Google’s Tensor G2 Is Made on Samsung’s 4nm LPE Process, Along With a New Exynos 5G Modem

The first-generation Tensor was made on Samsung’s 5nm process, so there was a possibility that Google would stick with the Korean manufacturer for its advanced manufacturing node dedicated to the Tensor G2. While a previous rumor stated that TSMC would not be Google’s chip supplier, we have received confirmation this time around.

Tensor G2 Also Features Samsung’s Exynos 5300 5G Modem, Though Little Data Is Available on It

It should be noted that Samsung has two 4nm technologies; one is the LPE variant, while the other is LPP. Sammobile reports that the Tensor G2 is mass produced on the 4nm LPE node instead of LPP, and it could be due to lowered manufacturing costs. Google is not known for ordering a large amount of Pixel smartphone shipments from suppliers, so it is possible that when demand for future models increases, the company will be able to give higher orders to Samsung and for better chip technologies.

Assuming Samsung does not give Google a better deal for mass producing future Tensor SoCs, the advertising giant could jump ship to TSMC. The Tensor G2 features two Cortex-X1 cores running at 2.85GHz, along with two Cortex-A78 cores operating at a 2.35GHz frequency. The remaining four cores belong to ARM’s Cortex-A55, and they run at 1.80GHz. As for the GPU, the Tensor G2 features a Mali-G710 GPU with seven cores.

Coming to the 5G modem, Samsung’s Exynos 5300 is integrated into the Tensor G2. There is little information surrounding the baseband chip, but we assume it is made on the 4nm LPE architecture, meaning it would be faster and more power-efficient than the 5G modem running in last year’s Pixel 6 and Pixel 6 Pro. For the time being, Google is expected to stay with Samsung for the Pixel 8 family for next year.

It has been reported that Google aims to use Samsung’s 3nm GAA technology for the Tensor G3, and there are significant advantages to using this manufacturing process. Samsung claims that next-generation chips will reduce power consumption by up to 45 percent, improve performance by 23 percent and reduce area by 16 percent when compared to the manufacturer’s 5nm technology. Perhaps by 2023, Google would have caught up to its rivals.