Jinke's AI unit takes 59% control of Beijing's Baihai in first deal
What's the deal? Jinke Real Estate GroupDealroom has a profile for this one. Try Dealroom → has acquired control of Beijing-based Baihai Technology, announcing the deal on November 17. Through its wholly owned subsidiary Chongqing Xingshi Zhisuan, Jinke bought roughly 40% of Baihai and secured voting control over a further 19%, reaching about 59% of voting rights. The deal value was not disclosed.
What does Baihai do? It positions itself as a next-generation AI software platform provider, selling compute services, AI development and fine-tuning tools, and private models through an "AI factory" model. Its core technology is Wind, a self-developed framework for scheduling compute resources across clusters. Baihai's platform is compatible with domestic servers from Huawei Ascend, Moore Threads, and others.
Why now? Jinke recently completed a judicial restructuring and is repositioning as a "technology-driven real estate operator." Xingshi Zhisuan, set up this year with former JD.com technology adviser Li Daxue as chairman, is extending from building and operating spaces to operating compute. This is its first acquisition.
What's the endgame? Jinke frames the move as a key step toward an "AI-plus" operations and investment platform, aiming to channel its real estate asset-management skills into the AI sector. China's compute demand is climbing sharply: the National Data Administration reported average daily token calls topped 140 trillion in March 2026, up more than 1,000 times from early 2024.
What could go wrong? This looks more like a strategic trial than a proven growth engine. China's intelligent compute infrastructure investment reportedly hit US$100.2B in the first three quarters of 2026, but most projects are not yet operational — underlining the gap between building capacity and earning stable revenue. Success hinges on Baihai's customers, revenue, and profitability, and on whether a post-restructuring Jinke can fund the new business.
The signal: As compute demand surges alongside cost pressure, scheduling efficiency is becoming the battleground. The industry's competitive focus is shifting from "model capability" to "compute scheduling and cost engineering" — and Jinke is betting a traditional property developer can find a second act there.
Read more: sohu.com
Image credit: "Virginia Tech - data center" by cbowns, licensed under CC BY-SA 2.0