Realtime AI News
NVIDIA RTX Spark Debuts at Bilibili World: Laptops Can Run 120B-Parameter Models Locally
NVIDIA showcased the RTX Spark chip at Bilibili World, a super chip integrating CPU and GPU in one package. It enables laptops to run 120B-parameter large models locally, supporting personal AI agents, privacy offloading, and gaming.
NVIDIA publicly demonstrated the RTX Spark chip at Bilibili World 2026 for the first time. This chip features a CPU-GPU integrated package, designed for edge AI computing. According to QbitAI, RTX Spark enables laptops to run large language models with up to 120 billion parameters locally, without relying on cloud inference.
RTX Spark supports on-device personal AI agents, allowing users to use intelligent assistants without internet connectivity while processing sensitive data. It also includes a privacy offloading mechanism to ensure data security. Additionally, RTX Spark is compatible with mainstream games, providing high-performance graphics for gamers.
NVIDIA also introduced the Spark Pro developer edition, aimed at AI application developers. It offers richer development and debugging tools to accelerate local AI deployment. The Bilibili World debut marks a key step for RTX Spark from concept to mass production.
Industry observers believe RTX Spark will drive the adoption of AI PCs and lower the barrier for local deployment of large models. More consumer devices may integrate such chips in the future. NVIDIA further strengthens its leadership in AI hardware with this move.
For developers, the Spark Pro edition means they can develop and test large model applications directly on laptops, improving iteration efficiency. Gamers will also benefit from locally AI-enhanced graphics rendering.
Overall, the debut of RTX Spark is a significant event in AI hardware in 2026, and its future release and ecosystem development are worth watching.
Why it matters
RTX Spark will accelerate AI PC adoption, making local large model deployment possible, with significant implications for privacy and real-time AI applications.
Nearby Updates
All07/12, 08:21
UK Invites Anthropic to Expand London Operations, Discusses Dual Listing
The UK government has invited AI company Anthropic to expand its London operations and is exploring dual listing options with the company. The move underscores Britain’s push to attract leading AI firms as part of its post-Brexit technology strategy.
07/12, 10:58
Zhipu Founder Sends Internal Letter: No Short-Term Monetization, Full Commitment to AGI Heights
Zhipu AI founder sent an internal letter to employees declaring the company will not pursue short-term monetization but will instead aim for all-humanity heights in AGI research. The announcement comes as Hong Kong-listed AI model companies show diverging market performance, signaling different investor assessments of competing AI strategies.
07/12, 07:06
ByteDance's Seedance 2.5 Leaks: Native 30-Second AI Video Generation, API Expected July 16
TestingCatalog has shared exclusive early footage of ByteDance's Seedance 2.5, the company's next-generation AI video model capable of generating native 30-second clips in a single pass. The API release has been delayed to July 16 following ongoing rights holder negotiations, with pricing already appearing on the Jimeng platform in China.
07/12, 02:38
Alibaba Open-Sources Open Code Review: An AI-Powered Review Tool for the AI Coding Era
Alibaba has open-sourced Open Code Review, an AI-powered code review tool that garnered over 5,000 GitHub stars in its first week. The tool uses a hybrid architecture combining deterministic engineering rules with AI agents to handle the growing review burden created by AI-generated code.