AI Roundup — July 21, 2026
Gritt Exits Stealth with $34M to Deploy Robots on Construction Sites
Construction robotics startup Gritt has emerged from stealth with $34 million in funding, according to TechCrunch. The company's stated goal is to automate some of the most demanding physical tasks found on construction sites, with an initial focus on building solar plants. Per the report, Gritt's longer-term ambition extends beyond solar infrastructure, positioning the platform as a general-purpose solution for construction site automation. The funding signals continued investor interest in applying robotics to physically intensive industries where labor demands are high.
Google Developing a New AI Chip to Improve Gemini Efficiency
Alphabet, Google's parent company, is reportedly developing a new custom chip specifically designed to improve the operational efficiency of its Gemini AI models, TechCrunch reports. Details about the chip's architecture or a release timeline have not been publicly confirmed, but the effort reflects a broader industry trend of AI companies investing in purpose-built silicon to reduce the computational costs associated with running large language models. Google has previously developed its own Tensor Processing Units (TPUs) for AI workloads, and this new chip appears to continue that in-house hardware strategy with Gemini as a primary target.
MCP Protocol Receives Usability Updates
The Model Context Protocol (MCP), described by TechCrunch as one of AI's most important integration protocols, is receiving updates aimed at making it easier to work with. According to the report, the key change involves adopting a looser, "stateless" approach to session ID handling on the server side. TechCrunch notes this brings MCP's behavior closer to how most standard websites already manage sessions, potentially lowering the implementation barrier for developers building tools and integrations on top of the protocol. MCP has gained traction as a standard for connecting AI models to external data sources and tools.
These developments span hardware, robotics, and developer infrastructure — areas that continue to see active investment and engineering progress across the AI industry. More details on each story are available via the linked TechCrunch coverage.