Qualcomm Inc. today detailed its newest 5G and Wi-Fi networking chips, which will start shipping with smartphones in the second half of the year.

The modules debuted at the MWC mobile industry event taking place in Barcelona this week. They employ hardware features not available in Qualcomm’s earlier silicon, notably satellite communications support, as well as onboard artificial intelligence software. New AI features are also arriving to several of Qualcomm’s existing chips in the form of a neural network bundle called the AI Hub.

The first addition to the company’s product portfolio is the Snapdragon X80 5G Modem-RF System. It’s a bundle of semiconductors that can be used to equip a mobile device with 5G connectivity. According to Qualcomm, handsets featuring the X80 will be capable of downloading data at a rate of up to 10 gigabits per second.

The module’s core capabilities are provided by a compact modem chip. The chip turns the 5G radio signals that a phone’s antennas pick up into ones and zeros the onboard processor understands. A mobile device’s modem handles outgoing communications as well: it encodes data such as texts into radio signals that can be wirelessly transmitted to the nearest cell tower.

Qualcomm X80 combines its modem with several complementary components. There are six antennas for sending and receiving wireless data, as well as an AI-optimized accelerator chip. Qualcomm says the chip helps improve the X80’s performance across more than a half-dozen core metrics, including speed, latency and power efficiency.

The chip also helps manage the X80’s antennas. To speed up 5G connections, the antennas perform so-called carrier aggregation, a process whereby radio signals are spread across multiple frequencies. This approach makes it possible to send multiple signals in parallel rather than one after another, which increases the amount of data that can be sent per second.

Qualcomm says handsets equipped with the X80 can connect to not only terrestrial 5G infrastructure but also internet satellites. The feature could help expand the adoption of satellite internet in the Android ecosystem, where the company’s chips are widely used. Apple Inc., which also uses Qualcomm silicon to power some iPhone features, has supported the technology since 2022.

“The Snapdragon X80 modem-RF system sets the stage for 5G Advanced and the era of intelligent computing everywhere,” said Durga Malladi, Qualcomm’s senior vice president and general manager of technology planning and edge solutions.

Qualcomm debuted the X80 today alongside a second wireless networking module, the FastConnect 7900, that likewise targets mobile devices. But whereas the former product is designed to enable 5G connectivity, the FastConnect 7900 focuses on Wi-Fi networking. It also supports two other wireless protocols: Bluetooth and Ultra Wideband, a niche networking technology used for tasks such as controlling smart home appliances.

The FastConnect 7900 supports Wi-Fi download speeds of up to 5.8 gigabits per second. Additionally, it uses 40% less power than previous-generation products. According to Qualcomm, the speed and efficiency improvements are partly made possible by onboard AI software that “adapts to specific use cases and environments.”

Handset makers can use the chip together with two new RF frontend modules that Qualcomm is rolling out in conjunction. RF frontend is a term for the circuitry that processes the incoming radio signals, in this case Wi-Fi signals, a device receives. One of the new modules is 50% smaller than Qualcomm’s previous-generation silicon while the other uses half the power.

“FastConnect 7900 is a technological feat, leveraging AI to raise the bar and deliver leading Wi-Fi 7 and Bluetooth capabilities while integrating Ultra Wideband all on a single 6nm chip,” said Javier del Prado, vice president and general manager of Qualcomm’s mobile connectivity group. 

Rounding out the company’s MWC announcement lineup is the Qualcomm AI Hub, a new machine learning software bundle. It comprises more than 75 AI models that the chipmaker has optimized to run on its processors. Qualcomm makes processors for not only handsets but also virtual reality handsets, smart cars and, as of last year, personal computers.

The AI Hub includes multiple large language models and AI image generators. They’re available alongside several multimodal models, neural networks that can take not only text but also media files as input. Qualcomm plans to add more AI models over time, as well as expand the software bundle’s operating system support. 

The neural networks in AI Hub interact with the Qualcomm processor on which they run via a software interface dubbed the AI Engine Direct SDK. It allows developers to customize the inference workflow in various ways, such as by specifying which of a system-on-chip’s compute modules should run a given AI model. In the background, the software automatically optimizes neural networks to improve their performance.

Image: Qualcomm