Recently, IDC, an authoritative consulting firm, released the "Market Share of Deep Learning Frameworks and Platforms in the Second Half of 2019" report, which provides an in-depth analysis of the current situation of the deep learning framework and platform market in China. The report shows that in the Chinese deep learning platform market, Google, Facebook, and Baidu's deep learning platforms have accounted for nearly 80% of the market share.
Baidu, which owns the PaddlePaddle deep learning platform, has experienced rapid market share growth in the past six months, with an increase of 5.98%. The next opportunity for AI is exploding at the edge. A few days ago, NVIDIA GTC CHINA 2019, an annual AI event, officially opened in Suzhou. Jensen Huang, the CEO and founder of NVIDIA, emphasized the new opportunities for edge intelligence. At the event, Baidu announced that its open-source deep learning platform, PaddlePaddle, has been adapted to the NVIDIA EGX edge computing platform. PaddlePaddle's deep learning inference models can run directly on the EGX platform, providing an end-to-end solution for edge-cloud collaboration. The collaboration between the two companies has created a solid technical foundation for the implementation of edge AI.
How powerful is PaddlePaddle? In the past year, the application of AI has flourished, bringing the intelligent era closer to people. While major technology companies are developing various AI applications, they are also investing heavily in core AI areas such as underlying algorithms, chips, and deep learning frameworks/platforms, aiming to provide stronger impetus for industrialized and scaled AI applications. Among them, deep learning frameworks, which integrate various models and applications at the top and connect with chips and various hardware at the bottom, have been highly valued as the "operating system of the intelligent era". Deep learning platforms, which are based on open-source deep learning frameworks, are becoming powerful tools to assist in the implementation of scaled AI applications.
Baidu has developed China's first and currently the only open-source and fully functional industrial-grade deep learning platform, PaddlePaddle. Since its official open-source announcement in 2016, PaddlePaddle has undergone large-scale practices in the development of China's industrial intelligence and has achieved remarkable results in the rapid progress of the entire AI industry in 2019.
For example, at the first WAVE SUMMIT 2019 Deep Learning Developer Summit held in April, PaddlePaddle released a panoramic view and announced 11 new features and services. Just half a year later, at the "WAVE SUMMIT+" 2019 Deep Learning Developer Autumn Summit held in November, PaddlePaddle released and upgraded 21 new product directions, including four end-to-end development kits for industrial application scenarios, PaddlePaddle Master mode that combines pre-training and transfer learning by integrating data and knowledge, Paddle Lite 2.0 for edge inference, EasyDL Professional Edition, and cutting-edge technology tools. The usability has been greatly improved, significantly reducing the development threshold and providing stronger support for industrial applications. In October, at the 6th World Internet Conference held in Wuzhen, PaddlePaddle was recognized by the organizing committee and represented Baidu in winning the "World Internet Leading Technology Achievement" for the fourth time.
Currently, PaddlePaddle has served more than 1.5 million developers, with over 65,000 enterprise users on its customized training platform, and has released 169,000 models, becoming the foundation for supporting technological innovation and intelligent transformation in various industries. In addition to its impressive development speed, Baidu's PaddlePaddle also has an important significance. At a time when the AI industry in China is booming, the dependence on foreign open-source deep learning frameworks for some of China's core AI technologies has caused concerns among relevant professionals.
Will the AI technology industry, which is built on open-source deep learning platforms, fall into the "lack of chips and soul" situation again? The existence and rapid growth of PaddlePaddle have to some extent dispelled these concerns. PaddlePaddle has broken the over-reliance of Chinese AI developers on foreign open-source deep learning frameworks, allowing AI technology developers and users in China to be independent of foreign platforms and further cultivate a self-controlled AI development and application ecosystem. In addition, PaddlePaddle is also collaborating with domestic chip manufacturers such as Huawei and Cambricon, optimizing both software and hardware to build the infrastructure for China's AI ecosystem.
PaddlePaddle, the open-source deep learning platform developed by Baidu, has announced its compatibility with the NVIDIA EGX edge computing platform. PaddlePaddle's deep learning inference models can run directly on the EGX platform, providing an end-to-end solution for edge AI implementation. Allowing AI to be executed where data is generated, whether it is in factory facilities, retail stores, or warehouses. In the era of intelligent interconnection, a large number of machine sensors and IoT devices will continuously generate data that needs to be processed. The latency caused by cloud-based AI processing cannot meet the real-time production needs. The collaboration between AI and 5G will push AI computing from data centers to the edge, enabling fast response and real-time decision-making through high-performance edge computing. This is the real demand of industrial applications and an important direction for industry development. Baidu's PaddlePaddle has achieved algorithm optimization for GPU platforms. Deep learning inference models obtained through PaddlePaddle can run directly on the NVIDIA EGX edge computing platform.
With the increasing number of edge devices carrying AI models, there is a need for dynamic management and continuous iteration of models. PaddlePaddle's service platform, EasyDL and EasyEdge, provide a customized AI model training and service management solution for edge-cloud collaboration. EasyDL offers rich and leading pre-trained models, full-process support for data management, model training and iteration, and service deployment. Developers can efficiently build high-precision AI models based on EasyDL. Through EasyEdge, developers can quickly obtain edge computing acceleration models compatible with various chip platforms. The produced SDK can be directly deployed on the EGX platform, including NVIDIA Jetson embedded products.
Currently, the Baidu PaddlePaddle + NVIDIA Jetson solution has been implemented in various industries, bringing tangible efficiency improvements. For example, Wuhan Jingce, a leading company in the quality inspection industry in China's intelligent manufacturing field, often faces the challenge of maximizing production capacity in limited cleanroom areas, requiring small-sized, low-power, and fast processing detection equipment. By combining the technical advantages of PaddlePaddle's industrial applications and the computing power of the Jetson platform, the algorithm was quickly ported and optimized on the Jetson platform. The solution perfectly solved the above problems while meeting the requirements of algorithm performance and real-time processing, greatly increasing the company's production capacity.
Alphabet Q1 paid subscribers reached 350M, fueled by YouTube & Google One. YouTube ad revenue slightly missed at $9.88B amid Premium growth. Total Q1 revenue hit $109.9B, beating forecasts.
Anthropic releases Claude connectors for Adobe, Blender, and Ableton, enabling AI workflow automation and app control. It also sponsors the Blender Development Foundation.
Apple introduces an App Store model splitting 12-month plans into monthly installments. On iOS 26.4+, it lowers upfront costs without reducing total price, adding billing tracking.
Google launches Aadhaar digital IDs in Wallet for India, enabling secure QR/NFC verification. A phased partner rollout targets key sectors, while passport-based credentials expand globally.
YouTube tests "Ask YouTube," an AI tool blending text and video for complex queries. Initially for US Premium users, it supports interactive follow-ups, expanding Google's AI search.
Google grants US DoD classified AI access, joining OpenAI and xAI in defense deals amid Anthropic's dispute. Contractual safeguards against surveillance and lethal weapons remain unenforceable.
Mistral AI launches Workflows, a Temporal-powered orchestration platform for secure, Python-driven enterprise AI automation featuring human-in-the-loop approvals.