You are currently viewing SemiWiki as a guest which gives you limited access to the site. To view blog comments and experience other SemiWiki features you must be a registered member. Registration is fast, simple, and absolutely free so please, join our community today!
OpenAI told investors it's targeting roughly $600 billion in total compute spend by 2030, trying to clarify its plans after touting a $1.4 trillion figure.
First instinct is to throw compute power at a problem. Second instinct is to optimize the software so it requires less hardware. I have heard Sam speak a few times, he was at the Intel event last year. Sam admitted there is room for improvement and that the OpenAI software is bloated. Then DeepSeek painfully proved the point. Hopefully this is part of that and it is a "less hardware will be needed" rather than "we can't afford enough hardware". Maybe both?
Sam Altman is a conman, OpenAI will never make any profit，it can only survive on constant money pouring from investors. Once investors realize that they will never get their money back，it will be the end of OpenAI
Sam Altman is a conman, OpenAI will never make any profit，it can only survive on constant money pouring from investors. Once investors realize that they will never get their money back，it will be the end of OpenAI
I would say he is a visionary/dreamer. I do not think he has a dishonest bone in his body. Making money might not be his thing but Sam has moved AI into the forefront of society for the greater good, absolutely.
I have been using ChatGPT for more than a year now and it is improving in leaps and bounds. SemiWiki is also integrated into OpenAI and the traffic we get as a result is astounding. We had 1M active readers before the integration and now we have 2.5M. SemiWiki traffic is already up 40% in 2026. I realize these are just numbers but the bottom line is that the knowledge gained by using AI will change the world. No more Googling around in garbage and advertisements.
Could Silicon Valley be overly optimistic? Nah that could never happen:
High Failure Rates: Approximately 95% of AI pilots fail to fully deploy, and 40% of projects are canceled, creating a massive discrepancy between market valuation and tangible results.
Declining Adoption: Evidence suggests corporate AI usage has slipped, with some surveys indicating a decline from 14% to under 12% in mid-2025.
My opinion: The US labs are just labs, AI needs to work in the real world, and the labs will cut you off if you do the things required to make it useful in the real world. The US labs are not serious about real-world applications. Because they lose control in that scenario.
There’s a reason Altman scooped up the OpenClaw creator in the blink of an eye. They’re still searching for the “killer app” for agentic. It’s all crazy Wild West duct tape and bubblegum right now, but whoever can build a functioning self-sufficient agent orchestrator for everyone’s devices is going to be a big winner.
I’m betting on Apple and Google because of the OS moat. Microsoft is incapable as evidenced by their “hurry up and make everyone hate everything you do” Windows strategy.
There’s a reason Altman scooped up the OpenClaw creator in the blink of an eye. They’re still searching for the “killer app” for agentic. It’s all crazy Wild West duct tape and bubblegum right now, but whoever can build a functioning self-sufficient agent orchestrator for everyone’s devices is going to be a big winner.
I’m betting on Apple and Google because of the OS moat. Microsoft is incapable as evidenced by their “hurry up and make everyone hate everything you do” Windows strategy.
Definitely Google in my opinion, Apple an Xai maybe. OpenAI only if Microsoft buys them. Xai because of Tesla, Twitter, SpaceX etc... The one with the most data and ability to monetize wins.
When ChatGPT shuts me down with "0 messages remaining" I just switch to another site. They need to find a better way to monetize it.
I would say he is a visionary/dreamer. I do not think he has a dishonest bone in his body. Making money might not be his thing but Sam has moved AI into the forefront of society for the greater good, absolutely.
Caution - just because Altman believes what he's saying, doesn't mean he isn't lying to himself. He is very good at marketing (look at his history), and he knows that he has to sell a vision to investors to keep the money flowing.
Our CEO recently admitted for the first time "no one has any idea how to monetize AI", last quarter his answer was "well no one can doubt Nvidia is making real money", so I think things are progressing. If it can't displace workers for profit, then it really is just a better correlation machine for pattern recognition and generation, just as Judea Pearl said years ago.
www.cnbc.com
