No. 67
2026-09-24
the AI pulse — brewed, not raw
 AI CRITIQUE
Görsel, yapay zeka iş yüklerinin etkisiyle 2030 yılına kadar küresel veri merkezi talebinin iki katından fazla artacağını öngören bir grafiği, üç farklı büyüme senaryosu ve bir konuşmacının fotoğrafıyla birlikte sunan…

NVIDIA executive warns against China's AI push

Read the article    English (automatic translation)

Summary

NVIDIA Vice President Rod Evans issues a stark warning that China's fully automated, robot-only factories are forcing Western industry to a 'change or die' juncture. Evans also rejects the AI bubble narrative, criticizing OpenAI and Anthropic for using security rhetoric to protect their closed-model market share.

Why it matters

The article frames a concrete industrial-geopolitical risk: if Chinese manufacturers are shipping higher-quality, cheaper products from zero-human facilities, the competitive window for Western automakers and industrial firms is rapidly closing. Evans also highlights physical infrastructure bottlenecks, such as 1 MW racks, 800 V DC power, and full liquid cooling by 2028. This infrastructure will determine which countries can scale 'AI factories,' turning the choice between sovereignty and consumer benefit into a near-term rather than long-term policy decision.

Highlights

  • Evans claims that AI model performance has increased by a factor of 1,000,000 over the past decade and that token costs have dropped tenfold annually. He argues that demand is real, not speculative.
  • Evans reports that Western executives returning from China are 'terrified' by 100% robotic factories. A Ford executive warns, 'We must change or die.'
  • NVIDIA's next-generation racks reach 1 MW (compared to today's 224 kW). This necessitates a transition to 800 V DC power and 100% direct liquid cooling with 45 °C inlet water.

Critical take

Every positive data point Evans relies on—consecutive quarters of 17% growth, the 26/26 hand-raising anecdote, the million-fold performance claim—comes from an executive at NVIDIA, the primary beneficiary of AI infrastructure spending. Consequently, the 'no bubble' framework structurally serves his own interests. The Ford warning and the story of 'terrified leaders' are second-hand, anonymous, and unverifiable, making them read as color for a promotional narrative rather than independent evidence.

Bu kritik, yerel yapay zeka modeli Qwen3.8-27B tarafından yazılmıştır. Haberi ben seçtim; kritik metnini yayımlanmadan önce tek tek denetlemiyorum.This critique was written by the local AI model Qwen3.8-27B. I chose the story; I do not review each critique before it is published.
Back to the issue