Compute Scaling AI News & Updates

UAE's G42 and Cerebras Deploy 8 Exaflops Supercomputer in India for Sovereign AI Infrastructure

G42 and Cerebras are deploying an 8-exaflop supercomputer system in India to provide sovereign AI computing resources for educational institutions, government entities, and SMEs. The project is part of broader AI infrastructure investments in India, including commitments from Adani, Reliance, and OpenAI, with the country targeting over $200 billion in infrastructure investment over the next two years.

Reliance Announces $110 Billion AI Infrastructure Investment in India Over Seven Years

Mukesh Ambani's Reliance has announced a $110 billion plan to build AI computing infrastructure in India over the next seven years, including gigawatt-scale data centers and edge computing networks. The investment is part of a broader trend of massive AI infrastructure spending in India, with Adani Group and global firms like OpenAI also committing significant resources. Reliance aims to achieve technological self-reliance and dramatically reduce AI compute costs, powered by its green energy capacity.

Runway Secures $315M Series E at $5.3B Valuation to Develop Advanced World Models for AGI

AI video startup Runway raised $315 million at a $5.3 billion valuation to develop next-generation world models, AI systems that create internal representations of environments to predict future events. The company, which recently released its Gen 4.5 video generation model that outperformed Google and OpenAI offerings, plans to expand world model capabilities beyond media into medicine, climate, energy, and robotics. This strategic shift positions Runway among competitors like Fei-Fei Li's World Labs and Google DeepMind in the race to build world models viewed as essential for surpassing large language model limitations.

Anthropic Pursues $20 Billion Funding Round at $350 Billion Valuation Amid Intense AI Competition

Anthropic is closing a $20 billion funding round at a $350 billion valuation, doubling its initial target due to strong investor demand, just five months after raising $13 billion. The round is driven by intense competition among frontier AI labs and escalating compute costs, with major participation from Nvidia, Microsoft, and leading venture capital firms. The company's recent successes include widely-praised coding agents and new models for legal and business research that have disrupted traditional data firms.

Tech Giants Commit Record Capital Spending to AI Infrastructure Despite Investor Concerns

Amazon and Google are leading massive capital expenditure increases for 2026, with Amazon projecting $200 billion and Google $175-185 billion, primarily for AI infrastructure and data centers. Despite the companies' conviction that controlling compute resources is essential for future AI dominance, investor sentiment has been negative, with stock prices dropping across the sector in response to these unprecedented spending commitments. The disconnect between tech executives' belief in AI's transformative potential and Wall Street's concerns about profitability reflects fundamental uncertainty about returns on these enormous investments.

SpaceX and xAI Merge to Pursue Orbital Data Center Network for AI Computing

SpaceX has filed plans with the FCC for a million-satellite data center network and formally merged with xAI, Elon Musk's AI venture, signaling serious intent to build orbital AI infrastructure. Musk argues that solar panels produce five times more power in space, making orbital data centers economically compelling by 2028, with predictions that space-based AI capacity will exceed Earth's cumulative total within five years. The merged SpaceX-xAI conglomerate is headed for an IPO, positioning to capture a share of the hundreds of billions spent annually on data center infrastructure.

Meta Launches Massive AI Infrastructure Initiative with Tens of Gigawatts of Energy Capacity Planned

Meta CEO Mark Zuckerberg announced the launch of Meta Compute, a new initiative to significantly expand the company's AI infrastructure with plans to build tens of gigawatts of energy capacity this decade and hundreds of gigawatts over time. The initiative will be led by three key executives including Daniel Gross, co-founder of Safe Superintelligence, focusing on technical architecture, long-term capacity strategy, and government partnerships. This represents Meta's commitment to building industry-leading AI infrastructure as part of the broader race among tech giants to develop robust generative AI capabilities.

Nvidia Unveils Rubin Architecture: Next-Generation AI Computing Platform Enters Full Production

Nvidia has officially launched its Rubin computing architecture at CES, described as state-of-the-art AI hardware now in full production. The new architecture offers 3.5x faster model training and 5x faster inference compared to the previous Blackwell generation, with major cloud providers and AI labs already committed to deployment. The system includes six integrated chips addressing compute, storage, and interconnection bottlenecks, with particular focus on supporting agentic AI workflows.

Nvidia Considers Expanding H200 GPU Production Following Trump Administration Approval for China Sales

Nvidia received approval from the Trump administration to sell its powerful H200 GPUs to China, with a 25% sales cut requirement, reversing previous Biden-era restrictions. Chinese companies including Alibaba and ByteDance are rushing to place large orders, prompting Nvidia to consider ramping up H200 production capacity. Chinese officials are still evaluating whether to allow imports of these chips, which are significantly more powerful than the H20 GPUs previously available in China.

Data Center Energy Demand Projected to Triple by 2035 Driven by AI Workloads

Data center electricity consumption is forecasted to increase from 40 gigawatts to 106 gigawatts by 2035, representing a nearly 300% surge driven primarily by AI training and inference workloads. New facilities will be significantly larger, with average new data centers exceeding 100 megawatts and some exceeding 1 gigawatt, while AI compute is expected to reach nearly 40% of total data center usage. This rapid expansion is raising concerns about grid reliability and electricity prices, particularly in regions like the PJM Interconnection covering multiple eastern U.S. states.