Former Google CEO Eric Schmidt just lately held a chat session at Stanford and informed college students that he believes the tech firm that can be a winner within the period of AI is not going to be his former employer, Microsoft or ChatGPT-maker OpenAI however the one on which all these depend on.
Schmidt stated that massive expertise firms are planning for giant investments into Nvidia-based AI information centres, which can price as a lot as $300 billion to construct.
“I’m speaking to the massive firms, and the massive firms are telling me they want $20 billion, $50 billion, $100 billion — very very arduous,” stated Schmidt, including that he’s a “shut good friend” to OpenAI CEO Sam Altman.
“If $300 billion is all going to Nvidia, what to do within the inventory market,” Schmidt stated, including, “That’s not a inventory suggestion.”
Schmidt’s indication that such an enormous quantity going to Nvidia – which makes the extremely in-demand information centre AI chips H100 – will make it a winner. Notably, has already seen income improve by greater than 200% for 3 straight quarters. And it even leapfrogged massive firms by way of valuation.
“For the time being, the hole between the frontier fashions — there are solely three — and everybody else seems to be getting bigger,” Schmidt stated.
“Six months in the past, I used to be satisfied that the hole was getting smaller, so I invested plenty of cash within the little firms. Now I’m not so positive,” he added.
Schmidt stated that massive expertise firms are planning for giant investments into Nvidia-based AI information centres, which can price as a lot as $300 billion to construct.
“I’m speaking to the massive firms, and the massive firms are telling me they want $20 billion, $50 billion, $100 billion — very very arduous,” stated Schmidt, including that he’s a “shut good friend” to OpenAI CEO Sam Altman.
“If $300 billion is all going to Nvidia, what to do within the inventory market,” Schmidt stated, including, “That’s not a inventory suggestion.”
Schmidt’s indication that such an enormous quantity going to Nvidia – which makes the extremely in-demand information centre AI chips H100 – will make it a winner. Notably, has already seen income improve by greater than 200% for 3 straight quarters. And it even leapfrogged massive firms by way of valuation.
“For the time being, the hole between the frontier fashions — there are solely three — and everybody else seems to be getting bigger,” Schmidt stated.
“Six months in the past, I used to be satisfied that the hole was getting smaller, so I invested plenty of cash within the little firms. Now I’m not so positive,” he added.
Tech giants making their very own AI chips
In a bid to scale back their reliance on Nvidia, tech giants are creating their very own chips. Google has developed chips known as Tensor Processing Models (TPUs) to compete with Nvidia’s processors.
Microsoft introduced the Azure Maia 100 AI chip final 12 months and it’s designed to run cloud-based AI workloads. In the meantime, Amazon can also be making ready Trainium chips and Fb-parent Meta revealed plans for “Artemis,” a second-generation Al chip surpassing their preliminary Meta Coaching and Inference Accelerator (MTIA) product launched final 12 months.
