Big Tech’s $674 Billion AI Spending Test Puts Samsung, SK Hynix on Edge Ahead of ‘Super Day’

Korean chipmakers brace for results from four US tech giants

AI data-center capex guidance is the key signal

Outlook is set to shape HBM and server DRAM orders

Samsung Electronics, SK Hynix seen as potential beneficiaries

South Korea’s semiconductor industry is bracing for earnings from Alphabet, Microsoft, Meta Platforms and Amazon. If the four US technology giants maintain the aggressive AI infrastructure spending plans they outlined at the start of the year, demand for Samsung Electronics Co. and SK Hynix Inc. products such as high-bandwidth memory, or HBM, and server DRAM could get another boost.

AI Spending Pace in Focus as Investors Watch Capex Guidance

Epic AI, an AI-based investment information service, said Alphabet, Microsoft, Meta and Amazon will report first-quarter results after the market closes on April 29, in what investors are calling a “super day.” The main focus is capital-expenditure guidance for AI data-center buildouts. More than a read on quarterly earnings, the reports are being treated as a key test of whether the AI infrastructure investment cycle still has room to run.

Based on foreign media reports, the four companies’ combined capital spending this year totals about $674 billion. Google has said it will spend $175 billion to $185 billion, Microsoft $140 billion, Meta $115 billion to $135 billion and Amazon more than $200 billion on AI infrastructure. That is more than 60% above last year’s combined $410 billion.

The industry’s view is that the companies will struggle to sharply slow investment in the near term. Competition in AI services has already spread into a race for data centers, power and semiconductors. Shinhan Securities said hyperscalers have little room to reduce capital spending because that could undermine valuations built on future growth.

Google Balances Cloud Momentum Against Search Ad Pressure

Consensus estimates compiled by Epic AI put Alphabet’s first-quarter earnings per share at $2.63, with revenue of $101 billion to $106.9 billion. In the previous quarter, the company posted EPS of $2.82 and revenue of $113.8 billion, beating market expectations by 7.63% and 8.65%, respectively. That has fueled expectations for another earnings surprise.

One key area is growth at Google Cloud Platform, or GCP. Some analysts have raised their estimate for this year’s GCP revenue to $84.8 billion, up 44% from a year earlier. The reasoning is that wider adoption of AI services is driving cloud usage higher, while corporate AI workloads such as model training and inference are increasingly moving onto cloud infrastructure.

Still, pressure is building on Google’s traditional profit engine in search advertising. As AI-powered search features spread, more users are getting answers directly on results pages in a “zero-click” format, reducing opportunities for ad exposure. Research firm eMarketer projects Google’s advertising business will grow 11.9% this year, far below Meta’s expected 24.1%. The market will be watching whether Google can reinforce its growth story through cloud and AI infrastructure while containing worries about softer search-ad momentum.

Microsoft’s Azure Growth Could Settle the AI Spending Debate

Microsoft’s first-quarter consensus stands at EPS of $4.07 and revenue of $72.8 billion, up 17.6% and 15.5% from a year earlier. The most closely watched metric is the growth rate of Azure, the company’s cloud service. Management previously guided for Azure revenue growth of 37% to 38% in constant currency.

Microsoft is widely viewed as the big tech company that moved first to commercialize AI services through its partnership with OpenAI. It quickly added generative AI features to Office, cloud offerings and developer tools, helping it seize an early lead in the enterprise AI market. But the cost of AI infrastructure has also climbed. When Microsoft reported earnings in January, the stock fell more than 5% even though net income jumped 60% from a year earlier. Investors focused less on the strong results than on concern that spending was accelerating too quickly.

If Azure growth meets or tops the company’s guidance this quarter, sentiment could shift. Investors would read that as a sign that heavy spending on AI servers is translating into cloud revenue. If growth misses expectations, questions about when Microsoft can recoup its AI infrastructure investment could intensify again.

Meta’s AI-Driven Ad Gains Face a Spending Test

Meta’s first-quarter consensus calls for revenue of $55.4 billion and adjusted EPS of $6.72, up 33.91% and 27.33% from a year earlier. Bank of America expects Meta to beat consensus with first-quarter revenue of $56 billion and EPS of $7.44. It also projects second-quarter revenue guidance of $57.5 billion to $60.5 billion.

Meta’s growth engine is improving advertising efficiency. Monetization of short-form ads on Instagram Reels has stabilized, while AI-based recommendation and targeting technology is lifting ad conversion. After Meta introduced its AI recommendation system, time spent watching Reels in the US rose more than 30%, according to the report. Longer user engagement can increase ad inventory and improve the performance of personalized ads.

Meta is also carrying a heavy AI infrastructure bill. Its capital spending this year is projected at $115 billion to $135 billion. That reflects rising server investment for more advanced recommendation algorithms, in-house AI model development and data-center expansion. The company’s advertising business generates enough cash to absorb near-term pressure, but investors are still set to scrutinize both the pace of spending growth and whether profitability holds up.

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. Revenue is projected to rise 13% to 14% from a year earlier. In the previous quarter, Amazon Web Services revenue increased 24% to $35.6 billion and advertising revenue rose 23% to $21.3 billion, both topping market expectations.

AWS remains Amazon’s core driver, according to market observers. As companies shift AI model training and inference work to the cloud, demand for AWS servers is rising as well. Morgan Stanley said corporate demand for AI computing is moving quickly toward AWS. In other words, more companies adopting AI are choosing cloud infrastructure such as AWS instead of building their own data centers.

Advertising has also emerged as another growth pillar for Amazon. Its ad products, built on purchase data gathered from its e-commerce platform, carry high margins. Goldman Sachs said improving ad margins are serving as a second engine alongside AWS.

Amazon has guided for capital spending of about $200 billion this year, up 60% from a year earlier and the highest on record. The figure includes logistics automation and data-center investment, but the market’s focus is on how much will go to AI infrastructure.

Will Memory Orders Keep Surging for Samsung and SK Hynix?

Korean chipmakers are watching these earnings closely because big tech investment plans ultimately feed through to memory orders. AI data centers require not only accelerators such as Nvidia GPUs, but also large volumes of HBM, high-capacity server DRAM and enterprise SSDs. Samsung Electronics and SK Hynix supply those products.

HBM in particular has become a core component in AI servers. It is packaged with GPUs and used for large-scale computing workloads. As AI models grow larger and inference demand increases, memory with higher bandwidth and larger capacity becomes more essential. If big tech companies keep spending on AI data centers, the medium- to long-term demand outlook for Korean memory makers will improve as well.

Samsung Electronics posted preliminary first-quarter revenue of 133 trillion won and operating profit of 57.2 trillion won, its highest quarterly results on record. DRAM contract prices jumped 90% to 112% from the previous quarter, while NAND prices rose 80% to 93%. If big tech companies’ AI infrastructure plans meet or exceed expectations this earnings season, that would strengthen the view that memory demand is unlikely to weaken easily after the second quarter.

SK Hynix is widely seen as the company benefiting most directly from the HBM boom. Its first-quarter operating margin was 72%, above Samsung Electronics’ 43%, Nvidia’s 65% and Taiwan Semiconductor Manufacturing Co.’s 58.1%. It also held the industry’s top HBM market share at about 59%. That is why investors link SK Hynix’s earnings structure so closely to big tech demand for AI accelerators.

“This round of big tech earnings will show whether enthusiasm for AI spending is translating into real industrial demand,” an industry official said. Epic AI said investors need to watch closely whether the earnings reports reaffirm the structural growth narrative across the AI value chain or instead provide a trigger for a short-term correction.

Hong Min-sung, Hankyung.com reporter mshong@hankyung.com