# Advancements in mobile and hybrid AI technologies > Live situation record from CLSTR: https://clstr.news/situations/advancements-in-mobile-and-hybrid-ai-technologies > Updated: 2026-10-16T17:00:00.000Z. Sources: 18. Developments: 3. The technology sector is advancing strategies to integrate artificial intelligence across multiple devices through hybrid computing and cloud-based memory. Google DeepMind has introduced a ‘Private AI Compute’ design intended to maintain conversational context across devices by storing encrypted memory in the cloud while keeping decryption keys on local devices. In parallel, Qualcomm and Samsung Electronics are developing a ‘Hybrid AI’ ecosystem. This approach distributes workloads across smartphones, smartwatches, and smart glasses, using lightweight models for wearables and offloading complex tasks to larger devices or the cloud to manage battery and thermal constraints. Recent developments highlight a shift toward cost-efficiency and specialized hardware. Intel CEO Lip-Bu Tan noted that Chinese AI models, including DeepSeek, Kimi, and GLM, are gaining rapid adoption because their performance is “already good enough” while offering significantly lower token costs compared to Western counterparts. DeepSeek is reportedly preparing a second round of financing totaling approximately 100 billion yuan, which could bring its pre-money valuation to roughly 500 billion yuan. The company is also accelerating plans for an IPO on the STAR Market, aiming for a listing in 2027. Looking toward 2028, Qualcomm CEO Cristiano Amon indicated that industry demand is shifting toward smartphones capable of running large language models with at least 100 billion parameters. This shift suggests a move toward AI-native devices that function as autonomous agents. Leading AI companies are reportedly outlining hardware requirements to Qualcomm to facilitate this development, with industry speculation pointing toward OpenAI and Meta. OpenAI has reportedly integrated its hardware team, io Products, and is working with designer Jony Ive’s studio, LoveFrom, on new products. Meta currently maintains a presence in AI-integrated hardware through smart glasses. Achieving these goals requires overcoming significant technical bottlenecks, including massive memory requirements—potentially requiring 50GB of RAM—DRAM supply constraints, high power consumption, and thermal management issues. ## Claims - AI companies are requesting hardware capable of running 100-billion-parameter models continuously on smartphones by 2028. (corroborated by 6 sources) - Some AI companies are already beginning to develop their own smartphones. (corroborated by 5 sources) - Running a 100-billion-parameter model via 4-bit quantization would require approximately 50GB of high-speed system memory. (corroborated by 2 sources) - Chinese AI models like DeepSeek, Kimi, and GLM offer sufficient performance for enterprise use at lower token costs. (corroborated by 2 sources) - DeepSeek is reportedly nearing a 100 billion yuan funding round with a pre-money valuation of approximately 500 billion yuan. (single source) ## Timeline ### 2026-10-16: AI companies may develop proprietary smartphones by 2028 Major AI firms are reportedly working with chipmakers like Qualcomm to develop smartphones capable of running massive 100-billion-parameter AI models by 2028. 2 sources. https://clstr.news/cluster/ai-companies-may-develop-proprietary-smartphones-by-2028 ### 2026-10-12: Qualcomm CEO reveals AI companies aim for 100B-parameter smartphone models by 2028 Qualcomm CEO Cristiano Amon reports that AI firms aim to run 100B-parameter models on smartphones by 2028, while Intel's CEO notes the rapid rise of low-cost Chinese AI models like DeepSeek. 14 sources. https://clstr.news/cluster/qualcomm-ceo-targets-100b-parameter-ai-models-in-smartphones-by-2028 ### 2026-09-28: Tech leaders advance hybrid AI and cross-device memory technologies Google and Qualcomm are advancing AI integration through cloud-based memory and hybrid computing strategies across smartphones, wearables, and smart glasses. 2 sources. https://clstr.news/cluster/tech-leaders-advance-hybrid-ai-and-cross-device-memory-technologies --- Cite as: Advancements in mobile and hybrid AI technologies. CLSTR, https://clstr.news/situations/advancements-in-mobile-and-hybrid-ai-technologies