🔍 Read the full analysis: Is The AI Boom Holding Its Pace? Six Numbers To Watch on ThorstenMeyerAI.com
Get business pricing on tech for your team
- Business-only prices and quantity discounts
- Tax-exempt purchasing
- Multiple users, one account, clear invoices
TL;DR
SK hynix reported preliminary third-quarter operating profit of 83.3 trillion won, 7.3 times its year-earlier result, while China’s STAR50 index has fallen 40% since July. Microsoft and Alphabet are due to report on October 28, followed by Amazon on October 29, putting investment, cloud growth and cash flow in focus. The figures point to strong chip demand but do not yet settle whether spending by major technology companies will produce sufficient returns.
SK hynix reported preliminary third-quarter operating profit of 83.3 trillion won on Wednesday, 7.3 times its result a year earlier, as investors await quarterly results from Microsoft and Alphabet and, on Thursday, Amazon. The contrast with a 40% fall in China’s STAR50 index since July captures the uneven AI market: chip demand is generating record profits, but the scale and payoff of technology companies’ spending remain under scrutiny.
SK hynix, a major supplier of high-bandwidth memory used alongside AI processors, said third-quarter revenue reached 104.6 trillion won, up 328% from a year earlier, and operating profit was 83.3 trillion won, up from 11.4 trillion won. The company’s release was preliminary and unaudited. Its reported operating margin was 80%. Net profit, however, fell 58% from the previous quarter to 39.8 trillion won, which the source account attributed to non-operating items.
The company said long-term agreements with 14 major customers cover more than half of its planned capacity in 2027 and 2028. SK hynix expects demand to exceed supply through 2030; that is the company’s forecast, not an established outcome. It also announced a 14 trillion won dividend and a 20 trillion won share buyback this quarter. The source reports a plan to return about 100 trillion won to shareholders in 2026, while noting that half of the buyback is for employee compensation and is not counted by the company as a shareholder return.
The wider earnings picture is strong but concentrated. FactSet estimated S&P 500 third-quarter earnings growth at 29.5% as of October 23. In a separate calculation dated October 19, FactSet said removing Micron and Nvidia would reduce the index’s estimated growth rate from 29.0% to 20.0%. These are estimates, and several major companies had not reported when the figures were published.
Infographic 01 · The boom in numbers
Record profits, narrow growth, rising bills.
Six numbers from company releases, FactSet and the Census Bureau, before Microsoft, Alphabet and Amazon report.
Who carries the earnings boom
Q3 2026 earnings growth, year over year. Blended: reported results plus estimates.
The Wall Street Journal (Dow Jones Market Data) puts the two chipmakers at 34% of the expected growth. Top five contributors per FactSet: Micron, Nvidia, Meta, EchoStar, Boeing.
What the buyers plan to spend in 2026
Capital spending guidance, US dollars. Bars show the top of each range.
Q3 consensus: Alphabet free cash flow −$2.3B on $55.2B capex; Amazon −$9.9B on $58.7B. Definitions differ slightly (Microsoft includes finance leases).
Shovel seller
- Revenue 104.6 trillion won, +328%
- Long-term deals cover >50% of 2027–28 capacity
- About 100 trillion won back to shareholders in 2026
Real economy
- Core capital goods orders +10.9% this year
- Computer orders +20.3% this year
- Census data do not label AI; semiconductors excluded from orders
Debt nerves
- Cost to insure Meta’s debt +177% this year, Amazon’s +129%
- Oracle: $117B long-term debt, rated BBB-
- Off-balance-sheet leases don’t show up as debt (S&P)
Sources: SK hynix 3Q26 release, 28 Oct 2026 (preliminary); FactSet, 19 Oct 2026; Alphabet Q2 call, 22 Jul 2026; Amazon via CNBC, 30 Jul 2026; Microsoft via CNBC, 29 Jul 2026; Meta Q3 release, 21 Oct 2026; FactSet consensus via CNBC, 27 Oct 2026; US Census Bureau, advance durable goods report, 27 Oct 2026. Capex sum and bar scaling: Thorsten Meyer.
Powered by Thorsten Meyer AI
Testing AI Spending Against Returns
The numbers matter because AI investment is now large enough to affect both company cash flows and broader earnings forecasts. Alphabet, Amazon, Microsoft and Meta are expected to spend a combined roughly $725 billion to $745 billion on capital projects in 2026, a total calculated from their guidance in the source account. Investors will compare that outlay with cloud growth, free cash flow and management’s plans for capacity.
Demand for equipment is already showing up beyond technology-company results. U.S. shipments of core capital goods rose 1.2% in September, while orders for computers and related products were up 20.3% this year, according to the Census Bureau figures cited in the source. Those measures provide evidence of investment activity, but do not show on their own how quickly AI-related spending will earn a return.
The issue is also concentration. If a small number of chipmakers account for a large share of expected earnings growth, overall market forecasts may depend heavily on their results. At the same time, the source says analysts expect Alphabet and Amazon to report negative free cash flow again for the third quarter. That expectation is not a reported result; their upcoming filings will provide a more current test.
Products Worth Considering
Bose S1 Pro+ Portable PA Speaker
Bose TV Speaker (2nd Gen) Compact Soundbar with Bluetooth and HDMI-ARC, Black
Amazon Kindle Scribe Colorsoft 64GB (1st Generation)
As an affiliate, we earn on qualifying purchases.
Chip Profits, Cloud Costs and China
The AI investment cycle links chip and memory suppliers to cloud companies building data centers. SK hynix’s high-bandwidth memory is used with AI processors, making its sales and margins one measure of demand for computing infrastructure. The company’s results show how strongly that demand has benefited a supplier, but they do not establish whether cloud providers will earn enough from AI services to cover their own investment.
The market signals are not uniform. Reuters reported that China’s STAR50 index, which includes AI and chip companies, had fallen about 40% in under four months as of October 26, erasing an estimated $480 billion in value. The index was still trading at about 100 times past earnings, according to the cited report. Those figures describe that market and period; they are not a direct measure of U.S. technology-company performance.
Corporate earnings measures also require care. FactSet’s broader 29.5% growth estimate, published October 23, and its 29.0% figure in the October 19 comparison come from different dated snapshots. The source also notes that some companies’ expected growth reflects one-time charges in the year-earlier quarter. Estimates can change as more companies report.
“Demand to exceed supply through 2030.”
— SK hynix
Payback and Supply Remain Unsettled
The main open question is whether revenue from AI services will grow enough to justify the data-center investment being made now. The source account says analysts expect negative third-quarter free cash flow at Alphabet and Amazon, but their actual results had not been released at the time of the preview. The figures for capital spending are also a sum of company guidance, not a reported final total.
It is not yet clear how much AI-related growth is reaching customers outside the largest technology and chip companies, or how quickly that demand may change. SK hynix’s forecast of demand exceeding supply through 2030 could be affected by customer needs, competing suppliers and market conditions. Its results are preliminary and unaudited, and the source does not provide a later update confirming the final figures.
The China index decline and U.S. earnings estimates also cannot settle whether AI stocks as a whole are overvalued. They cover different markets and measures. The upcoming company reports may add evidence on spending and sales, but they will not by themselves establish the long-term return on AI investment.
This Week’s Earnings Tests
Microsoft and Alphabet are scheduled to report quarterly results on October 28, 2026, with Amazon due on October 29. Investors will be watching capital-spending guidance, cloud revenue growth relative to that spending, free cash flow and any comments about computing capacity. Those disclosures will help clarify whether infrastructure investment is being matched by customer demand and operating cash.
Management commentary will need to be distinguished from reported results. Plans and forecasts can indicate how executives see demand developing, but they are not proof of future sales or returns. The same applies to the spending total: company guidance may be updated, and actual outlays can differ from plans.
The source material describes this article as an October 28 preview and says a dated update would be added after the results. No such update or the companies’ eventual figures are included in the material provided here. Until those reports are available, the confirmed picture remains mixed: record preliminary profit at SK hynix, a steep fall in China’s AI-linked index, and unanswered questions about how quickly Big Tech’s spending will pay off.
Key Questions
What is the clearest sign that AI hardware demand is strong?
SK hynix reported 83.3 trillion won in preliminary operating profit for the third quarter, up from 11.4 trillion won a year earlier. The figures are company-reported and preliminary, and show a strong quarter for this memory supplier rather than the performance of the entire AI market.
Why are Microsoft, Alphabet and Amazon’s results important?
The companies are among the major buyers of data-center infrastructure. Their results and guidance can show how cloud growth, capital spending and free cash flow are changing as they invest in AI. Microsoft and Alphabet are scheduled to report October 28, followed by Amazon on October 29.
Does the S&P 500 earnings estimate mean every company is growing quickly?
No. FactSet estimated third-quarter S&P 500 earnings growth at 29.5% as of October 23, but its October 19 comparison said removing Micron and Nvidia would lower a separate growth estimate from 29.0% to 20.0%. The figures are estimates and show that a few companies contribute substantially to the aggregate.
What does the fall in China’s STAR50 index show?
Reuters reported that the index of Chinese AI and chip leaders had fallen 40% in under four months as of October 26. It indicates a sharp reversal in that market over the stated period, but does not by itself establish the outlook for U.S. companies or AI investment globally.
Source: ThorstenMeyerAI.com
