Post Thumbnail

From renting servers to building them for other companies

OpenAI company seems to be preparing to become a player in the data center market. OpenAI CFO Sarah Friar openly stated plans to monetize their experience in building computational infrastructure. The company plans to create data centers for clients and lease capacity, as Amazon does with AWS.

OpenAI has accumulated unique experience in building infrastructure for artificial intelligence. Now they’re ready to monetize this knowledge. Imagine — the company that used to only rent other people’s servers will now build them for others. And there are also words that they might even abandon Nvidia in favor of cheaper hardware.

Sam Altman already warned that in the near future the company plans to spend $1 trillion on these data centers. OpenAI is possibly betting on inference optimization instead of expensive model training. And this could become the key to the company’s long-awaited profitability.

Unfortunately, it might be that OpenAI is actually acknowledging the failure of their main business model. If ChatGPT were potentially profitable, why get into infrastructure? Isn’t this diversification from desperation?

Autor: AIvengo
For 5 years I have been working with machine learning and artificial intelligence. And this field never ceases to amaze, inspire and interest me.

Latest News

How xAI competes with OpenAI in developer tools

xAI is launching Grok Code Fast 1. This is a compact agentic model for coding. $0.20 for 1 million input tokens, $1.50 for output — and just $0.02 when using cache!

Battle for browsers: Anthropic vs Perplexity and OpenAI

It seems Anthropic wants to transform the concept of browser technologies. Claude for Chrome enters closed testing for 1,000 Max-plan subscribers.

Chinese autonomous tractor without steering wheel and cabin works in fields

Chinese company Shiyan Guoke Honghu Technology introduced the fully autonomous tractor Honghu T70. Which independently moves across fields and performs the entire spectrum of agricultural tasks without any human participation.

Nvidia introduced Jetson AGX Thor: 2560 cores for robots

Nvidia company presented a development for physical AI - Jetson AGX Thor. This isn't just a chip, this is literally a brain for future robots. Imagine — 2560 Blackwell cores and 128 GB of RAM in one compact device!

GPT-5 optimizes costs

The Register reveals OpenAI's strategy and according to them, GPT-5 turned out to be not a revolution of capabilities, but genius cost optimization.