• AI KATANA
  • Posts
  • Meta in talks for $200 billion AI data center project

Meta in talks for $200 billion AI data center project

Also: Register for Jensen Huang's NVIDIA GTC 2025 Keynote

Author

AI KATANA
February 26, 2025

Hello!

The AI industry continues its rapid expansion, with Meta reportedly planning a $200 billion AI data center campus, despite denying the figure. Meanwhile, Alibaba has open-sourced its AI video generation model, intensifying competition with proprietary AI solutions. DeepSeek, a rising Chinese AI startup, is accelerating the release of its next-generation model, reinforcing China’s strategic push into AI. A McKinsey report explores why generative AI adoption remains slower than expected, highlighting enterprise challenges. In education, AI usage among UK university students has surged, prompting institutions to rethink assessment policies. OpenAI has also expanded its Deep Research tool to ChatGPT Plus users, making advanced AI-powered research more accessible. Finally, India’s AI startup ecosystem is growing, but experts debate whether it can produce 100 unicorns in the next decade.

Sliced just for you:

  • 🏗️ Meta in talks for $200 billion AI data center project

  • 🎥 Alibaba makes AI video generation model free to use globally

  • 🚀 DeepSeek rushes to launch new AI model as China goes all in

  • 🤖 When will we see mass adoption of gen AI?

  • 🎓 Surge in UK university students using AI to complete work

Meta is reportedly in discussions to build a massive AI-focused data center campus with an estimated cost exceeding $200 billion, considering locations in states like Louisiana, Wyoming, or Texas. This move aligns with the ongoing surge in AI infrastructure investments, with Meta planning to spend up to $65 billion in 2025 to expand its AI capabilities. While Meta has denied the $200 billion figure, the report underscores the intense competition among tech giants, as Microsoft and Amazon also plan significant investments in data centers. The AI boom, sparked by OpenAI’s ChatGPT in 2022, continues to drive unprecedented spending, reshaping the landscape of cloud computing and AI infrastructure development.

Alibaba has announced that it is making its latest AI-powered video generation models, part of the Wan2.1 series, freely available worldwide. These models, capable of generating images and videos from text and image inputs, will be open-sourced via Alibaba Cloud’s Model Scope and Hugging Face, making them accessible to researchers, academics, and commercial users. This move positions Alibaba in direct competition with proprietary AI solutions like OpenAI’s and aligns with a broader trend among Chinese firms favoring open-source AI development. The announcement follows the recent impact of DeepSeek, another Chinese firm that gained attention for training an AI model at lower costs on less advanced hardware. Open-sourcing AI technology fosters innovation and community engagement but also raises questions about whether AI models will become commoditized. Alibaba's stock has surged 66% in 2025, reflecting investor optimism about its AI advancements and growing government support for China’s private sector.

DeepSeek, the Chinese AI startup that disrupted global markets with its cost-effective AI reasoning model, is accelerating the launch of its next-generation R2 model, aiming for an earlier release than its initial May target. The company's R1 model, built using less advanced Nvidia chips, proved highly competitive against Western AI giants, sparking a shift in industry dynamics. DeepSeek's unconventional approach, favoring a flat management structure and prioritizing AI research over consumer-facing applications, has gained favor with Chinese authorities, despite initial regulatory scrutiny over its massive chip acquisitions. The firm’s success stems from strategic investments in computing power and efficient AI architectures, including Mixture-of-Experts and Multihead Latent Attention, which reduce operational costs while maintaining performance. As China rallies behind DeepSeek, integrating its models into government and corporate systems, Western regulators may see the firm as a growing geopolitical challenge, potentially prompting further restrictions on AI-related exports.

The widespread adoption of generative AI is taking longer than expected, but its development is accelerating rapidly, according to experts from McKinsey and industry leaders. While AI is already transforming sectors like customer service and content creation, mass adoption requires more time for users to learn how to integrate AI into daily workflows. Consumer adoption is progressing faster than enterprise adoption, where concerns over data security, legal risks, and talent shortages slow implementation. Experts predict AI will evolve into "AI teammates" that augment human capabilities rather than merely automating tasks. Businesses that embrace AI as a tool for amplifying productivity and problem-solving will gain a competitive edge, while those hesitant to adapt risk falling behind. However, concerns over AI-generated misinformation, job displacement, and regulatory challenges remain. The next wave of AI integration will likely reshape industries by shifting business models, requiring enterprises to rethink their approach to digital transformation.

The use of AI among UK university students has surged dramatically, with over 90% now incorporating it into their studies, compared to two-thirds a year ago. A report from the Higher Education Policy Institute found that 88% of students have used generative AI like ChatGPT for assessments, with science students being the most frequent users. The main motivations are saving time and improving work quality. Despite this widespread adoption, there is still confusion over AI policies, with students receiving mixed messages about its acceptability. While a growing number of students find it acceptable to include AI-generated text after editing, only 6% believe unedited AI content should be allowed. Universities are under pressure to rethink assessment methods and staff training to adapt to this rapid shift, as concerns grow over how AI will reshape higher education.

🛠️ AI tools updates

OpenAI has made its Deep Research feature, previously available only to Pro users for $200 per month, accessible to all ChatGPT Plus subscribers at a significantly lower cost of $20 per month. The tool allows users to conduct complex, multi-step research tasks, autonomously browsing the internet, analyzing sources, and generating detailed reports. The update also introduces improved handling of uploaded files, ensuring better integration of user-provided documents with real-time web information. While Plus, Team, Edu, and Enterprise users now have access to 10 deep research queries per month, Pro users see an increase from 100 to 120 queries. OpenAI plans to expand the feature with enhanced visual outputs, such as embedded images and data visualizations, reinforcing its commitment to making advanced AI research capabilities widely available.

💵 Venture Capital updates

India’s venture capital market is recovering, with AI emerging as a key focus for investment. Despite an increase in overall VC funding to $11.2 billion in 2024, AI startups in India continue to struggle with securing sufficient capital, and Krutrim AI remains the country’s only AI unicorn. The government’s IndiaAI Mission has received 67 proposals, with 20 focused on large language models, signaling a shift towards foundational AI development. Peak XV Partners has allocated ₹16,000 crore, adding to the $20 billion available for AI startup investment, while global investors such as Accel and Antler continue backing Indian AI firms. India ranks second in GitHub contributions to generative AI, showcasing its strong talent pool, though many top professionals are based in the US. With support from incubators like Google, NVIDIA, and JioGenNext, along with government-backed GPU access, India’s AI ecosystem is growing. However, achieving 100 AI unicorns in the next decade will require sustained investment, regulatory clarity, and a shift towards deep-tech innovation.

🫡 Meme of the day

⭐️ Generative AI image of the day