Рубрика: AI

Abhishek Singh, MeitY additional secretary and CEO of IndiaAI Mission, has warned that India’s tech and IT services industry could face serious trouble if engineering talent is not upgraded fast for the age of artificial intelligence.

Speaking at the Bengaluru Tech Summit on November 19, Singh said the rise of AI coding tools poses a direct threat to the country’s long-held advantage in software services.

“The ability to use brain power of Indians to solve world problems that has led to the boom in the IT industry… is facing a challenge from… AI code generators and the AI tools that OpenAI, Anthropic, and Google are making. If we don’t turbocharge our engineers with those AI skills, we run a huge risk and we’ll have a lot to lose,” he said.

He pointed out that India built its reputation as the tech garage of the world, yet the skills needed today are shifting fast. AI, data science, and advanced computing have become essential for the next leap in global technology. A slow response from companies could leave them exposed.

The IndiaAI Mission has started fellowships for students working on AI across fields like engineering, medicine, law, and liberal arts. Singh said data labs are coming up in partnership with states and industry to train data annotators, data analysts, and data scientists in tier 2 cities.

MeitY under the IndiaAI Mission, has also launched ‘YUVA AI for ALL’, a first-of-its-kind free course that introduces the world of AI to all Indians, especially the youth.

The mission is also creating tools focused on AI safety. These include systems for bias testing, ethical certification, deepfake detection, and stress testing, which will sit inside the AIKosh platform.

Singh said India’s long term tech strength depends on how fast companies lift the capability of their engineering teams. If the upskilling does not happen at speed, the country could lose ground despite new investments in compute infrastructure and model development.

India’s own push to build large language models is also moving ahead. Singh said Bengaluru based Sarvam is closing in on the launch of its foundation model. Sarvam is one of twelve foundation model projects supported under the IndiaAI Mission, where the government pays for all compute needed.

The post Indian IT Faces Threat From AI Coding Tools from OpenAI & Google, Says IndiaAI CEO appeared first on Analytics India Magazine.

When three entrepreneurs building India’s own foundation models under the IndiaAI Mission came together for a panel discussion, the venue at the Bengaluru Tech Summit 2025 had to be obviously packed. The audience was eager to listen to what they had to say about the country’s AI ecosystem.

Moderated by Kalika Bali from Microsoft Research, the session included Vivek Raghavan from Sarvam AI, Sashikumar Ganesan of Zenteiq, and Ananth Nagaraj of gnani.ai. They discussed why India cannot depend on global systems and must build its own base models, and what this means for the next 800 million Indians who don’t live inside Bengaluru’s ring roads.

The three startups, among the 12 empanelled under the IndiaAI Mission, have all taken up different verticals for building foundational models, ranging from a linguistic focused model to voice and material discovery models.

Raghavan of Sarvam AI arrived at the point directly. “If we don’t put an effort in foundational models, we will become a digital colony,” he said. He explained the risks of depending on open models whose origins and training data are unclear.

He warned that even an open model “can actually be poisoned with a very small amount of data,” and pointed out that many top-performing open models today come from China, like DeepSeek. The implication was obvious. For a country the size of India, relying blindly on such systems made no sense.

He also mentioned figures to back his argument. “Certain early open models had maybe sub 1% Indian data,” he said. Sarvam’s upcoming model, built under the IndiaAI Mission, will be a 120 billion parameter system with 17 trillion tokens, around 15-20% from Indian sources.

Raghavan said that distillation and domain-specific SLMs will become normal as applications scale for agentic AI.

The Missing Link

Sashikumar Ganesan of Zenteiq widened the discussion with his calm approach. “If you do not know how to build the foundational models, then definitely in the next wave we will be behind,” he said. According to him, India missed the supercomputing wave and paid the price. Missing the foundation model wave would be another generational loss.

ZenteiQ.ai, formerly Zentech AI, is building BrahmAI, a scientific foundational model for engineering intelligence, scientific computing, and industrial innovation. Its approach differs from typical LLMs: the model will understand and validate physics-based scientific questions.

“Our focus is Industry 5.0—applications in aerospace, automotive, EVs, energy, and pharma. India relies on foreign software for industrial R&D. BrahmAI is about sovereign scientific AI,” Ganesan had earlier told AIM.

Initial phases aim for 35 billion parameters, eventually scaling to 80 billion depending on compute availability. ZenteiQ has been allotted 2,128 H200 GPUs for the first year, but wishes to scale it further after that.

Ananth Nagaraj of gnani.ai pulled the conversation away from the labs and into a village 200 kilometers from Bangalore. “AI is a necessity for the next 800 million people,” he said. The internet in his village is still used for WhatsApp and YouTube, but the real potential is elsewhere.

“Unless we control the whole tech stack,” he said, “we are prone to all sorts of attacks.” His concern was beyond inclusion, about security, and sovereignty in the most literal sense. And it was about building tech that solves problems for people who don’t have the luxury of English, typing, or even silence around them.

Gnani.ai is developing a 14-billion-parameter multilingual voice AI model with 1.3 crore GPU hours.

Read: India’s AI Push Might Be Pointless Without National Language Standardisation

India’s Unique Problems

Nagaraj explained the everyday problems with a clarity that cut through the noise. Indian conversations are noisy, code-switched, full of background chaos. People talk from speakerphones in buses, fields and markets. “We handle close to 10 crore voice calls. One lakh audio calls every second,” he said.

They need to give a response in 150 milliseconds. It is a different universe. Western benchmarks simply don’t apply.

This is why Indic foundation models must be voice-first and robust enough for railway stations, farms and government offices. Bali reminded the audience of a previous ASR deployment that collapsed instantly because the model had never seen noise like Indian railway stations. Ananth nodded. This wasn’t a theory. This was daily reality.

Ganesan broke down how scientific foundation models work and why they can’t rely on standard transformers. “You cannot mix match. The next operator is not a probabilistic operator,” he said. Physical laws matter. Equations matter. Encoders must understand scientific structure, not only tokens.

India needs such models for materials, energy, climate and manufacturing. He said India’s biggest challenges — energy, EVs, climate threats, materials — won’t be solved by generic chatbots. They need scientific reasoning, not autocomplete.

Raghavan also added that use cases matter, but only platforms change the country. India Stack succeeded not because it solved one problem, but because it created rails for the future. AI, if built right, can play the same role. “AI is an accelerant,” he said.

He warned about the AI divide being even worse than the digital divide.

The only way to avoid that future is to make sure every citizen gets access to it. Not a few thousand engineers. Not a few million users. Everyone.

The post India’s AI Mission Isn’t About Chatbots. It’s About 800 Million People. appeared first on Analytics India Magazine.

Luma AI, a leading frontier AI company focused on multimodal general intelligence (AGI), has raised $900 million in a Series C funding round led by HUMAIN, a public investment fund (PIF) that delivers global full-stack AI solutions, according to the release. AMD Ventures, Andreessen Horowitz, Amplify Partners, and Matrix Partners also participated in the round.

The funding will support Luma AI’s partnership with HUMAIN to become a customer of Project Halo, a 2-gigawatt AI supercluster in Saudi Arabia, one of the largest compute infrastructure buildouts in the world.

The supercluster will enable Luma AI to train and deploy next-generation AI systems capable of understanding and operating in the physical world, going beyond large language models (LLMs) to learn from video, audio, and language data at an unprecedented scale.

“HUMAIN is the perfect partner for this next stage in Luma AI’s explosive trajectory,” said Amit Jain, CEO and co-founder of Luma AI.

Jain explained that to create AI that can help humanity in the physical world and expand understanding of the universe, it is necessary to build systems that can learn from a quadrillion tokens of information, roughly the collective digital memory of humanity, contained in video, image, audio, and language.

HUMAIN is deploying frontier compute infrastructure at impressive speed, and this is critical to achieving Luma AI’s mission.

To this, Tareq Amin, CEO of HUMAIN, mentioned that “Our investment in Luma AI, combined with HUMAIN’s 2GW supercluster, positions us to train, deploy, and scale multimodal intelligence at a frontier level. This partnership sets a new benchmark for how capital, compute, and capability come together.”

The supercluster will support Luma AI in training peta-scale multimodal data, 1,000 to 10,000 times more information than current frontier LLMs, making AI more applicable for real-world tasks. It will also feature next-generation inference systems capable of serving these models globally in real-time.

Luma AI’s flagship model, Ray3, has already demonstrated the company’s ability to transform foundational research into commercial products, deployed across studios, advertising agencies, and brands, including integration within Adobe’s global products.

With this new round, Luma AI plans to expand into simulation, design, and robotics while maintaining leadership in entertainment and advertising.

The company was also the first to launch models within HUMAIN Create, a regional initiative for building sovereign AI models tailored for the Arabic world.

These models are designed to understand cultural context, visual nuance, and linguistic diversity, enabling creators, enterprises, and governments to adopt AI solutions that reflect their identity, values, and sovereignty.

The post Luma AI Raises $900 Million Series C to Power 2GW AI Supercluster in Saudi Arabia appeared first on Analytics India Magazine.

OpenAI has released GPT-5.1-Codex-Max, a new agentic coding model designed for long-running software development tasks, and made it available across all Codex surfaces.

The model is built on an updated reasoning foundation and is trained on agentic tasks across software engineering, math, research, and more, according to OpenAI. It is the company’s first system trained to operate across multiple context windows through a process called compaction, enabling it to maintain coherence over millions of tokens during a single task.

OpenAI said the model can independently run for hours, adding that internal tests saw Codex-Max “persistently iterate on its implementation, fix test failures, and ultimately deliver a successful result” over tasks that ran for more than 24 hours.

The new model is accessible to users on ChatGPT Plus, Pro, Business, Edu, and Enterprise plans. Developers using Codex CLI via API key will receive access when API support rolls out. GPT-5.1-Codex-Max will now replace GPT-5.1-Codex as the default in all Codex interfaces.

OpenAI said 95% of its internal engineering team uses Codex weekly and that engineers “ship roughly 70% more pull requests since adopting Codex.”

Higher accuracy and better token efficiency

GPT-5.1-Codex-Max outperforms previous versions on several real-world and benchmark coding evaluations. On SWE-Lancer, it reached 79.9% accuracy, compared with 66.3% for GPT-5.1-Codex. On SWE-bench Verified, Codex-Max achieved higher accuracy at the same reasoning level while using 30% fewer thinking tokens.

OpenAI said the efficiency gains translate to lower costs for developers. In one example, the model generated a full browser-based CartPole reinforcement learning sandbox requiring 27,000 thinking tokens, compared to 37,000 for the earlier Codex model.

The company is also introducing a new extra-high reasoning mode for non-latency-sensitive tasks, which allows the model to think longer before producing output.

Long-horizon work and Windows support

Because of compaction, GPT-5.1-Codex-Max can handle complex refactors, multi-hour debugging, and extended agent loops that previously failed due to context limits. It is also the first Codex model trained to operate inside Windows environments. The system now includes tasks specifically designed to improve collaboration inside the Codex CLI.

Safeguards and cybersecurity

OpenAI said GPT-5.1-Codex-Max “does not reach High capability on Cybersecurity” under its Preparedness Framework, but is the most capable cybersecurity model the company has deployed so far.

OpenAI said it is preparing additional safeguards as agentic capabilities evolve, noting that it has already disrupted attempts to misuse its models in cyber operations.

Codex runs in a restricted sandbox by default, with limited file access and no network connectivity unless explicitly enabled. OpenAI recommends keeping these limits in place to avoid prompt-injection risks.

“Codex should be treated as an additional reviewer and not a replacement for human reviews,” the company said, adding that developers should examine all generated changes before deployment.

The post OpenAI Releases Agentic Codex Model That Can Work for More Than 24 Hours appeared first on Analytics India Magazine.

India has been actively advancing its semiconductor mission through aggressive investments and the development of its own fabrication plant, alongside multiple existing OSAT facilities. While this push has brought the industry to the forefront in India and globally, it cannot rely solely on building fabs and packaging plants.

This shift must also come into play while building Indian product companies that actually use these facilities, says Raja Manickam, founder of iVP Semi and former head of Tata Electronics’ OSAT operations. In an exclusive interview with AIM, Manickam says he believes that India will miss its moment if the country continues to rely on foreign chip suppliers for core components used in power, mobility, and consumer systems.

The end goal should be complete independence and not just manufacturing for its sake, he adds. “We have 100% dependency today. Not even one chip is made in India.” For him, the question driving India’s semiconductor story is as simple as who controls the basic chips that power everyday systems?

iVP Semi is his attempt to answer that question by building foundational power chips in India, starting with MOSFETs and other basic devices. These chips are not cutting-edge, but they sit at the heart of everyday products like EVs, renewable energy systems, consumer appliances, and industrial equipment. They also dominate India’s electronics bill of materials.

Acquisitions are Very Important

India has seen public funding for fabs and OSATs, but Manickam warns that these facilities may ultimately rely on foreign customers. He argues that this defeats the purpose.

Meanwhile, the CEO of Tessolve, Srini Chinamilli, told AIM that the customer’s nationality should not be the metric of success. “If it’s made in India and sold to the rest of the world, that is something you can be proud of,” he said, adding that global demand is essential for competitiveness.

He pointed out that India already has significant domestic demand, not just from Indian firms but also from multinational companies manufacturing for the Indian market, and that the first fab will still meet many local needs.

According to him, the upcoming facility can meaningfully serve segments such as power management, analogue components and low-end microcontrollers, even if it cannot provide the advanced nodes needed for data centres.

On the contrary, Manickam pointed to the example of CG Power’s OSAT, which will initially run on business from Renesas, a Tokyo-based company. He sees no issue with foreign customers, but questions the larger vision. “The government has put probably 70% into this capex, who’s the beneficiary of that?” His concern is that India may build capacity, but fail to link it to Indian product companies.

He also believes that India will need strong incentives for domestic use. “Can I get 20% favourable pricing?” he questioned, as an Indian semiconductor company. His argument is that if public money bridges early losses and learning cycles, then Indian companies could gain an advantage that allows them to scale.

Manickam is also sceptical that small fabless start-ups will drive meaningful growth. He added that the combined chip sales of Indian semiconductor startups are “under $10 million.” For context, the global market exceeds $500 billion.

Instead, he advocates a coordinated plan where the government funds not just facilities, but also integrated Indian companies or strategic acquisitions. “There has to be an element of acquisitions,” he suggests, citing the example of Nexperia, the Dutch company bought by Chinese investors for $200 million that has grown to more than $2 billion in sales.

His broader view is that India has everything needed for a semiconductor industry, except the ecosystem to retain and deploy its own talent. “We are training for the world, when will we help ourselves?”

A Domestic Market India Can Win

Manickam believes that iVP Semi can claim as much as a quarter of India’s power device market in the next decade. He said that the domestic market is large enough for Indian companies to scale. “If I cannot win the Indian market, forget about going global.”

He also believes that Indian businesses will buy local chips if they are competitive. “Japanese will buy from Japanese, Chinese will buy from Chinese, Germans will buy from Germans.”

“I just hope that Indians will buy from Indians.”

iVP Semi is also collaborating with institutions such as IIT Bhubaneswar and other state and private colleges to enhance design capacity. He says the collaboration is a “win-win” because academic teams get to see their designs enter commercial production rather than stay within research labs.

The company is preparing to launch new power modules and traction products, which will sit above individual chips. “Instead of selling one chip, we’re selling a subsystem,” he updated. These products are undergoing testing and will be announced in the coming months.

As India advances with its semiconductor efforts, Manickam remains focused on his concern. “The overall picture will be independence,” he said. For him, this means designing, manufacturing and consuming Indian chips, not only building factories or signing large partnerships.

It is “simple” he said: “If India cannot do semiconductors, I don’t think any other country can.”

Start With Power Blocks India Uses Every Day

Manickam sees the power and electrification segment as the most practical entry point for India’s chip ambitions. His view is that India should first target segments where chips are technologically achievable and in high domestic demand.

He also believes that India’s semiconductor industry must accept that it cannot jump straight to advanced processors or AI chips. “Do what you can do,” he said. “Once you build the foundation, then slowly other things will build on top of it.” His approach begins with “very simple building block chips” that universities and small teams can design and implement.

This aligns with iVP Semi’s product roadmap. The company currently buys wafers from Taiwan and Japan and will begin shifting to Indian OSATs once commercial packaging capacity becomes available. Testing will remain in-house. “I want my quality, reliability of the product that I ship guaranteed,” he said.

“Hopefully, when the Tata fab comes, we will move what we are buying from the wafer fabs outside,” he confirmed.

The post ‘We’re Creating Chip Factories for Foreign Customers, What’s the Point?’ appeared first on Analytics India Magazine.

Only 16 months after CrowdStrike brought the digital world to a standstill, and just a month after the global AWS outage, this week another key player in the cybersecurity space—Cloudflare—suffered a global disruption. Websites across the world, including Spotify, ChatGPT, X, and others, were hit with error messages as the outage rippled through the internet’s backbone.

These failures are million-dollar events.
According to the 2025 Observability Forecast, high-impact outages can cost Indian organisations between $1 million and $3 million per hour in lost revenue. And with the world entering peak Black Friday season, the hit is likely even worse.

Rob Newell, Senior Vice President and General Manager, Asia Pacific & Japan at New Relic, framed the moment bluntly, and told AIM: “With global outages becoming more common, observability is no longer an engineering tool, it is a business-critical practice.”
In an exclusive conversation on Front Page by AIM Network, New Relic CEO Ashan Willy pointed out that during the recent global AWS outage, New Relic’s engineers identified the issue 27 minutes before AWS informed its customers, giving teams a crucial early window to stabilise systems, isolate failures and protect revenue.

Against this backdrop of cascading outages, New Relic announced one of its most consequential integrations yet.

Azure Gets an Observability Boost

New Relic introduced new agentic AI integrations with Microsoft Azure, bringing its intelligent observability insights directly into the Azure SRE Agent and Microsoft Foundry. Powered by New Relic’s AI model context protocol (MCP) server and Azure Monitor, the integration embeds observability inside the workflow of developers, ITOps, DevOps, SREs and platform engineers — exactly where AI-driven troubleshooting now happens.

“AI agents are poised to transform how IT and development teams work, but leaders and practitioners need intelligent observability within their workflows to realise the full potential of agentic AI,” said New Relic chief product officer Brian Emerson.

“With our new integrations, we bring our AI-strengthened observability directly into Microsoft Azure products and services so teams can automate workflows and surface actionable insights, without having to context-switch. Together with Microsoft, we are helping more businesses harness the power of AI for growth.”

Microsoft reinforced the importance of this shift.

“Microsoft Azure helps IT teams and developers build AI-powered solutions that scale and inspire,” said Julia Liuson, president, Developer Division at Microsoft.

“These teams deserve a seamless workflow without switching between tools. Our latest integrations with New Relic mean that teams receive intelligent insights from Azure’s AI agents within their workflows so they understand exactly what’s going on during incidents. We’re driving an accelerated time to value and helping teams do more, faster.”

With the integration, the Azure SRE Agent can now tap New Relic’s MCP Server whenever an alert fires or a deployment happens, pulling the observability intelligence required to automate incident detection, root-cause analysis, and remediation across an organisation’s environment, from backend systems to mobile and browser experiences.

Inside Microsoft Foundry, New Relic’s telemetry is said to become a native part of how teams design, build and manage AI apps. Logs, metrics, deployment impacts, dependency graphs and configuration changes flow seamlessly into one real-time picture, giving teams clarity over how AI agents behave in complex, non-deterministic environments.

A Unified View in a Fragmented, Failure-Prone World

New Relic Azure Autodiscovery now claims to allow platform engineers to view entire service dependency maps and overlay configuration changes directly on performance graphs, making it possible to identify root causes in minutes instead of hours. And with New Relic Monitoring for SAP Solutions now available on the Microsoft Marketplace, the team said that the Azure customers can unify SAP and non-SAP systems without deploying agents inside SAP or disrupting mission-critical workflows.

As global outages increase in frequency, complexity, and financial impact, the message from New Relic is unmistakable: observability is not an option anymore. By wiring itself directly into Azure’s agentic ecosystem, New Relic is positioning its platform at the centre of how AI-driven incidents will be detected, understood, and resolved in real time.

“The world is seeing a dramatic rise in failure modes across the stack. Companies that treat observability as optional will pay for it—sometimes in minutes, sometimes in millions,” concluded Ashan Willy, CEO of New Relic, capturing the urgency perfectly.

The post The Internet Is Breaking. New Relic Wants to Fix It. appeared first on Analytics India Magazine.