Best of NVIDIA — October 2025

1. 1
   Article

   Where's Your Ed At·32w

   OpenAI Needs $400 Billion In The Next 12 Months

   OpenAI has committed to building 33GW of data center capacity through deals with NVIDIA, AMD, and Broadcom, with first deployments promised by late 2026. Analysis shows this requires approximately $400 billion in the next 12 months—more than global venture capital raised in 2024. The timelines are physically impossible: building 1GW of capacity costs $50 billion and takes 2.5 years, yet OpenAI promises three separate 1GW deployments in 18 months without disclosed construction sites or power infrastructure. The company burns $9.2 billion semi-annually against $4.3 billion revenue, faces a $20 billion funding clawback if it doesn't convert to for-profit by October 2026, and has announced plans for 250GW by 2033—requiring $10 trillion, or one-third of the US economy.

   40

   7
2. 2
   Article

   Ars Technica·32w

   Nvidia sells tiny new computer that puts big AI on your desktop

   Nvidia launched the DGX Spark, a $4,000 desktop AI workstation featuring one petaflop of computing power and 128GB of unified memory in a compact form factor. The system can run AI models with up to 200 billion parameters locally and fine-tune models up to 70 billion parameters, addressing the need for developers who want to avoid cloud services. Built on the GB10 Grace Blackwell Superchip with ConnectX-7 200Gb/s networking, it targets AI developers working with large language models and media synthesis applications. Orders begin October 15 through Nvidia's website and select retail partners.

   35

   2
3. 3
   Article

   NVIDIA·32w

   Elon Musk Gets Just-Launched NVIDIA DGX Spark: Petaflop AI Supercomputer Lands at SpaceX

   NVIDIA launched DGX Spark, a desktop-sized AI supercomputer delivering one petaflop of performance with 128GB unified memory, capable of running models up to 200 billion parameters locally. CEO Jensen Huang personally delivered the first unit to Elon Musk at SpaceX's Starbase facility. The system targets developers, researchers, and creators who need supercomputer-class AI performance in a portable form factor, with general availability starting October 15, 2025.

   30

   6
4. 4
   Article

   Where's Your Ed At·33w

   The AI Bubble's Impossible Promises

   An in-depth analysis of the AI infrastructure bubble reveals the impossibility of OpenAI's trillion-dollar data center promises. The piece examines critical power supply shortages, GPU depreciation economics, and physical constraints that make gigawatt-scale data centers unfeasible within promised timelines. Stargate Abilene currently has only 200MW of power for a planned 1.2GW facility, requiring at least 1.7GW total. With transformer shortages, electrical steel scarcity, and multi-year construction timelines, the article argues that AI companies' infrastructure commitments are fundamentally unrealistic, despite driving 92% of recent GDP growth through speculative investment.

   19

   2
5. 5
   Article

   80 LEVEL·31w

   NVIDIA "Went From 95% Market Share to 0%" in China, CEO Says

   NVIDIA CEO Jensen Huang revealed that US export restrictions caused the company to lose its entire Chinese market, dropping from 95% market share to zero. Huang criticized the policy impact, noting that China represents 50% of global AI researchers and the second-largest computer market. He argues that excluding Chinese AI developers from American technology harms both nations and slows worldwide AI development, expressing hope for future policy changes to restore market access.

   17

   2
6. 6
   Article

   Where's Your Ed At·30w

   Big Tech Needs $2 Trillion In AI Revenue By 2030 or They Wasted Their Capex

   Major tech companies have spent over $776 billion on AI infrastructure between 2023-2025, yet none are showing meaningful revenue from AI services. Microsoft reports only $13 billion annual recurring revenue from AI, with much of Azure's AI revenue coming from OpenAI's discounted compute costs. Analysis suggests these companies need to generate $2 trillion in AI revenue by 2030 to justify their capital expenditures, while currently every AI service provider except GPU manufacturers is losing money. The high costs of GPUs, data centers, and rapid hardware depreciation compound the challenge of achieving profitability.

   12

   1