A company with $46.5 billion in net cash on its balance sheet has chosen to raise HK$80 billion from the market. This decision, at first glance, appears contradictory, but a deeper look reveals a deliberate strategic move that goes beyond mere capital needs.
This fundraising effort is best understood as a "confidence policy" written to the market. Rather than signaling financial distress, it represents a proactive move to secure ammunition for AI computing power ahead of time. For A-shares, this serves as a shot in the arm, but it is by no means a signal for blind buying. The judgment here is that as industry giants vote with their real money, the A-share market will only reward hardcore suppliers with actual orders.
The logic becomes clear when dissecting the details. Alibaba (NYSE: BABA) holds approximately $46.5 billion in net cash after excluding long-term debt, yet it opted for equity financing instead of tapping into its cash reserves, committing 100% of the proceeds to AI. This is a statement to investors that the company itself believes in the returns from AI computing power. The subscription was oversubscribed within an hour, with sovereign funds showing long-term interest, indicating the market also recognizes this logic. While share dilution may suppress short-term sentiment, channeling funds into AI is a long-term positive.
This move aligns with a broader industry trend of accelerated capital expenditure in AI. In June 2026, Google's parent company Alphabet raised $84.75 billion through a stock issuance for AI infrastructure. In July, Zhipu completed a roughly HK$31.4 billion placement. By August, Alibaba's HK$80 billion followed, demonstrating that top players have shifted from verbal commitments to tangible financial backing. What's more significant is the shift in narrative from "stories" to "payback periods."
During the earnings call, CEO Wu Yongming stated that based on current average gross margins of AI products, capital expenditure could be recouped within three years. With improved gross margins and a higher proportion of self-developed chips, this cycle could be compressed to 2.5 years. CFO Xu Hong provided supporting data, citing that V100 chips purchased in 2018 and A100 chips from 2020 are still operating near full capacity. This suggests that the actual lifespan of AI computing assets is significantly longer than theoretical depreciation periods, meaning machines can continue generating positive cash flow after the three-year payback period. This focus on payback periods signals that AI computing power is being pulled into a phase of performance delivery.
Where to start
Does Alibaba's investment truly benefit the entire A-share computing power chain? The HK$80 billion raised will primarily target AI infrastructure. The three-year, 380 billion yuan investment plan announced in February 2025 has seen cumulative investment of 190 billion yuan by June 2026, with progress on track. Midstream sectors like IDC, servers, optical modules, and liquid cooling could all receive incremental orders. However, a closer look at Alibaba's financials reveals a quarterly capital expenditure of 67.676 billion yuan, a 75% year-over-year increase, and free cash flow net outflow expanding to 44.67 billion yuan. While spending is aggressive, the key question is whether orders will translate into revenue for specific A-share companies.
Market expectations that increased spending by giants equates to broad sector gains are misguided. Suppliers with major tech company certifications and proven delivery records will secure orders; others merely riding on the trend will be identified and left behind by capital. Industry participants face a common dilemma: AI computing and optical communications have rallied repeatedly, making it hard to decide whether to chase the rally or risk missing out. The answer lies not in whether to chase, but in which segment of the chain to focus on. The certainty of the optical module sector may be overstated, as the 1.6T volume ramp-up is accompanied by persistent price war shadows. While everyone focuses on how much Alibaba raised, the real variable is the placement discount and lock-up period. Lower discounts and longer lock-ups indicate stronger endorsement from long-term funds like sovereign wealth funds, translating into sustained three-year order support for the A-share computing power chain. Conversely, higher discounts and shorter lock-ups signal institutional cooling.
Why just a few players?
The highest certainty lies with companies already embedded in Alibaba's supply chain with verifiable orders. Zhongji Innolight's 1.6T optical modules hold a prime position, but its TTM price-to-earnings ratio sits around 74 times, reflecting high expectations. Foxconn Industrial Internet's AI server utilization acts as a barometer for the sector, with a TTM P/E of about 26 times, notably below the industry median, indicating relatively restrained valuation. Cambricon is in the sample validation window for domestic training chips, with a dynamic P/E exceeding 140 times, purely expectation-driven, where yield rates remain the critical hurdle. Zhongji Innolight's strength in optical module volume is also its weakness due to price pressure on margins; Foxconn Industrial Internet's thin foundry margins serve as a safety cushion with low valuation and stable orders. Envicool in liquid cooling leverages its barriers to secure profit margins, trading at a TTM P/E of around 143 times, which is not cheap. Runze Technology in the IDC sector sees its order flow determined by cabinet delivery pace, with a TTM P/E of about 20 times, below the industry median and relatively reasonable. Suppliers know that major tech orders may look attractive, but payment terms and resource allocation are never evenly distributed. Three key hurdles remain: external constraints on advanced GPUs and HBM, ongoing price wars in optical modules, and intense competition in servers and liquid cooling. The opportunity is real but belongs only to a select few with certifications and orders.
Final assessment
The signal value of Alibaba's HK$80 billion fundraising outweighs the capital itself. Opportunities exist for leading computing infrastructure suppliers already in the supply chain with confirmed orders, who can leverage a three-year order base. Risks include the "sucking effect" suppressing the broader market, the tendency for widely anticipated positives to open high and close low, and corrections for fake AI concept stocks. Key indicators to track are the placement discount and lock-up period announcements, which reveal institutional sentiment, and the pace of Alibaba's capital expenditure deployment.