SoftBank Japan cashing out on NVIDIA. Is this a 1st step down the hype curve into the trough? Although they are also highly invested in OpenAI.

• NVIDIA’s latest earnings highlight a dramatic surge in AI demand, driven by a shift from simple one-shot inference to more complex, compute-intensive reasoning tasks.
• Reasoning models require hundreds to thousands of times more computational resources and tokens per task, significantly increasing GPU usage, especially for AI coding agents and advanced applications.
• Major hyperscalers like Microsoft, Google, and OpenAI are experiencing exponential growth in token generation, with Microsoft alone processing over 100 trillion tokens in Q1—a fivefold year-over-year increase.
• Hyperscalers are deploying nearly 1,000 NVL72 racks (72,000 Blackwell GPUs) per week, and NVIDIA-powered “AI factories” have doubled year-over-year to nearly 100, with the average GPU count per factory also doubling.
• To meet this unprecedented demand, more than $300 billion in capital expenditure is being invested this year in data centers (rebranded by NVIDIA as “AI factories”), signaling a new industrial revolution in AI infrastructure.

Smart strategy by Nvidia.

Relative recent article on alternatives for optirun, primusrun, etc

... and that will be also the day when most of the simultaneous interpreters will go out of business https://en.wikipedia.org/wiki/Simultaneous_interpretation

In my tests using Standard Engine, in combo with G-Sync Compatible Driver, I get more screen flickering during menus.