Forecast trend report by period
South Korean chipmaker braces for performance of four US tech giants
AI Data Center CapEx Guidance Signals Important Signal
Outlook is set to determine HBM and server DRAM orders
Samsung Electronics and SK Hynix are seen as potential beneficiaries.
South Korea’s semiconductor industry is bracing for gains from Alphabet, Microsoft, Metaplatforms and Amazon. Demand for Samsung Electronics and SK Hynix products, such as high-bandwidth memory (HBM) and server DRAM, could further increase if the four U.S. technology giants maintain the aggressive AI infrastructure investment plans they outlined earlier this year.
As investors focus on capital spending guidance, focus on pace of AI spending
Epic AI, which provides AI-based investment information services, announced that Alphabet, Microsoft, Meta, and Amazon will release their first-quarter results after the market closes on April 29, which investors refer to as “Super Day.” The main focus is on capital expenditure guidance for building AI data centers. These reports are being treated not just as quarterly earnings readings, but as important tests to determine if there is still room to run in the AI infrastructure investment cycle.
According to foreign media reports, the four companies will make a total of about $674 billion in capital investment this year. Google announced it would spend $175 billion to $185 billion on AI infrastructure, Microsoft $140 billion, Meta $115 billion to $135 billion, and Amazon announced more than $200 billion. This is more than 60% higher than last year’s total of $410 billion.
The industry view is that companies will struggle with a significant slowdown in investment in the short term. Competition in AI services has already expanded to include competition in data centers, power, and semiconductors. Shinhan Securities said hyperscalers have little room to cut capital spending, which could hurt valuations built on future growth.
Google balances cloud momentum with search advertising pressures
Consensus forecasts compiled by Epic AI project Alphabet’s first-quarter earnings per share to be $2.63 and revenue between $101 billion and $106.9 billion. For the last quarter, the company’s EPS was $2.82 and revenue was $113.8 billion, beating market expectations by 7.63% and 8.65%, respectively. This raised expectations for further unexpected results.
One key area is the growth of Google Cloud Platform (GCP). Some analysts have raised their expectations for GCP revenue this year to $84.8 billion, up 44% from a year ago. That’s because cloud usage is increasing due to the growing adoption of AI services, while enterprise AI workloads such as model training and inference are increasingly moving to cloud infrastructure.
Still, pressure on Google’s traditional revenue stream in search advertising is mounting. As AI-powered search capabilities become more prevalent, more users are getting answers directly on results pages in a “zero-click” format, reducing ad exposure. Research firm eMarketer predicts that Google’s advertising business will grow 11.9% this year, well below Meta’s forecast of 24.1%. The market will be watching to see if Google can strengthen its growth story through its cloud and AI infrastructure while quelling concerns about slowing search advertising momentum.
Microsoft’s Azure growth could settle AI spending debate
Microsoft’s Q1 consensus is for EPS of $4.07 and revenue of $72.8 billion, up 17.6% and 15.5% year-over-year. The most notable indicator is the growth rate of the company’s cloud service, Azure. Management previously expected Azure revenue growth to be between 37% and 38% at constant currency.
Microsoft is widely recognized as the first major technology company to commercialize AI services through its partnership with OpenAI. The company quickly added generative AI capabilities to Office, cloud products, and developer tools, helping it gain early leadership in the enterprise AI market. But the cost of AI infrastructure is also rising. When Microsoft reported its results in January, its stock price fell more than 5%, even though net income was up 60% from a year earlier. Investors focused more on concerns that spending was accelerating too quickly than strong performance.
Sentiment could change if Azure’s growth meets or exceeds the company’s guidance this quarter. Investors will likely interpret this as a sign that heavy spending on AI servers is turning into cloud revenue. A weaker-than-expected growth could renew questions about when Microsoft will be able to recoup its AI infrastructure investments.
Meta’s AI-driven ad profits face spending challenge
Meta’s Q1 consensus is projecting revenue of $55.4 billion and adjusted EPS of $6.72, up 33.91% and 27.33% year-over-year. Bank of America expects Meta to beat consensus on first-quarter sales of $56 billion and EPS of $7.44. The company also expects second-quarter revenue to be between $57.5 billion and $60.5 billion.
Meta’s growth engine is improved advertising efficiency. Monetization of short-form ads on Instagram Reels is stable and ad conversions are increasing due to AI-based recommendation and targeting technology. According to the report, the time spent watching Reels in the US has increased by more than 30% since Meta introduced its AI recommendation system. Longer user engagement increases ad inventory and improves the performance of personalized ads.
Meta also has significant AI infrastructure costs. Capital spending this year is expected to be between $115 billion and $135 billion. This reflects increased server investment for more advanced recommendation algorithms, in-house AI model development, and data center expansion. The company’s advertising business is generating enough cash to absorb short-term pressures, but investors will still scrutinize both the pace of spending growth and whether profitability can be maintained.
Amazon relies on AWS and advertising as dual growth engines
Amazon’s first-quarter consensus calls for revenue of $177.2 billion and adjusted EPS of $1.63. Sales are expected to increase 13-14% year-on-year. Last quarter, Amazon Web Services’ revenue rose 24% to $35.6 billion, and advertising revenue rose 23% to $21.3 billion, both of which beat market expectations.
AWS remains a core driver for Amazon, market participants said. Demand on AWS servers is also increasing as companies move their AI model training and inference work to the cloud. Morgan Stanley said that enterprise demand for AI computing is rapidly shifting to AWS. In other words, more and more companies implementing AI are choosing cloud infrastructure such as AWS rather than building their own data centers.
Advertising has also emerged as another growth pillar for Amazon. Advertising products built on purchase data collected from e-commerce platforms offer high profit margins. Goldman Sachs said improved advertising margins are acting as a second engine alongside AWS.
Amazon plans to spend about $200 billion in capital spending this year, an increase of 60% from a year ago, the highest on record. This figure includes investments in logistics automation and data centers, but the market is focused on how much investment will be made in AI infrastructure.
Will Samsung and SK Hynix’s memory orders continue to soar?
South Korean chipmakers are watching these profits closely as their big tech investment plans ultimately translate into memory orders. AI data centers require not only accelerators such as Nvidia GPUs, but also large amounts of HBM, large server DRAM, and enterprise SSDs. Samsung Electronics and SK Hynix are supplying those products.
In particular, HBM has become a core component of AI servers. GPUs are packaged and used for large-scale computing workloads. As AI models grow larger and inference demands increase, higher bandwidth and larger memory become more important. If big tech companies continue to spend on AI data centers, the medium- to long-term demand outlook for South Korean memory makers will also improve.
Samsung Electronics recorded preliminary sales of 133 trillion won and operating profit of 57.2 trillion won in the first quarter, marking the highest quarterly performance ever. DRAM contract prices rose 90% to 112% quarter over quarter, and NAND prices rose 80% to 93%. If big tech companies’ AI infrastructure plans meet or exceed expectations this fiscal year, it would strengthen the view that memory demand is unlikely to ease easily after the second quarter.
SK Hynix is widely recognized as the company most directly benefiting from the HBM boom. Operating profit margin for the first quarter was 72%, higher than Samsung Electronics’ 43%, Nvidia’s 65% and Taiwan Semiconductor Manufacturing’s 58.1%. Additionally, HBM market share was at the top of the industry at approximately 59%. That’s why investors are closely linking SK Hynix’s earnings structure to the large technology demand for AI accelerators.
“These results from big tech companies will show whether enthusiasm for AI spending is translating into actual industrial demand,” said an industry source. Epic AI said investors should closely monitor whether the earnings report reaffirms the structural growth story across the AI value chain or rather triggers a near-term correction.
Hong Min-sung Hankyoung.com reporter mshong@hankyung.com
