AI computing power requirements clash again with cash return logic! Alphabet (GOOGL.US) invests $15 billion to expand AI data center in Malaysia
The Zhitong Finance App learned that Google's parent company Alphabet (GOOGL.US) said it plans to invest 15 billion US dollars to expand AI data centers and traditional cloud computing-related businesses in Malaysia, highlighting that the unprecedented boom in artificial intelligence applications is driving the accelerated growth in demand for AI computing power resources and digital infrastructure.
According to a statement released on Wednesday, Google, a global cloud computing and technology leader under Alphabet, will build a new computing center and invest in local water security, power generation facilities, and training for Malaysian residents. The company said these investments are planned to be implemented in 2027 and 2028.
Google has joined the ranks of global technology companies expanding the size of the Asia-Pacific market in Malaysia. Malaysia has attracted these tech companies with its relatively low comprehensive operating costs, reliable power and energy infrastructure, and proximity to international business hubs such as Singapore.
According to information, Malaysia has become a major data center hub in Southeast Asia, attracting companies such as Oracle, Amazon, Alibaba Group, and ByteDance. As the number of users of data-intensive AI services continues to grow in Asia and the Pacific, these companies need larger storage capacity and AI computing processing capabilities.
Google said that the construction phase of the new park is expected to add 1.1 billion US dollars to the Malaysian economy in the next two years and support more than 37,000 local jobs every year. Once operational, the data center is expected to support more than 9,000 local jobs each year, including technicians, mechanical and electrical engineers, and logistics and security managers.
Ruth Borat, chief investment officer of Google's parent company Alphabet, said at an event in Kuala Lumpur that the investment “reflects our confidence in the strength of the Malaysian leadership, skilled workforce, and enthusiasm for advancing digital transformation.” Malaysian Prime Minister Anwar Ibrahim attended the event.
Anwar said that Google's investment respects Malaysia's resources and also shows trust in the Malaysian people. “Our country has a stable environment, talent and ambition, and is capable of becoming a leading digital power in Southeast Asia,” he said in an interview.
Earlier this week, Google announced a $160 million investment to build shared water infrastructure in Negeri Sembilan state south of Kuala Lumpur to support planned data center operations while also serving the local community.
The stock and bond market is re-evaluating the AI trading theme, while Google is determined to expand AI computing power infrastructure
Google, a subsidiary of Alphabet, plans to invest US$15 billion from 2027 to 2028 to expand its data center and cloud business in Malaysia, and simultaneously invest in water resources security, power generation facilities, and talent training. Close to Singapore, relatively low cost, and reliable infrastructure, makes Malaysia an important fulcrum for Google to meet the digitalization and AI needs of Southeast Asia. This huge investment by Google, a subsidiary of Alphabet, can be described as revealing the clearest differences in the current market — technology companies are still building production capacity for future needs, yet investors are increasingly concerned about when construction expenses can be converted into cash returns sufficient to cover capital costs.
Global AI transactions are experiencing intense repricing between growing demand and return on capital — CoreWeave shares are under pressure, CDS implied default risks are rising, credit protection costs for Meta, Amazon, and Alphabet are also rising, and investors are beginning to demand higher compensation for huge financing needs and long payback cycles.
Market adjustments have been made through US technology stocks and the Asian chip supply chain. In October, the Nasdaq 100 index fell by about 6.7%, while SK Hynix and Samsung Electronics fell by about 20.2% and 16.2%, respectively. Despite the recent rebound, the Philadelphia Semiconductor Index and South Korea's KOSPI are still about 28% and 38% lower than the June closing high, respectively, and are in the technical bear market range, based on Wednesday's closing point.
The impact at the individual stock level was even more severe: Alphabet A shares fell about 12.6% on October 29, the biggest one-day decline since listing; Intel fell 28.58% on October 30, the biggest one-day decline since 1974. The real money pricing in these markets all seems to be highlighted. As AI financing costs continue to rise, the growing demand for AI computing power is no longer sufficient to support the valuations previously given by the market alone.
The credit market has further amplified this prudent pricing. CoreWeave's stock price fell by a cumulative total of about 37% in October; in mid-October, the yield on some corporate bonds once rose to 15% to 16%, and the cumulative default probability implied by the five-year CDS price exceeded 60%. This probability includes a risk premium and relies on a recovery rate assumption. It cannot be directly understood as predicting the actual probability of default, but it is sufficient to show that investors are demanding a significant increase in risk compensation. The cost of default protection for Meta, Amazon, and Alphabet, the parent company of Google, is also rising, reflecting the combined pressure of large-scale financing supply, capital expenditure recovery cycles, and hedging demand.
The interest rate environment makes this competition more expensive. Although the yield on US 10-year Treasury bonds fell back to about 4.90% during the Asian session on November 4, the global AI financing cost benchmark can be described as still at an all-time high. For new cloud companies such as CoreWeave that rely on external financing, construction payments occur before, and electricity delivery and customer repayment occur after; widening financing spreads, accelerated equipment updates, or project delays will reduce the cash buffer. Google has a wider range of revenue sources, but the additional computing power must also prove that future operating returns can cover construction investment, depreciation, energy, and financing costs.
AI orders prove demand, and water and electricity delivery and actual capacity utilization determine returns
The expansion of Google's parent company Alphabet has actual business support. In the third quarter, Google's cloud computing business — that is, Google Cloud's revenue reached US$30.7 billion, up 103% year over year; the growth came from corporate AI infrastructure, AI solutions, and core cloud services, and TPU system sales also contributed substantially. By the end of September, Google Cloud's contract orders for which revenue had not been confirmed reached $56.8 billion. Orders increase future revenue visibility, but whether the company can deliver on time and when customers use services still determine how quickly revenue is realized.
However, Google's parent company's cash flow indicators explain why the market was not fully convinced by rapid growth: operating cash flow in the third quarter was 42.290 billion US dollars, lower than capital expenditure of 51,355 billion US dollars, and free cash flow was negative 9.065 billion US dollars, which was negative for the second consecutive quarter; free cash flow for the past 12 months was still positive at 19.747 billion US dollars. The company maintains the 2026 capital expenditure guideline of US$195-205 billion and is expected to increase significantly in 2027. Investment in Malaysia means that expansion continues to advance, and it also means that cash returns will still have to wait for production capacity to be put into operation and utilization to increase.
The project focus of the newly announced 15 billion dollar large-scale AI data center investment is to build a computer room together with public infrastructure. Google has announced an investment of 160 million US dollars to build shared water facilities in Negeri Sembilan, and has promised to cover all of the park's electricity consumption and direct grid access costs; media also reported that the expansion plan will add more than 1 gigawatt of clean power generation capacity to the local power grid.
The positive meaning behind Alphabet's investment plan for the tech giant's AI ambitions is that the AI industry chain opportunities are being extended at an accelerated pace to power supply, water treatment, and engineering construction. Corresponding to this are large-scale AI data centers that can supply electricity and water on schedule, operate stably and control unit computing power costs, and achieve efficient energy consumption, making it easier to turn demand into profit; the scale of AI chip procurement is only part of the project economy.
Wall Street strategists' unanimous bullish views after the collapse of technology stocks are also revolving around this time gap between investment and return. Deutsche Bank said that the market is too concerned about the cost of AI financing and is recommending overallocation of technology stocks; Mike Wilson, a senior strategist at Morgan Stanley, believes that after the uncertainty of financing, liquidity, and elections is digested, the market is expected to have buying opportunities. Pershing Plaza, where Ackman is at the helm, pointed out that analysts have fully taken into account current construction costs, but have placed too low expectations on future increases in revenue, profits, and free cash flow.
The growth already seen in 2026 is due in large part to investment from 2024-2025. Google's Malaysia plan undoubtedly provides a new investment basis for determining AI application demand and the long-term growth prospects of AI computing power infrastructure. The final verification still depends on delivery progress, customer usage, and return on capital.