Netweb is betting on AI infrastructure, with new initiatives spanning agent security, GPU optimisation, and sovereign computing.Read more →
Nov 12, 2026
Markets can do little more than speculate about possible outcomes until the investigations yield conclusive results. But what if Huang is ousted as CEO?Read more →
Nov 11, 2026
Coforge CAICO Anup Nair says the IT services firm is skipping a Palantir-style platform and betting on context, trained engineers and repeatable projects.Read more →
Nov 11, 2026
AI agents today—the ones that can use your web browser and autonomously perform tasks—are getting more intuitive. But are they safe enough?Read more →
Nov 11, 2026
The new infrastructure is designed to help organisations deploy and scale AI and HPC workloads across on-premises data centres, hybrid environments, and colocation facilities.
While 60% of Indian organisations report positive returns on AI investments, declining security readiness and infrastructure gaps are creating challenges to scaling AI.
The move could help financial services firms and other regulated businesses meet data localisation requirements while deploying OpenAI applications.
The offering runs on PTC's Windchill platform and covers IP management, supplier collaboration, product introduction, quality, and regulatory compliance.
IBM and Microsoft have launched Imagine Enterprise, an agentic AI platform combining agents with enterprise cloud, data, and security technologies.
4 min readThe health AI company’s Parrotlet-A 2.5 Pro speech model and Lipi agentic harness convert doctor-patient consultations into structured digital health records.
4 min readThe Japanese investment giant has written off billions of dollars after changing its plans for US data centre infrastructure, while also reporting losses linked to its OpenAI deal.
4 min read
Supreeth Koundinya
IBM had historically used high-skilled visa programmes to address talent shortages, but its participation had been declining for several years.
3 min readThe Series A funding will support engineering and enterprise expansion as the startup launches PuppyGraph 1.0 graph query engine.
3 min readThe expansion will place millions more teenagers under Instagram's Teen Account restrictions, even if they registered as adults.
4 min readAs AI changes engineering delivery, larger IT services firms are buying specialist capabilities. The challenge is turning those acquisitions into new business.
4 min readThe Power plan is listed at ₹1,799 and the Maximum tier at ₹8,999, although Meta has not confirmed an India launch date or the prices.
4 min read
Supreeth Koundinya
TPU 8tx is built for confidential AI inference, allowing enterprises and governments to process sensitive data.
3 min readAI-led contracts accounted for 66% of quarterly deal wins, while utilisation fell to 82% and operating cash flow turned negative.
3 min readThe National Stock Exchange has introduced natural language stock screening, AI-assisted corporate announcement summaries, and market data access.
4 min readSalesforce is positioning Slack as a work operating system for enterprises, with Slackbot bringing together organisational data, applications, and AI agents.
4 min readThe NVIDIA CEO’s admission has triggered a board-level probe into allegations of Chinese intelligence involvement and potential risks to sensitive company information.
4 min read
Mukundan Sivaraj
LTM will use its internal deployment of 1,500 ServiceNow AI agents as a reference for customer engagements.
3 min readThe growing adoption of enterprise AI is creating a different set of requirements for technology partnerships.
3 min readThe stories sparking conversations this week. Join the discussion.
Build your routine with some of our top newsletters or view them all here.
Make sense of the day's AI news and breakthroughs with our morning briefing.
Receive a roundup of AI adoption stories by industry vertical, curated for professionals.
By signing up, you agree to our Privacy Policy.
The model scored 71.8 across three benchmarks, recording a 44 ms response time on an NVIDIA B200 GPU.
The company will offer affected employees at least six months’ salary and consolidate more product teams in Sydney as part of an organisational restructuring.
The models support multiple Indian languages and target applications ranging from agentic AI and banking to secure, offline deployment across critical sectors.
Coforge’s 53.6% Q2 revenue jump includes Encora. The company is moving from selling engineering hours to selling outcomes.
"Self-hosting GLM brings our blended cost of tokens to 30 cents per million. Our invoice is now one-tenth of what we were spending on Anthropic and OpenAI,” said David Abitbol, who leads AI operation
Enmovil's expansion addresses the need for AI across complex supply chains, where solutions must fit unique customer workflows and operational limits.
The platform provides a single API to access open AI models, including NVIDIA Nemotron, Sarvam, DeepSeek and Qwen, through NVIDIA GPU infrastructure hosted in India.
The platform has not only added 4 million CPU cores and 160 PB of internal storage, but it is also rebuilding core Git primitives by decoupling compute from storage.
CFO Kevan Parekh also confirmed that India contributed to Apple’s best fiscal year for iPhone upgrades.
4 min readThe divide is visible in the difference between the resources available to urban and rural students.
4 min readThe company will use the capital to expand its AI team and develop its Large Logistics Spatio-Temporal Model, which it uses to improve delivery allocation and reduce idle time for riders.
4 min read
Mukundan Sivaraj
The offering combines specialised agents and conversational assistants for system administration, observability and optimisation, with human oversight.
3 min readBy modifying just 0.016% of the model’s weights, Astra increased its score on the MATH-500 benchmark from 7.9% to 48.1%.
3 min readNeuroDX’s Sphota claims 50x greater training data efficiency and 4.5x better memory efficiency than a comparable Liquid AI model.
4 min readThe prototype generated its first AI response using an onboard TPU, as Google prepares to test the hardware against radiation and thermal stress.
4 min readThe certification recognises how data science teams work across the company and global client programmes.
4 min read
Ridhika Basnet
The text-to-speech model powers more than one million calls daily and delivers approximately 30% better pronunciation accuracy on selected terms, according to the company.
3 min readIndian engineers found that replicating San Francisco’s model wasn’t enough to scale. Success required a deeper understanding of user behaviour and engagement, a gap Suhas Motwani is working to bridge.
3 min read
The company has launched Bivia, a multimodal decision model designed to help enterprise AI agents make domain-specific decisions and verify evidence across different data formats.
Industry leaders say the focus is moving from building increasingly capable models to enterprise readiness, governance, reliable execution and measurable returns.
Nemotron 3.5 Transcribe supports speaker attribution, speech translation and custom transcript formats. NVIDIA reports a 4.71% mean word error rate across nine English benchmark test sets.
Legacy systems, fragmented data and infrastructure gaps could slow Indian enterprises as they move AI agents from pilots into production.
Unlimited access to premium articles, exclusive analysis, and the stories shaping India's AI economy.
Excellence in AI Journalism
AI Education
AIM Americas
1190 Miraloma Way Suite P, Sunnyvale, CA 94085
AIM India
1st Floor, Sakti Statesman, Bellandur, Bengaluru – 560103
Contact
[email protected]© 2026 AIM. All rights reserved.
