Google has hired former Intel executive Uri Frank to lead its custom chip division. Apart from Google, many companies have taken to chipmaking in the last few years to build competitive moats.
Ambitious Move
The Intel veteran will serve as the Vice-President of Engineering for server chip design at Google. Uri Frank has over two decades of experience in custom CPU design and delivery experience. His expertise in design engineering at Intel will come in handy for Google.
“I look forward to growing a team here in Israel while accelerating Google Cloud’s innovations in compute infrastructure,” Frank shared on LinkedIn.
Amin Vahdat, vice president of systems infrastructure at Google, said Frank would lead a “world-class team” in Israel focusing on systems-on-chip [SoC] designs.
“Together with our global ecosystem of partners, we look forward to continuing to innovate at the leading edge of compute infrastructure, delivering the next generation of capabilities that are not available elsewhere, and creating fertile ground for the next wave of yet-to-be-imagined applications and services,” he said.
Custom chips have been an integral part of Google’s strategy to build efficient computing systems. The company has plans to move away from buying motherboard components from different vendors to building its own “system on a chip” or SoC. Google has been sourcing CPUs, networking, storage devices, custom accelerators and memory — all from different vendors. SoC involves putting multiple functions on the same chip or multiple chips onto one package.
Google believes the trick is in streamlining. Meaning, a deeper integration of workloads is required to gain higher performance at less power consumption.
Compute at Google is at an important inflection point. To date, the motherboard has been our integration point, where we compose CPUs, networking, storage devices, custom accelerators, memory, all from different vendors, into an optimised system, Vahdat said.
Instead of integrating components on a motherboard, we are turning to Systems on Chip designs where multiple functions sit on the same chip, or on multiple chips inside one package. In other words, the SoC is the new motherboard, he added.
Google counts on Frank’s expertise to work with customers and partners to build new custom chip architectures and take them to the next level.
History
Almost five years ago, the search giant launched the Tensor Processing Unit (TPU) chips. These custom-designed chips for driving deep neural networks revolutionised the AI industry. Later, the company launched Video Processing Units (VPUs) and OpenTitan, to make computing cheaper and more secure. Google has also designed chips such as Titan M and Pixel Neural Core for its phones.
Google has constantly been on the lookout for hiring experienced candidates. It had in the past poached talents from the likes of Intel, Qualcomm, Nvidia etc.
Why Focus On Chips
Google forayed into chipmaking with the intention was to reduce costs and boost the profit margin. By developing own chipsets, Google cut back the production cost of its Pixel devices. It also allowed the company to improve the Android experience by customising chipsets.
Further, designing a custom server chip will help Google Cloud to take on Microsoft Azure and AWS. Amazon Web Services and Microsoft Azure are at different stages of developing and deploying processors in-house for new cloud instances.
The SoC-based approach will help Google improve the latency and bandwidth between different components by “orders of magnitude.
Looking Forward
Google will design individual components of SoCs when necessary. The company aims to build ecosystems that benefit the entire industry, is also open to buying components from other vendors, Vahdat said.
That said, Google’s commitment to custom chips is bad news for companies such as Intel and Nvidia, as they will miss out on one of their largest potential customers.
Recently, Intel’s new CEO Pat Gelsinger revealed plans to revitalise the chip company by doubling down on manufacturing. The company has committed around $20 billion on semiconductor plants.
- Anthropic Makes $2.08 in Revenue for Every $1 Spent on Compute — Here’s the Full IPO Story
- OpenAI Says AGI is Here. But Not Everyone is Buying It
- Why Oracle is Bringing Post-Quantum Security to Java
- What Really Happens When You Turn Your Selfie Into a 1980s AI Pic?
- AI Agents Are Making Data Security A Bigger Enterprise Problem
- OpenAI is Getting Nervous About Reinforcement Learning