richer than Shah Rukh Khan and even the Kapoor family. The chairman of auto components maker Sona Comstar, Sunjay Kapur's net worth was estimated to be $1.2 billion (Rs 10,300 crore), as per a report by Forbes. He booked his spot in the prestigious World's Billionaire List in 2022. Sanjay Kapur's wealth makes him the richest person when compared to some prominent Bollywood celebrities. According to reports and as quoted on Hindustan Times, Shah Rukh Khan, considered one of the richest actors in the Hindi film industry, has a net worth of $ 880 million (Rs 7700 crore). Additionally, Sanjay Kapur's wealth is more than the entire extended Kapoor family, too. Their combined net worth which is around Rs 2000 crore which is just a fifth of what Sunjay owns. As for Karisma Kapoor, her net worth is reportedly Rs 120 crore. Sanjay Kapur, armed with a business degree from the University of Buckingham, entered the Sona Comstar business in 2015 after his father Surinder Kapur's death. Surinder Kapur launched Sona Comstar in 1995, which became a key player in the auto parts sector. Under his son's leadership, the company evolved into a global name in auto component manufacturing. 'Sona' translates to gold in Hindi and the name was inspired by his grandfather's old jewellery business. According to Forbes, Sona Comstar, headquartered in Gurugram, runs 12 factories with over 5,000 employees working for the organisation in India, China, Mexico, Serbia, and the US. The global rise in electric vehicle demand has boosted the company's growth. In the year ending March 2025, Sona Comstar earned 35.5 billion rupees (around $412 million), marking a 12% jump in revenue. EV components made up roughly 36% of that income. Sunjay Kapur's death was confirmed by Sona Comstar and the Automotive Component Manufacturers Association of India (ACMA). In a statement, they wrote, 'A visionary leader, Kapur played a pivotal role in shaping Sona Comstar into a global mobility technology company built on innovation, sustainability, and purpose.' 'Kapur, a respected leader and visionary, was a passionate advocate for the automotive components industry. Under his stewardship, ACMA witnessed significant progress in global competitiveness and EV readiness,' the statement added. Sunjay Kapur got married to Priya Sachdev, a former model and entrepreneur, after his divorce from Karisma Kapoor in 2016.
Time of India
2 hours ago
- Time of India
Scale AI may lose partnership with Microsoft, Google, ChatGPT-maker OpenAI and Elon Musk's xAI after Facebook's $14 billion investment
Google, the largest customer of AI data-labeling firm Scale AI, plans to end its relationship with the company following news that rival Facebook-parent company Meta is acquiring a 49% stake, a report said. Google had reportedly intended to pay Scale AI around $200 million this year for human-labeled training data, essential for developing advanced AI models like its ChatGPT competitor, Gemini. Citing five sources familiar with the matter, news agency Reuters reported that the move stems from concerns that Meta's significant investment could expose Google's proprietary AI research and roadmap to a key competitor. Reportedly, the search giant has already begun discussions with several of Scale AI's rivals this week, looking to shift this substantial workload. Not only Google, The unease isn't limited to Google. Other major tech companies using Scale AI's services, including Microsoft and Elon Musk's xAI, are also reportedly distancing themselves. The report also says that OpenAI, another significant AI developer, had already scaled back its engagement with Scale AI months ago, though its spending was considerably less than Google's. However, OpenAI's CFO recently stated that the company would continue working with Scale AI as one of its many data vendors. Why these companies are scaling back their involvement in Scale AI The primary concern among companies competing with Meta in cutting-edge AI model development is the risk of their research priorities and technical blueprints being exposed to a direct rival. Customers often share proprietary data and even prototype products with Scale AI's workers for data-labeling services. With Meta now holding a substantial stake, these AI firms fear their chief competitor could gain sensitive business and technical intelligence. What it may mean for Scale AI For Scale AI, this potential loss of Google's business comes as a blow, despite its recent valuation jump to $29 billion following Meta's investment (up from $14 billion pre-deal). Scale AI's core business heavily relies on a few major customers, making it vulnerable to such departures. While Scale AI intends to continue operations, its CEO, Alexandr Wang, along with some employees, are moving to Meta. The bulk of Scale AI's revenue comes from providing human trainers to annotate complex datasets for generative AI model makers, a service that can cost as much as $100 per annotation for sophisticated examples. Its largest revenue stream is from its partnerships with generative AI model makers.
Mint
3 hours ago
- Mint
Google Search is fading. The whole internet could go with it.
Experience a random pain in the 21st century and an internet search usually comes before a call to the doctor. Googling 'chest pain," 'high fever," or 'skin rash" calls up a series of blue links followed by a frenzied trip across the web. A similar pattern plays out, minus some anxiety, for 'today's weather," 'restaurants near me," and 'high-yielding dividend stocks." Roughly one in five visits to the world's top internet sites begin on search engines, according to data from analytics firm Semrush. At Wikipedia, search generates 63% of global visits. For travel site Tripadvisor, it's 58%; for local review site Yelp, it's 51%. But internet search traffic has been falling for much of the past year as web surfers experiment with artificial-intelligence-powered search from OpenAI's ChatGPT and AI start-up Perplexity AI. So far, referrals from AI search engines have replaced about 10% of the traditional search losses, according to Similarweb data. Google is pushing back by adding AI-powered summaries to the top of its search results, de-emphasizing its traditional blue links and thereby further reducing search traffic. May could prove to be a tipping point. Last month, search referrals to top U.S. travel and tourism sites tumbled 20% year over year, according to the latest data from Similarweb. E-commerce companies saw their referrals fall 9%. For news and media sites, search traffic dropped 17%. The finance, lifestyle, and food-and-drink categories all saw similar types of declines on the month. Across the web economy, the trend is clear: Search is drying up, and Google is no longer the clear-cut way to drive audiences to websites. The changes have begun to force a reckoning across various industries. Late last month, Business Insider, a leading digital news publication, cut 21% of its staff, citing traffic drops that were 'outside of its control." 'Business models are under pressure, distribution is unstable, and competition for attention is fiercer than ever," Business Insider CEO Barbara Peng wrote to employees. Reddit, the social-media site and source of answers to many random questions, which gets 57% of its visits from search, is making deals with AI firms and rolling out its own AI-driven search engine. Chegg, a homework-help company, worth $15.1 billion at its peak in 2021, said earlier this year that traffic declines had given it no choice but to explore strategic alternatives, including a possible sale. On Wall Street, no company has faced greater worries about the future of search than Google itself. Shares of parent Alphabet are down 7% on the year; the company now gets counted as a value stock in some investor benchmarks. But Google has countermeasures. For one, it has diversified itself into a cloud-computing giant, and it's a winner in the nascent category of autonomous driving. Google itself is also no slouch in the generative-AI world, with massive resources to build and improve its Gemini large-language models. Instead, as traditional search fades in importance, it's the rest of the internet that will suffer. In May, monthly U.S. search traffic to fell for the first time in at least two years, according to Similarweb, down 14%. A year ago, search referrals to Schwab were up 179%. TripAdvisor's search tumbled 34% on the month, while Starbucks saw a 41% decline to its website. Search to Netflix, a pioneer in digital strategies, was down 23%. The traffic conversation has the feel of the 1990s and early aughts before Google arrived and companies were still trying to figure out how to attract audiences across the World Wide Web. Executives are talking up deals with OpenAI's ChatGPT, Perplexity, and other AI-driven search tools. 'We're partnering with AI search companies to ensure our brands show up well across customer queries," said Expedia CEO Ariane Gorin in May, 'and building new experiences to connect with travelers outside our ecosystem." There's a long way to go. Based on Similarweb's U.S. estimates, Expedia got 88,000 referrals from AI search engines in May. It got 34 million referrals from search. ChatGPT was something of a novelty when the model made its public debut in November 2022, generating a wave of songs, poems, and essays across the web. But the latest models, which are more sophisticated and promise humanlike reasoning, have spurred a surge of new use. ChatGPT had 500 million weekly active users in March, rising from 300 million in December. Many of them pay $20 a month for service; parent company OpenAI says it reached an annualized revenue run rate of $10 billion this month, up from $5.5 billion at year end. Another start-up, Perplexity, has taken on Google more directly. 'A direct line to the world's knowledge—compressed, cited, and made clear," Perplexity says on its about page. 'No gimmicks. No fluff. Just answers that make sense." (Barron's owner Dow Jones has sued Perplexity for copyright infringement.) As AI pressure mounts on Google, the company has moved to defend its 89% U.S. market share in search. A year ago, it launched so-called AI Overviews atop Google search results, promising condensed AI-generated answers to search queries. Those overviews, which initially appeared on a small number of searches, have been appearing more frequently. An analysis by research firm Ahrefs said that the prevalence of AI Overviews have more than doubled from March 12 to May 6. The AI summaries have spurred debate across the internet, with publishers worried about a search query that delivers answers in a few paragraphs, with no need to click for more info. According to Similarweb data from March, searches with AI Overviews resulted in a click 23% of the time. For searches without the overviews, the click rate was 36%. 'Looking at search results that do show an AI answer, comparing that with search results that do not show an answer, we found a crazy drop-off," said Kevin Indig, a search-engine optimization, or SEO, consultant and author of the Growth Memo blog. 'This is a click killer." Google told Barron's that third-party data offer an incomplete picture of search trends. In February, online education platform Chegg said search trends had crushed its business. CEO Nathan Schultz told investors: 'We would not need to review strategic alternatives if Google hadn't launched AI Overviews, retaining traffic that historically had come to Chegg, materially impacting our acquisitions, revenue, and employees." Asked for comment, the company directed Barron's to a lawsuit it filed in February against Google. It alleges that Google is using its search dominance to 'coerce online publishers like Chegg to supply content that Google republishes without permission in AI-generated answers that unfairly compete for the attention of users on the internet." Chegg shares have tumbled 99% since 2021. Google says its AI Overviews have improved the search experience and are being embraced by users. A Google spokesperson says that AI Overviews show more links to a wider range of sources on results pages. 'More than any other company, Google prioritizes sending traffic to the web, and we continue to send billions of clicks to websites every day," the Google spokesperson told Barron's. In April, during Google's earnings call, an analyst asked company executives about the impact that AI Overviews was having on click-through rates. 'I don't think this is the moment to go into the details of click-through rates and conversion and so on," said Philipp Schindler, Google's chief business officer. 'But overall, we're happy with what we're seeing." Reddit has become a battle ground and flashpoint in the argument about search's future. The stock is down 28% so far this year as investors worry about slowing user growth on the social-media site. It's a trend the company has attributed to an evolution in search. The company remains a standout in search, and it points out that 'Reddit" is the No. 6 searched term on Google. Still, traffic trends have notably shifted in recent months. In May 2024, Reddit's search referrals soared 78%, according to Similarweb. This past May, searches to the site were up 14%. Reddit's daily user growth, meanwhile, has gone from 37% in the first quarter of 2024 to 31% in 2025. Reddit Chief Operating Officer Jennifer Wong said in an interview with Barron's that search is under 'heavy construction." Wong is confident about the long-term opportunity for Reddit, noting that its human-generated content will be especially sought after to train the large language models that run AI. Reddit has a deal with OpenAI. For now, that kind of licensing is a small part of the business, accounting for less than 9% of revenue in the most recent quarter. 'Nobody knows how Google is going to cross this canyon and the kind of ripple effects that they'll have across the internet," Wong says. 'What I do know is that I think human intelligence is still going to be worth a lot, and it's going to go up and up in value, and I think Reddit is the place for that." Wall Street generally agrees. Across 29 analysts covering Reddit, the average price target is $152, 31% above its recent close. Going forward, investors should pay attention to a company's search exposure. Across the categories, certain brands are far less dependent on Google's referrals. Airbnb, for instance, got 14% of referrals from search in March, versus the 58% for travel firm Tripadvisor, according to Semrush. DoorDash and Uber Technologies were 13%. Social-media apps like Pinterest and Meta Platform's Instagram also tend to be far less search-dependent than much of the internet. Their sites drew 23% and 17% of referrals from search, respectively. In May, Pinterest's CEO told investors that 85% of users come directly to the company's mobile app. Meta Platforms, which essentially shares the online advertising pie with Google, is in an enviable position. As search traffic falls and traditional search advertising wanes, businesses will be compelled to advertise on Instagram and Facebook, Meta's social networks that are insulated from search disruption. Meanwhile, Meta can use AI to improve ad effectiveness and personalization. As Google defends itself against AI, Meta is free to fully embrace it. Ultimately, the best hedge against AI disruption is the company empowering it all: Nvidia. As AI explodes, Nvidia will sell more AI chips and the infrastructure to power data centers. Newer reasoning-focused AI models are particularly profitable for Nvidia because they require 100 times the computing resources compared with prior AI chatbots, Nvidia CEO Jensen Huang told investors on Nvidia's recent earnings call. 'Reasoning models are driving a step-function surge in inference demand," he said. As Google Search comes under threat for the first time in two decades, a U.S. district judge is determining the company's fate. Judge Amit Mehta ruled in August 2024 that Google was a monopoly in general search services and general text advertising. Alphabet and its investors are now waiting for the judge to determine potential remedies. A penalty phase of the trial concluded last month, and Mehta is expected to issue a final decision in August. The timing is notable. Alphabet was ruled a monopoly around the same time that it was pushing out AI Overviews. Less than a year later, search is facing a massive competitive shift as users increasingly embrace chatbots. Apple's senior vice president of services, Eddy Cue, testified that searches in Apple browsers fell for the first time in April as people increasingly turned to AI for search queries. Google pays billions of dollars to Apple annually to make its search engine the default option on Apple browsers. But don't expect Google's recent weakness to affect Mehta's decision. Antony Haynes, partner at Dorf Nelson & Zauderer, told Barron's that even though these new competitive threats have become more prevalent since the judge made his ruling, it's unlikely those threats will affect his decisions regarding remedies. 'We're not thinking about a remedy for potential future technology changes. We're looking at what they did in the past," Haynes said. That past includes one of the best business models ever created. Last year, Alphabet's operating margin was 32% versus an aggregate 14% for S&P 500 index companies. Meanwhile, since Alphabet went public in 2004, the stock has returned an annualized 23.7%, versus 10.6% for the S&P 500. A post-search world will probably weigh on those margins, but Alphabet stock already reflects the next phase. It trades at 17.8 times expected earnings for the next 12 months, below the S&P 500's 22.5 multiple. Barron's has remained bullish on the stock, including in a November 2024 cover story and a follow-up last month. Google will be fine. It's the rest of the internet that should be worried. Write to Adam Levine at Tae Kim at and Angela Palumbo at
