Latest news with #SarvamM
Time of India
27-05-2025
- Business
- Time of India
RBI's new directive to NBFCs; Oyo's third IPO attempt
RBI's new directive to NBFCs; Oyo's third IPO attempt Also in the letter: RBI tightens default loss guarantee rule; NBFCs to exclude cover on fintech-sourced loans What's changing? Strengthen underwriting practices. Curb systemic risks. Avoid over-reliance on fintechs. Who's affected? MobiKwik Paytm Moneyview When's the deadline? The impact: NBFCs: Higher provisions, leading to reduced appetite for fintech-originated credit. Higher provisions, leading to reduced appetite for fintech-originated credit. Fintechs: Likely decline in originations and income. Likely decline in originations and income. Borrowers: Stricter access to unsecured credit. Oyo to meet bankers next week for third IPO attempt; eyes $5-7 billion valuation Verbatim: More details: The company is in talks with Indian and global banks. It aims to file its draft red herring prospectus (DRHP) between August and September, with a public listing targeted for March or April 2026, according to sources. This will be Oyo's third attempt at going public. It first filed with the Securities and Exchange Board of India in 2021, aiming for an Rs 8,430 crore IPO, but withdrew in 2022. A second filing, made confidentially in 2023, was also withdrawn in 2024. Recent fundraise: Sponsor ETtech Top 5 & Morning Dispatch! Why it matters: The opportunity: Reach a highly engaged audience of decision-makers. Boost your brand's visibility among the tech-savvy community. Custom sponsorship options to align with your brand's goals. What's next: Info Edge Q4 revenue rises 14% to Rs 750 crore; net profit surges 8x to Rs 678 crore Financials: Operating revenue (Q4): Rs 750 crore. Rs 750 crore. Net profit (Q4): Rs 678 crore, up from Rs 88 crore a year ago. Rs 678 crore, up from Rs 88 crore a year ago. Revenue (FY25): Rs 2,849 crore, up 12% YoY. Rs 2,849 crore, up 12% YoY. Net profit (FY25): Rs 1,310 crore, more than double the Rs 594 crore reported in FY24. Rs 1,310 crore, more than double the Rs 594 crore reported in FY24. Total expenses (FY25): Rs 539 crore, compared to Rs 469 crore in FY24. Employee benefits: Rs 331 crore. Advertising and promotion costs: Rs 100 crore. Rs 539 crore, compared to Rs 469 crore in FY24. Revenue breakdown: Recruitment solutions (including : Rs 542 crore, up 13% YoY. : Rs 542 crore, up 13% YoY. 99Acres (real estate portal): Rs 106 crore, up 14%. Rs 106 crore, up 14%. Other businesses (including and : Rs 101 crore, up nearly 20%. Also Read: Sarvam AI unveils multilingual LLM; low traction poses questions on India's AI scene About the model: The model supports a hybrid reasoning mode for tackling complex logical reasoning problems, as well as mathematical and coding tasks. Additionally, it features a non-think mode for general-purpose conversation. According to the company, Sarvam M outperforms similarly sized models on coding and math benchmarks. No traction: Frinks AI raises $5.4 million: Contineu raises $1.2 million: Technology workforce in Bengaluru crosses one-million mark; IT city among 12 global tech hubs: CBRE Details: The city ranks fourth among the 12 tech markets in terms of the share of its working-age population. 75% of Bengaluru's falls in this productive age group. Between 2019 and 2024, Bengaluru saw a 2.4% increase in its working population. The city also leads in terms of AI development talent. In 2024, Bengaluru attracted 140 venture capital (VC) deals worth $3.3 billion. The RBI has tightened rules on default loss guarantees (DLGs), a move that is likely to hit digital lenders hard. This and more in today's ETtech Top 5.■ Info Edge Q4 results■ Sarvam's LLM gets muted response■ Bengaluru tech workforce crosses 1 millionThe Reserve Bank of India (RBI) has directed non-banking finance companies (NBFCs) to exclude default loss guarantees (DLGs) provided by fintech partners when provisioning for bad will no longer factor in the typical 5% guarantee from digital lending partners to reduce provisioning on stressed loans. This marks a significant shift that could dent both origination volumes and fee income for RBI wants NBFCs to:The move follows cases where fintechs failed to honour DLGs, leaving NBFCs exposed to lending partners such as:These firms act as lending service providers and typically offer DLGs of up to 5%, often backed by fixed deposits lien-marked in favour of NBFCs. These guarantees act as credit cushions and are usually factored into expected credit losses (ECL) must comply by September 30, treating fintech-originated loans as if there is no credit enhancement. Some NBFCs have already begun adjusting provisions from Q4 directive is part of the RBI's broader crackdown on hidden risks in India's rapidly growing digital lending ecosystem. It also serves as a nudge for NBFCs to shoulder the risk they underwrite, instead of outsourcing Agarwal, CEO, OyoOravel Stays Ltd, the parent company of hospitality startup Oyo, is set to formally review proposals from merchant bankers next week as it prepares for a fresh attempt at an initial public offering (IPO), according to multiple people familiar with the development.'During the preliminary discussions, some bankers proposed that the company could get valued as high as $10 billion for its public issue, but the company's realistic expectation is around $6-7 billion,' a source August, Oyo raised Rs 1,457 crore from a group of investors at a significantly reduced valuation. In December, it secured $825 million in debt from Deutsche Bank to fund its $525 million acquisition of US motel chain Motel Top 5 and Morning Dispatch are must-reads for India's tech and business leaders, including startup founders, investors, policy makers, industry insiders and Reach out to us at spotlightpartner@ to explore sponsorship Bikhchandani, cofounder, Info parent Info Edge posted a 14% year-on-year (YoY) increase in operating revenue for the March quarter, driven by robust growth across both recruitment and non-recruitment AI startup Sarvam AI, the first company chosen by the government to build a homegrown foundational model, has launched its open-source large language model (LLM) . However, the early reception has been muted, with only a few hundred downloads in the initial claims its LLM, Sarvam M, performs well on benchmarks in mathematics, programming, and 11 Indian languages, including Hindi, Gujarati, Kannada, and model, released on Hugging Face, recorded just over 300 downloads at launch. As of May 27, the number had increased to 1,200. The lukewarm response has reignited debate over India's place in the global AI race, particularly with rivals such as DeepSeek and AI has raised $5.4 million in a new funding round led by Prime Venture Partners. Founded by IIT Hyderabad alumni Aditya Agrawal, Dharmgya Sharma, and Subhra S Bhattacherjee, Frinks AI is a deep-tech startup developing next-generation vision AI systems for industrial automation and quality startup Contineu has raised $1.2 million in a seed funding round led by SenseAI Ventures, with Piper Serica Angel Fund participating. Founded in 2023, the startup automates data entry on construction sites through its platform, utilising helmet-mounted cameras and 3D computer vision is now among the top 12 global technology hubs, joining the ranks of Beijing, Boston, London, New York, and Toronto, as its technology workforce has crossed the one million to a report by real estate consultant CBRE, Bengaluru's tech talent scale rivals that of the US hubs of San Francisco and New York.
Time of India
27-05-2025
- Business
- Time of India
Sarvam AI unveils multilingual LLM; tepid response poses questions over India's AI chops
Indian artificial intelligence (AI) startup Sarvam AI , the first company picked by the government to build a sovereign foundational model under the India AI Mission , dropped its multilingual open-source large language model (LLM) Sarvam M last week. It claims the LLM has achieved benchmarks across mathematics, programming, and a range of Indian languages. But the launch has thrown up questions about the country's AI scene, with the model getting only a few hundred downloads in the first two days. Let's take a deep dive to understand why. About the model Sarvam M is a 24-billion-parameter model, built on top of French AI company Mistral's model Small. The startup said the model is "built for versatility" and designed to support a wide range of applications, including conversational agents, translation and educational tools. It supports 10 Indian languages, including Hindi, Bengali, Gujarati, Kannada and Malayalam. Live Events — SarvamAI (@SarvamAI) "This is the first in a series of contributions as we help build out the Sovereign AI Ecosystem in India," the startup said in a post on X. Discover the stories of your interest Blockchain 5 Stories Cyber-safety 7 Stories Fintech 9 Stories E-comm 9 Stories ML 8 Stories Edtech 6 Stories Key features Hybrid thinking mode: This mode supports complex logical reasoning, solves mathematical problems and handles coding tasks. It also has a 'non-think' mode for general purpose conversation. Indic skills: It is post-trained in Indian languages along with English to better suit local needs. Reasoning capabilities: The company claims Sarvam M outperforms similarly sized models on coding and math benchmarks. No traction In the first two days after the launch, the model had just over 300 downloads. As of May 27, it had 1,200 downloads on Hugging Face, can be tested on Sarvam AI's playground, and accessed through its APIs. This sparked a debate on social media platforms about the standing of India's AI scene when compared with global rivals OpenAI and China's DeepSeek. Social media backlash In a scathing post, Deedy Das, a venture capitalist at Menlo Ventures, hit out at "India's biggest AI startup $1B Sarvam" for a muted launch with only 23 downloads in 2 days. "In contrast, 2 Korean college trained an open-source model that did ~200K last month (sic)," his post on X read. — deedydas (@deedydas) Das added in another post that he had nothing against Sarvam or India trying to build AI, but was "disappointed at their direction". Many came to the defence of the model. "Before Sarvam M came out, people were complaining about the lack of IndicLLMs. After Sarvam M came out, people are still complaining," one user wrote. Another person said we must offer encouragement, not shame entrepreneurs. "Teams in India are trying and they are working to build the muscle and culture of what good or amazing looks like." — svembu (@svembu) Sridhar Vembu of Zoho also jumped to the defence of Sarvam, saying that there was no such thing as an "instant hit". "Even when we were the first mover in a new market and we had done a lot of technical work, we only got slow traction. Instant success is neither necessary nor sufficient to succeed long term," he said in a post on X, telling the Sarvam team to "fight the good fight". Sarvam's trajectory The Bengaluru-based startup was the first to be selected to build an indigenous foundational model under the Rs 10,000 crore IndiaAI Mission. It was provided access to 4,096 Nvidia H100 graphics processing units (GPU) for six months from the mission's common compute cluster to train its model. According to Tracxn, the company has so far raised $53.6 million in total and is valued at $111 million.
Entrepreneur
26-05-2025
- Business
- Entrepreneur
Sarvam AI Launches 24B Parameter Open-Source LLM for Indian Languages and Reasoning Tasks
This release follows Sarvam's selection by the Indian government to build a sovereign LLM under the IndiaAI Mission, marking the first step in strengthening the country's domestic AI capabilities You're reading Entrepreneur India, an international franchise of Entrepreneur Media. Bengaluru-based AI startup Sarvam AI has introduced its flagship large language model (LLM), Sarvam-M, a 24-billion-parameter open-weights hybrid model built on Mistral Small. Designed with a focus on Indian languages and advanced reasoning capabilities, Sarvam-M is intended to power applications such as conversational agents, machine translation, and educational tools. According to Sarvam, the model has been fine-tuned through a three-step process involving Supervised Fine-Tuning (SFT), Reinforcement Learning with Verifiable Rewards (RLVR), and inference optimisation. The SFT process involved crafting diverse, high-quality prompts to train the model in both general dialogue and complex reasoning. RLVR further improved its instruction-following and mathematical capabilities using custom reward engineering and curated datasets. Inference was optimised using FP8 post-training quantisation and techniques like lookahead decoding. Sarvam-M has demonstrated strong performance in multilingual and reasoning benchmarks. It achieved a significant 86 per cent gain on a romanised Indian language version of the GSM-8K math dataset and showed average performance boosts of 20 per cent on Indian language benchmarks, 21.6 per cent on math, and 17.6 per cent on programming tasks. It outperforms Llama-4 Scout and is on par with models like Llama-3.3 70B and Gemma 3 27B, though it slightly lags in English benchmarks such as MMLU. The model is accessible via Sarvam's API, its own playground, and available for download on Hugging Face. This release follows Sarvam's selection by the Indian government to build a sovereign LLM under the IndiaAI Mission, marking the first step in strengthening the country's domestic AI capabilities.
