Musk's xAI to raise up to $12 billion in debt for AI expansion: Report
's artificial intelligence startup
xAI
is working with a financier to secure up to $12 billion more for its expansion plans, the Wall Street Journal reported on Tuesday, citing people familiar with the situation.
Valor Equity Partners, an investment firm whose founder, Antonio Gracias, has close ties to Musk, is in talks with lenders to raise the capital, according to the report.
The money would be used to buy a massive supply of advanced
Nvidia chips
that would be leased to xAI for a new huge data center meant to help train and power its
AI chatbot Grok
, the WSJ reported.
Valor and xAI did not immediately respond to Reuters' requests for comment.
Some lenders want the debt to be repaid within three years and to cap the amount of money borrowed in order to limit their risk, the report said.
Training and deploying advanced AI systems is expensive, requiring costly hardware, intensive compute and skilled engineers in a highly competitive market where rivals such as OpenAI, Google and China's DeepSeek are trying to dominate the technology.
XAI is training Grok on 230,000 graphics processing units, including Nvidia's 30,000 GB200 AI chips, in a supercluster, with inference handled by the cloud providers, Musk said in a post on X on Tuesday.
He added that another supercluster will soon launch with an initial batch of 550,000 GB200 and GB300 chips.
The AI startup was expected to burn through about $13 billion over the course of 2025, according to media reports.
Earlier this month, the Financial Times reported that xAI was in talks to raise more money in a deal that could value the AI startup between $170 billion and $200 billion.
In response, Musk had said xAI was not seeking funding. "We have plenty of capital."
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
Time of India
39 minutes ago
- Time of India
Agra police holds training session on AI for cops
Agra: In an initiative aimed at technologically empowering its personnel and enabling "effective use of artificial intelligence (AI) in day-to-day policing and curbing crime", Agra police on Wednesday held a special training session on prompt engineering based on AI. Tired of too many ads? go ad free now The session focussed on application of Large Language Models (LLMs) such as ChatGPT, Gemini and Perplexity in policing activities, including FIR and report writing, understanding and interpreting BNS sections, giving direction in cybercrime investigations, creating awareness material for cybersecurity and public outreach, drafting documents and summaries for analysis and interpretation, among others. During the session, Agra police also showcased its in-house developed AI apps, including FAI, EBITA and AI RAGBOT for UP police circular, which have already been integrated into the system. Agra DCP (city) Sonam Kumar said policemen participated in hands-on activities, crafted their own prompts, and gained a better understanding of the capabilities of ChatGPT, Gemini and Perplexity AI. "Those who completed the 'prompt engineering' training were awarded the title of AI commandos and were also provided a one-month paid subscription to Perplexity AI, enabling them to perform traditional duties with greater speed, accuracy and efficiency using new technology," the DCP said. The personnel were also instructed not to share any sensitive or confidential departmental information on AI models. "This training marks a significant step toward building a tech-savvy police force in Agra, and more such technical programmes will be organised in future," said Kumar. Agra police commissioner Deepak Kumar said, "In modern policing, the use of technology and AI is no longer optional — it has become a necessity. Such initiatives will enhance the efficiency and response time of the police personnel."
Hindustan Times
an hour ago
- Hindustan Times
AI must aid human thought, not become its replacement
Watching the recent resurgence of violence in Kashmir, I find myself grappling with questions about the role of technology, particularly Generative Artificial Intelligence (GenAI), in warfare. India is built upon the philosophy of live and let live, yet that doesn't mean passively accepting aggression. As someone deeply invested in responsibly applying AI in critical industries like financial services, aerospace, semiconductors, and manufacturing, I am acutely aware of the unsettling dual-use potential of the tools we develop: The same technology driving efficiency and innovation can also be weaponised for harm. We stand at a critical juncture. GenAI is rapidly shifting from mere technological advancement to a profound geopolitical tool. The stark division between nations possessing advanced GenAI capabilities and those dependent on externally developed systems poses serious strategic risks. Predominantly shaped by the interests and biases of major AI-developing nations, primarily the US and China, these models inevitably propagate their creators' narratives, often undermining global objectivity. Consider the inherent biases documented in AI models like OpenAI's GPT series or China's Deepseek, which subtly yet powerfully reflect geopolitical views. Research indicates these models minimise criticism of their home nations, embedding biases that can exacerbate international tensions. China's AI approach, for instance, often reinforces national policy stances, inadvertently legitimising territorial disputes or delegitimising sovereign entities, complicating fragile diplomatic relationships, notably in sensitive regions like Kashmir. Historically, mutually assured destruction (MAD) relied on nuclear deterrence. Today's arms race, however, is digital and equally significant in its potential to reshape global stability. We must urgently reconsider this outdated framework. Instead of mutually assured destruction, I advocate for a new kind of MAD: mutual advancement through digitisation. This paradigm shifts the emphasis from destructive competition to collaborative development and technological self-reliance. This evolved MAD requires nations, particularly technologically-vulnerable developing countries, to establish independent, culturally informed AI stacks. Such autonomy would reflect local histories, cultures, and political nuances, making these nations less susceptible to external manipulation. Robust, culturally informed AI not only protects against misinformation but fosters genuine global dialogue, contributing to a balanced, multipolar AI landscape. At the core of geopolitical tensions lies a profound challenge of mutual understanding. The world's dominant AI models, primarily trained in English and Chinese, leave multilingual and culturally diverse nations like India, with its 22 official languages and hundreds of dialects, in a precarious position. A simplistic AI incapable of capturing nuanced linguistic subtleties risks generating misunderstandings with severe diplomatic repercussions. To prevent this, developing sophisticated, culturally aware AI models is paramount. Multilingual AI systems must leverage similarities among related languages such as Marathi and Gujarati or Tamil and Kannada to rapidly scale without losing depth or nuance. Such culturally adept systems, sensitive to idiomatic expressions and contextual subtleties, significantly enhance cross-cultural understanding, reducing the risk of conflict driven by miscommunication. As GenAI becomes integrated into societal infrastructure and decision-making processes, it will inevitably reshape human roles. While automation holds tremendous promise for efficiency, delegating judgment, especially in life and death contexts like warfare, to AI systems raises profound concerns. I am reminded of the Cold War incident in 1983 when Soviet Lieutenant Colonel Stanislav Petrov trusted human intuition over technological alarms, averting nuclear disaster — a poignant reminder of why critical human judgment must never be relinquished to machines entirely. My greatest fear remains starkly clear: A future where humans willingly delegate judgment and thought to algorithms. We should not accept this future. We share collective responsibility as innovators, technologists, and global citizens, to demand and ensure that AI serves human wisdom rather than replaces it. Let's commit today: never allow technology to automate away our humanity. Arun Subramaniyan is founder and CEO, Articul8. The views expressed are personal.
The Hindu
an hour ago
- The Hindu
Google Play, Android ecosystem generated ₹4 lakh crore in revenue for Indian app publishers in 2024: Google
Google on Wednesday (July 23, 2025) said, based on third-party evaluations, the Google Play and Android ecosystem generated ₹4 lakh crore in revenue for app publishers and the wider economy in India in 2024, and supported the creation of 35 lakh direct, indirect, and spillover jobs in the country. According to Google, the Indian ecosystem comprises the second largest number of active developers on Google Play across the world, contributing over 10 lakh developer jobs in 2024. Building compelling offerings on Android, this ecosystem has played an important role in shaping the impact and inclusiveness of India's digital economy. Approximately 72% of surveyed Indian users claimed their first access to the internet was on an Android device; some 85% say their Android phone is one of the main ways they access digital public services, and 69% report first using AI through an app on their Android device. Device makers have also been able to make devices affordable due to the open-source nature of Android, which has assisted in estimated savings of ₹25,200 crore in development and operating costs. At Google I/O Connect India 2025, an annual developer conference held in Bengaluru, Google announced new initiatives and AI capabilities to support the Indian developer and start-up ecosystem in the country. These included localising data processing of its high-performance thinking model Gemini 2.5 Flash, new Agentic AI tools in Firebase Studio, a new training programme with one of the world's leading gaming engine platforms – Unity, and collaborations with 3 India AI Mission start-ups in building innovative solutions toward the development of India's Make-in-India AI models. Dr. Manish Gupta, Senior Director for India and APAC at Google DeepMind, who opened the event, said, 'Indian developers are writing the next chapter of India's success story, using AI capabilities to build real-world applications that are reaching millions of businesses and people across the country and the world.' Preeti Lobana, Country Manager, Google India, who held a fireside chat with Subrata Mitra, Partner, Accel, on building enduring AI-driven companies during the event's opening keynote, said, 'India's developers are shaping how the world will use AI, and we're proud to stand with them. We're giving India the best tools, the most open platforms, and the strongest support to build boldly for the world, fostering a profound 'AI Productivity Leap' across businesses and start-ups.' Google said it was closely partnering with BharatGen at IIT Bombay to build indigenous Indic language Automatic Speech Recognition (ASR) and Text-to-Speech (TTS) models that would aim to make Gemma (a set of lightweight, generative artificial intelligence open models) more helpful in the Indic context, especially for historically underrepresented languages and governance applications.
