Shezhea NET is the capital layer for AGI connects Machine Learning with the Open Market by binding Quantitative Algorithms with Deep Learning. The vision behind Shezhea NET is to create a decentralized way for Artificial General Intelligence to profit from financial markets and become self-financing. The self-financing activity using the models developed by Shezhea NET in the financial sector will generate profits that will be automatically reinvested in the development of the AGI.
Microsoft stated that the speech AI model MAI Realtime 1 outperformed GPT Realtime 2.1 in public benchmarks and internal Microsoft product evaluations.
Microsoft unveiled MAI Cyber 2 Flash at the Ignite conference, with plans to integrate it into its security products.
According to TechFlow Research, a July 13 report from JPMorgan pointed out that Starbucks building its own AI tools to replace Microsoft and IBM software, Microsoft using its self-developed MAI to replace OpenAI and Anthropic models, and Meta developing cloud business to sell AI compute collectively illustrate a trend: the software profit pool is shifting downstream from the model layer. DigitalOcean's Q2 remaining performance obligations exceeded $800 million, a 10-fold year-over-year increase, with AI inference already accounting for a significant portion; over half of Cloudflare's requests originate from AI agents, and it launched a crawler paywall to open new revenue sources. Morgan Stanley believes model providers face pressure of being replaced, infrastructure layer demand remains strong but the structure is changing, enterprise customers' bargaining power is rising, and investment logic needs to shift from "model as winner" to "infrastructure and intermediary layers".