Rene Haas says Arm CPUs are in every mobile phone built, including all iPhones and Android phones.
“we are in literally every single mobile phone uh built. We're in all the iPhones, we're in all the Android phones”
Person
Rene Haas
1 episodes · 21 quoted claims
Rene Haas says Arm started in a barn in Cambridge with the ambition to be the global standard for CPUs.
“started in a a barn in Cambridge with the ambition to be the global standard for CPUs, and uh, that mission, they accomplished it.”
Rene Haas says Arm began as a joint venture between VLSI Technology and Apple, seeking a chip for the Newton PDA.
“it was a joint venture between chip company called VLSI Technology and and Apple. And they were looking for the first chip for a first PDA known as the Newton.”
Rene Haas says the Newton needed a low-power CPU that could run off a battery, which is why Arm was conceived.
“what it needed was a low-power CPU, something that could run off a battery”
Rene Haas says Arm entered general-purpose cloud data centers years ago with Amazon, Microsoft and Google.
“ARM had gotten into the cloud data centers a number of years ago, general purpose cloud with Amazon and Microsoft and uh, and Google.”
Rene Haas says Nvidia chose Arm CPUs to pair with the accelerators in its Grace Hopper and Grace Blackwell chips.
“when NVIDIA made the choice with their Grace Hopper chip and then Grace Blackwell, they needed a CPU to pair up with the accelerator, and and they chose ARM”
Rene Haas says the CPU manages data and prepares it for the GPU accelerator in AI data centers.
“the CPU does all of the the management of the data, it prepares the data that goes in the accelerator.”
Rene Haas says wearable glasses have only about two watts of power, so AI there must run on the CPU rather than a GPU.
“But when you go to a very small application where wearable glasses, where you have maybe two watts of power, you can't put a GPU there.”
Rene Haas says a major bottleneck in drug discovery today is human trials, which he expects AI to accelerate.
“If you look at one of the bottlenecks in drug discovery today, it's actually around trials, human human trials.”
Rene Haas says OpenAI spends three to eight months training a large model such as ChatGPT.
“Months and months and months are spent training a model, training chat GPT, and OpenAI takes three months, six months, eight months to train this giant model.”
Rene Haas says that today companies still spend more compute time training models than running inference.
“First things today, still researching companies spend more time training than they do on inference.”
Rene Haas says during COVID a replacement car key fob took fifty-two weeks because of semiconductor shortages.
“And when you lost your key fob, it took fifty two weeks to get a new one.”
Rene Haas says TSMC is by far the main provider of AI compute chips, including everything Nvidia builds.
“Uh, TSMC is by far and away the provider of all these AI compute chips. Everything NVIDIA builds is from there.”
Rene Haas says ASML, a Dutch company, is the only maker of EUV chip-manufacturing machines.
“for EUV, it's a company called ASML Dutch company. They're the only people in the world who build EUV machines.”
Rene Haas says Zeiss of Germany is the only company making the mirrors used in EUV machines.
“There are a set of mirrors that go into that EUV machine made from a company in Germany called Zeiss. They're the only company in the world that makes those mirrors.”
Rene Haas says hyperscalers including Google, Microsoft, Meta and AWS are building their own AI chips, with Google's TPUs spanning several generations.
“You'll have hyperscalers doing their own chips, whether that's Google TPUs, and that's gone on for a number of generations, Microsoft, Meta, AWS,”
Rene Haas says an AI chatbot's 'thinking' delay is the computer waiting for data, which is a memory bandwidth limit.
“It's it's not the computer that's thinking, it's the computer is waiting for data to come into the computer to figure out how to make the answer. That's memory bandwidth.”
Rene Haas says he read a New York Times article comparing Waymo's conventional accident counts with those of human drivers.
“I was looking at Waymo, uh, and the number of accidents that Waymo had conventional accidents compared to human drivers.”
Rene Haas predicts that within five to ten years large parts of factory work will be replaced by robots.
“I think in the next five to ten years, and again, ARM will be involved. You're gonna see large, large sections of factory work replaced with robots.”
Rene Haas says customers are asking Arm to build chips for them instead of waiting for Arm's designs, and Arm continues to explore this.
“It's not unnatural for customers to come to us and say, if the computer and models are all running on your platform anyway, instead of waiting for your IP, why don't you just build something for us?”
Rene Haas predicted that no one will be talking about an AI bubble a year after the Fortune Brainstorm AI conference.
“And I said, no one will be talking about an AI bubble a year from now.”