Tag: infrastructure

  • China Just Set a Target to Quadruple Its AI Power

    China’s industry ministry laid out a plan to grow the country’s AI computing power almost fourfold by 2030, from roughly twenty four hundred exaflops today to nearly ten thousand. To put a dollar figure on it, the blueprint commits close to three point eight trillion yuan, something like five hundred thirty billion dollars, in infrastructure…

  • OpenAI Just Bet 30 Billion on Compute

    OpenAI Just Bet 30 Billion on Compute

    OpenAI announced a $30 billion data center investment this week, and it is a statement about where frontier AI lives. Not in software. Not in clever ideas. In raw, permanent, physical infrastructure. Thirty billion dollars in servers and power. That is what it costs now to train and run frontier models at scale. That is…

  • Google’s Chip Just Got Very Specific

    Google’s Chip Just Got Very Specific

    Google is building a server chip called Frozen v2 that hardwires Gemini’s architecture directly into silicon. Not weights. Architecture. The blueprint itself becomes hardware. The payoff: 6 to 10 times the token throughput per watt compared to general-purpose TPUs. This is a bet on a settling architecture. Once your model design is baked into silicon,…

  • The Chip That Powers AI Infrastructure Just Went Public

    SK Hynix listed on Nasdaq on July 10, raising 26.5 billion dollars in the largest IPO by a foreign company in U.S. history. The stock opened at 170 dollars and closed at 168 dollars after surging past the 149 dollar offering price. Why does this matter for AI? SK Hynix makes high-bandwidth memory. HBM. The…

  • Anthropic Just Locked In 401 Megawatts

    Anthropic Just Locked In 401 Megawatts

    Anthropic signed a $19 billion, 20-year data center lease with TeraWulf. That is one number. The actual story is what it means. The facility is in Hawesville, Kentucky on the site of an old aluminum smelter. 401 megawatts of computing power. Initial capacity starts in the second half of 2027, fully online in 2028. This…

  • Meta’s Muse Spark 1.1 Just Got Serious

    Meta’s Muse Spark 1.1 Just Got Serious

    Meta released Muse Spark 1.1 this week as a paid AI model, and the move is significant because it means Meta is willing to compete on price and capability, not just platform reach. Muse Spark 1.1 is built for multi-agent workflows and coding, positioned directly against Claude and GPT-5.5. Mark Zuckerberg called it a strong…

  • A Memory Chip IPO Just Signaled the Infrastructure Wave

    A Memory Chip IPO Just Signaled the Infrastructure Wave

    SK Hynix, a South Korean semiconductor manufacturer, is raising $28 billion on the Nasdaq. Trading begins July 10. It would rank among the largest initial public offerings in history. Why now? AI. The company’s revenue jumped 198 percent year-over-year in Q1 2026. The entire growth story is built on one thing: high-bandwidth memory became a…

  • Korea’s Building Its Own AI Supercomputer

    Korea’s Building Its Own AI Supercomputer

    NAVER just announced something quiet but significant. They’re building sovereign AI infrastructure at gigawatt scale with NVIDIA, starting at 55 megawatts in Korea. This matters because it’s not just about compute. NAVER is taking NVIDIA’s Nemotron 3 model and fine-tuning it with Korean data and expertise to build HyperCLOVA X, their own frontier model that…

  • Meta’s Building the Next Cloud Monopoly

    Meta’s Building the Next Cloud Monopoly

    Meta just announced a cloud business to sell AI computing power. This is not about selling spare GPU cycles. This is Meta saying: everyone else is paying $10,000 a month to run GPUs. We’re going to undercut them. Here’s the play. Meta built extraordinary amounts of compute for its own models. Now it’s realizing what…

  • SoftBank just committed 10 gigawatts to US AI

    SoftBank just committed 10 gigawatts to US AI

    SoftBank announced SB Neo, a new company to operate AI cloud computing at gigawatt scale in the United States. They’re planning 10 gigawatts of capacity, launching fiscal 2027. The US is scrambling to build compute. Blackstone committed $30 billion to AI data centers. Microsoft and OpenAI are stacking infrastructure. Nvidia and TSMC are selling chips…