Huawei has just announced in Berlin the new SoC for smartphones called to accelerate adoption of the fifth generation of mobile communications.

chipset Kirin 990 is already a reality. What has been presented this morning Richard Yu, CEO of Huawei Mobile, during the keynote of the manufacturer with the occasion of IFA 2019, which is being held in Berlin, Germany.

The manager has stressed the important evolution that it brings to the telecommunications industry thanks to the inclusion in the own encapsulated by a modem 5G, which will allow for designs that do not compromise the space needed for this type of components, as well as the energy consumption. In this sense, the chipset is manufactured using a process of 7-nanometer EUV using a total of 10.300 million transistors, a true achievement of the miniaturization.



But it is the combination of the 5G with the capabilities of Artificial Intelligence embedded in the three cores NPU, which will allow the ecosystem of developers to design a new generation of mobile applications never seen before until now.

Richard Yu, CEO of Huawei Mobile, in the presentation of the chipset Kirin 990 5G

keep in mind that this first generation with modem 5G will be able to support download speeds of up to 2.3 Gbps and upstream 1.25 Gbps always and when the connection is made with base stations 5G, whose deployment is already a reality in major urban centers.

Yu has also confirmed that the first terminal governed by Kirin 990 will be the Huawei Mate 30, scheduled for launch the 19th of September in Munich, an event that they will attend Silicon to tell the tale first-hand.

Entering into the details of the technical specifications of the new SoC, there will be two versions, one with 5G and the other without this modem. The first will be with two cores Cortex-A76 to 2.86 GHz, 2 Cortex-A76 at 2,36 GHz and 4 Cortex-A55 to 1.95 GHz. The graphic section will be well covered with 16 cores Mali-G76, while the neuronal processing will be performed on 3 cores NPU, one of them of low power consumption for tasks of AI less complex.