Key Performance Period

Iran's strikes have paused again, oil prices have retreated, and risk appetite has rebounded once more. Although fuel costs remain high, Cathay Pacific's performance in the first half of the year has been impressive, with its stock price reaching new highs, primarily because the company can pass on the increase in oil prices and capture passenger flow from Middle Eastern hubs, while its dividend yield is also competitive with banks and other dividend stocks.

This week, U.S. stocks enter a period of intensive earnings disclosures, with the seven major tech giants and core companies in chips and storage set to release their results.

Among them, although Apple faces uncertainty regarding its AI strategy, its ability to raise prices to pass on cost pressures, continuous market share growth, the September iPhone launch catalyst, the introduction of the foldable screen in C4Q, and the rollout of Apple Intelligence in China have kept its stock price at a new high. Of course, while the replacement cycle is expected to start, the valuation is not cheap, and the space driven solely by earnings upgrades should be limited.

Additionally, the capital expenditures (capex) and AI business statements from Microsoft, Meta, Arm, Qualcomm, and Amazon will influence market sentiment towards AI hardware.

Microsoft's overall expectations are relatively low, with Azure expected to accelerate for the second consecutive quarter. Although Copilot is still in its early stages, its penetration rate is also accelerating. The risks of AI replacing Office and rising capex have mostly been priced in, but there are concerns that AI depreciation may eat into profits more than expected. The ideal scenario is that earnings can prove that the cycle of capex converting into revenue growth is viable; Meta has more potential catalysts, depending on whether it can provide a specific revenue timeline for Neocloud/MaaS and whether its cash burn rate slows down.

Currently, compared to the strong capex intensity overseas, domestic large cloud vendors have better flexibility in capital expenditures.

Starting from July 31, Alibaba Cloud will switch to a new billing model in batches, reducing the unit price for pay-as-you-go billing, focusing on lowering the operating costs of users' AI batch tasks, with the core aim of exchanging price for volume to stimulate more demand for AI computing power.

In contrast, Tencent officially began charging for the usage of its cloud database MySQL local disk instances that exceed the purchased storage space starting August 1, only charging for the portion that exceeds the purchased quota, directly boosting revenue and forcing users to plan their resource capacity reasonably.

The strategies of the two companies are clearly contrasted: Alibaba Cloud is pursuing a price-for-volume approach to compete in the AI computing power market; Tencent is directly commercializing the previously uncharged usage scenarios. However, both share the common question of when capex investment will translate into revenue/profit growth.

On a positive note, the rapid decline in costs for domestic large models may lower the threshold for enterprises to deploy AI applications. If AI agents and enterprise-level applications begin to be deployed on a large scale, the rapid growth in inference demand may provide unexpected elasticity in revenue growth for domestic cloud vendors.

Regardless, this week's capex/cash flow combination statements from U.S. stocks will provide direction for the valuation anchoring of Hong Kong's tech stocks