In addition to better AI, what other surprises can 5G mobile phones bring to people?

As a technology enthusiast, the CES and MWC conferences held at the beginning of each year can be said to be the two most unmissable technology events. The former brings together innovative technology products in various fields including automobiles, consumer electronics and other fields, displaying the latest scientific and technological achievements; the latter focuses on mobile communications technology, bringing together top companies in the communications industry, and is a benchmark for global communications development.

This year's MWC conference was held in Barcelona from February 26th to 29th, and a series of exciting technologies were released, such as concept products such as Lenovo Transparent Notes, Samsung Smart Rings, and Motorola Curved Phones, as well as forward-looking technologies such as 5G and AI. Innovations, some of which may well determine the direction of our technological development over the next five years.

Accelerate evolution, 5G further evolves

2024 is already the fifth year since 5G has entered commercial use. In the past 5 years, 5G networks have gone from a novelty to the infrastructure that connects smartphones and other smart devices today. If you look back at the video posted by Mr. He 5 years ago, you will find many predictions mentioned in it. All have become a reality today.

The large bandwidth and low latency characteristics of 5G have led to the explosive growth of mobile Internet usage scenarios. Short videos and mobile live broadcasts will become mainstream applications in the 5G era, changing people's work and lifestyle.

According to statistics, during Taylor Swift's concert at AT&T Stadium in Texas, the peak mobile data transmission volume was as high as 28.9TB, which was almost unimaginable in the 4G era. 5G allows more and more people to participate, record and share their lives in new ways.

However, some of our previous imaginations about 5G are still in the conceptual stage. Applications such as autonomous driving and remote surgery require extremely high speed and extremely low latency, and the existing 5G network is still difficult to fully support it. Industry customers are also calling on operators to improve network performance and achieve faster content access.

In order to unleash the full potential of 5G, the global communications industry is committed to the continuous evolution of 5G technology. In 2021, 3GPP named the next-generation evolution standard of 5G 5G Advanced and launched standardization work.

The highly anticipated 5G Advanced provides higher speeds, lower latency and greater capacity. Compared with 5G that was just commercialized five years ago, the bandwidth speed of 5G Advanced has increased by 10 times, reaching 10Gbps in the downlink and 1Gbps in the uplink.

At the same time, 5G Advanced’s latency is reduced to 1/10 of the original, connection density is increased by 10 times, and positioning accuracy is improved from 5G’s sub-meter level to centimeter level. In addition to huge improvements in performance, 5G Advanced also further reduces power consumption and achieves more efficient network control through deep integration with AI technology.

5G Advanced is expected to be commercially available by the end of 2024, and Qualcomm, which lays the foundation for the advancement of mobile communications by providing underlying technology products, is already prepared for the arrival of next-generation networks.

At the MWC 2024 conference, Qualcomm released its latest achievement in the 5G field – the Snapdragon X80 5G baseband, which is expected to be used on Android flagship models in the second half of this year.

The Snapdragon X80 5G baseband integrates a dedicated 5G AI processor and 5G Advanced-ready architecture, which can significantly improve the peak rate, range and stability of the 5G network. In fact, this is already the third generation of AI-powered 5G modems launched by Qualcomm, namely Snapdragon X70, X75 and X80.

You may be wondering, how can 5G and AI be combined?

Qualcomm said that with the AI ​​processor in the 5G modem, the Snapdragon X80 can improve user experience in different use cases, including multi-antenna management, situation-aware quality of service and delay optimization, as well as the average user experience. connection speed, network coverage, battery life, etc. Judging from the data, this is indeed the case – Snapdragon X80 improves positioning accuracy by 30%, reduces the time to select the best base station by 20%, increases link acquisition speed by 30%, and increases CPE service acquisition speed by 60% (millimeter wave). Power consumption is reduced by 10% in networking mode (millimeter wave).

The combination of 5G and AI is actually the general trend. To give a simple example, through the efficient machine learning capabilities of AI, the beam management of the 5G system can be improved, allowing the base station to achieve more accurate beam prediction in space, thus improving the accuracy of data transmission and reducing energy consumption. Simply put, when AI recognizes that you have entered the transmission range, it can predict your course of action, accurately transmit high-speed network signals to your device, and improve network transmission efficiency.

In addition, Snapdragon X80 has also brought groundbreaking breakthroughs in other features: the first to achieve downlink 6 carrier aggregation in the Sub-6GHz frequency band, the first to support 6-channel reception for smartphones, the first AI-based multi-antenna management, the first To support AI-enabled extended-range communications for CPE, the NB-NTN satellite communications function is integrated into the 5G modem for the first time.

The signs are beginning to show, AI is accelerating its popularity

The next revolution in smartphones will be comprehensive AI. This view has become the consensus of almost all mainstream mobile phone manufacturers.

According to IDC's definition, the new generation of AI mobile phones is characterized by NPU computing power greater than 30TOPS, equipped with an SoC that can support faster and more efficient end-side generative AI models, and supports generative models including Stable Diffusion and various large language models. The AI ​​model runs on the device side.

IDC further predicts that global shipments of new-generation AI mobile phones will reach 170 million units in 2024, accounting for nearly 15% of total smartphone shipments, a significant increase from approximately 51 million units shipped in 2023. AI phones will soon become the mainstream of the mobile phone market, bringing consumers a refreshing AI experience.

Since the second half of last year, AI large model functions have almost become standard features on flagship mobile phones. Huawei released the Hongmeng 4.0 system and announced access to the Pangu large model; OPPO and vivo have successively released the Find X7 series and X100 series with terminal-side large models, realizing AIGC elimination function and AI large model voice summary and other functions; Samsung released Galaxy AI The flagship Galaxy S24 series brings video AI processing, image processing, real-time call translation and other functions.

The just-released Xiaomi 14 Ultra is also equipped with Xiaomi AISP, the first AI large-model computational photography platform. By fully integrating CPU, GPU, NPU and ISP computing power, it can achieve 60TOPS computing power, bringing "super Capture" and "Super Film" functions.

The reason why AI can become so popular in such a short period of time is inseparable from the long-term layout of hardware manufacturers. Judging from the development of Qualcomm Snapdragon series mobile chips in recent years, the rate of progress in AI computing power has far exceeded that of CPU and GPU, becoming another important indicator of chip computing power.

The third-generation Snapdragon 8 released by Qualcomm in the second half of last year is the first mobile chip to support the running of 10 billion parameter AI large models on the device side, and can smoothly run complex generative AI models such as Stable Diffusion on the device side. To a certain extent, mobile phones equipped with the third-generation Snapdragon 8 are already at the starting line of AI mobile phones.

At MWC 2024, Qualcomm also released the latest progress in the field of AI-the new Qualcomm AI Hub. AI Hub provides more than 75 mainstream AI and generative AI models, including Whisper, ControlNet, Stable Diffusion, Baichuan-7B, etc., allowing developers to more easily combine AI models with applications and improve development efficiency.

It is worth noting that the much-anticipated LoRA model on Stable Diffusion can now run on mobile phones. Qualcomm demonstrated the first LoRA model running on Android smartphones, which can generate high-quality AI images based on users' personal preferences.

Qualcomm's continued efforts in large-scale end-side models have given space for the revolutionary technology of generative AI to be implemented. This is because a large amount of private information is stored on our mobile phones. Only by completing information processing on the terminal side can privacy leaks be avoided. Not only that, the large end-side model can also improve the stability of operation, reduce computing costs, and truly bring widespread benefits to technology.

With the powerful understanding capabilities of large models, the human-computer interaction model we are accustomed to will undergo tremendous changes. Daily applications such as photography, photo editing, video shooting, and even shopping and recording will be subverted by AI technology.

Arrive at the Intelligent Future in advance

With the development of next-generation 5G technology, the mobile Internet is entering a new stage of development. How to achieve fast and stable connections between devices connected to the Internet of Everything has become a brand new topic. At the same time, AI technology is also advancing rapidly in various fields, and mobile devices have increasing requirements for computing power.

From the 2G era to the 5G era, Qualcomm is involved behind every technological evolution. Nowadays, 5G deployment is accelerating, and Qualcomm has begun to explore around 6G.

At MWC 2024, Qualcomm demonstrated a number of innovative technologies leading the development of 6G, such as ultra-large-scale MIMO systems for the 13GHz frequency band, which will help develop new spectrum resources in high-frequency bands and prepare technical reserves for 6G. Qualcomm also demonstrated demonstrations of wireless AI interoperability and efficiency improvements, as well as digital twin network technology. These innovations lay the foundation for 6G and push wireless technology to further push physical limits.

At the same time, Qualcomm is also continuing to promote the in-depth development of 5G, expand application areas, and promote the implementation of the 5G Advanced standard. Qualcomm, together with its partners, demonstrated 5G technology innovations for automobiles, AR, RedCap (lightweight 5G), space communications and media.

For example, in the automotive field, Qualcomm expands its cloud platform to support more vulnerable road users and uses vehicle trajectory information to optimize wireless coverage to ensure seamless connectivity of cars.

In the field of AR, Qualcomm and its partners have demonstrated a dynamic distributed computing solution that can switch between cloud and local computing in real time based on network conditions to provide a seamless AR experience. These efforts will allow 5G to benefit more vertical industries and unleash its potential to connect hundreds of billions of devices.

AI technology is a universal tool in this era and will penetrate into all walks of life. But there are only a few companies that can truly combine AI capabilities with wireless connectivity to develop innovative solutions. Qualcomm is leading the development of wireless AI with its technological advantages accumulated over the years.

In recent years, Qualcomm has increased its investment in AI software and hardware, making the Qualcomm Hexagon AI processor an industry leader in performance and efficiency. Qualcomm's continuous optimization of large models allows the generative AI experience to be seamlessly transferred to mobile devices, making AI that was originally costly and difficult to obtain ubiquitous.

In today's world, billions of devices are connected to each other, generating massive amounts of data. Establishing communication is the first step in transmitting information. As a pioneer in wireless communication technology, Qualcomm is leading the integration innovation of wireless technology and AI, connecting hundreds of billions of devices, and bringing us closer to a future where everything is connected and intelligent computing is ubiquitous.

# Welcome to follow the official WeChat public account of aifaner: aifaner (WeChat ID: ifanr). More exciting content will be provided to you as soon as possible.

Ai Faner | Original link · View comments · Sina Weibo