SHANGHAI, 11 September 2026 — On the evening of 10 September, Envision Greenwise Holdings Limited ("Envision Greenwise"; HKEX: 01783) announced that its wholly-owned subsidiary, Shanghai Yovole Cloud Computing Co., Ltd. ("Yovole Cloud Computing"), will purchase more than 100 high-performance servers from Shanghai Shunquan Technology Co., Ltd. ("Shanghai Shunquan") for a total consideration of RMB 380 million.

Just one month earlier, on 6 August, Envision Greenwise had disclosed that Yovole Cloud Computing would purchase more than 400 high-performance servers from Shanghai Shunquan for a total consideration of RMB 1.2884 billion. Placing another order after an interval of only one month not only underscores the strong customer demand flowing to Yovole Cloud Computing, but also validates the efficient execution of its "order-driven procurement" (sales-to-procurement) model — a clear signal that Yovole Cloud Computing's computing capacity is now scaling at an accelerated pace.

This latest announcement disclosing the additional purchase of more than 100 high-performance servers also reveals two highly critical pieces of information.

First, on the hardware side. The servers procured this time by Yovole Cloud Computing come with the following specifications: 8 processing units, approximately 17,000 CUDA cores, and 141 GB of memory. This configuration effectively pins down the GPU model — one of the most sought-after high-end chips in today's AI computing market. Thanks to the bulk order volume and the qualifications of the counterparty, the actual transaction price came in below the publicly quoted market level, broadly in line with the unit price paid in August when more than 400 units were purchased. In other words, within less than two months, Yovole Cloud Computing has locked in two batches totalling more than 500 high-performance server units of computing assets at below-market pricing.

Second, on the capacity and revenue side. Envision Greenwise disclosed directly in the announcement that the company's AI computing business is currently providing 1.5 MW of computing power to customers, with an additional 7.0 MW already contracted. The company expects that, on the premise that it is able to procure the relevant computing hardware and to enter into contracts with suitable customers, the computing capacity it can provide in the short term will increase to 16.0 MW.

This indicates that Envision Greenwise's AI computing business will continue to advance on a "lock in contracts first, then match hardware" basis: server procurement is initiated only after binding customer commitments have been obtained, with orders driving capacity expansion so as to minimise inventory risk and delivery mismatches. The 1.5 MW already in operation today is the clearest evidence that this model is working as intended.

The fact that short-term available capacity can grow to 16.0 MW implies the company is about to sign new contracts covering at least 7.5 MW of incremental capacity — and the 16 MW figure has already sketched out a deterministic return profile for the capital markets.

________________________________________

Revenue Outlook

Industry insiders forecast that, based on the GPU model of the high-performance servers purchased by Yovole Cloud Computing and the ultra-high rack-up rate generated by the company's "order-driven procurement" model, a 16 MW short-term computing capacity could generate nearly RMB 1 billion of AI computing revenue for Envision Greenwise in the second half of FY2027 (the six-month period from 30 September to 31 March).

RMB 1 billion of revenue in a single half-year already equals roughly 40% of Envision Greenwise's total revenue of HKD 2.462 billion reported for FY2026 (the 12 months ending 31 March). And these short-term 16 MW alone are expected to deliver at least approximately RMB 2 billion of revenue in FY2028. If the company's available computing capacity continues to expand thereafter, AI computing is expected to vault into Envision Greenwise's largest business segment next year, becoming the core engine driving the company's high-speed growth.