BiyaPay data showed a clear split in U.S. equities last week. The Dow Jones Industrial Average posted a modest gain, while the S&P 500 slipped and losses were deeper in the Nasdaq Composite and the Russell 2000. With inflation data moving higher, the market has returned to a familiar set of questions: how much AI computing infrastructure will cost, who will absorb the financing burden, and when revenue can cover capital spending.
Hong Kong stocks were little changed on the week, though technology shares were weaker. Bitcoin traded near $81,000 in a range. This week, China’s third-quarter economic data is already out, and earnings from Meta, Tesla and Netflix are due. The next test for markets centers on three things: profits, inflation and financing costs.
Last week: the Dow rose while tech and small caps fell back
As of the Oct. 16 close, the S&P 500 stood at 7,659.44, down about 1.9% for the week. The Nasdaq Composite closed at 26,140.03, down about 4.5%. The Dow finished at 52,002.41, up about 0.7%, while the Russell 2000 fell about 3.0%. The Nasdaq logged its weakest week since June, including a 2.1% drop on Friday alone.
That set of moves does not describe a market where all U.S. stocks weakened in tandem. The Dow still rose over the week, which suggests the pressure was concentrated in technology and small-cap shares, especially in areas more exposed to financing conditions and assumptions about future growth. Stress inside the S&P 500 technology sector was also broad. At Friday’s close, only 7 of its 75 constituents had avoided losses. Micron, AMD, Broadcom and Intel fell 5.6%, 6.0%, 4.5% and 5.8%, respectively, on Friday. Those figures refer to single-day declines, not weekly losses.
U.S. inflation data added pressure to rate expectations. September CPI rose 3.7% from a year earlier, above August’s 3.4%, and increased 0.6% from the prior month. Gasoline prices climbed 9.1%, accounting for more than half of the monthly increase. The next day, final-demand PPI for September rose 1.3% month over month and 6.3% year over year. CPI tracks prices paid by consumers, while PPI measures prices at the producer level. Their baskets and methodologies differ, so PPI should not be treated as a direct forecast for future CPI. Read together, though, they still point to cost pressure worth watching.
AI demand remains, but the market is now focused on returns
The main issue behind last week’s tech pullback was not whether AI applications still have demand. It was whether the returns on AI infrastructure spending can match the scale of the investment. Data centers, chips and power facilities all require heavy upfront capital. Companies expanding in those areas often rely on a mix of operating cash flow, equity financing and debt financing. When rates and credit spreads rise, one of the first things to change is the cost of funding and the time needed to recover that investment.
Pressure on both the equity and debt of CoreWeave brought that issue into sharper view. Yields on its bonds in the secondary market moved higher, prompting investors to reassess the relationship between compute contracts, equipment values and debt maturities. Volatility in one company’s bonds does not prove that the broader AI investment cycle has peaked. It does, however, remind the market that order volume is not the same as cash collection, and backlog is not the same as realized profit.
In looking at this kind of cross-market transmission, BiyaPay said it compares U.S. tech stocks, Hong Kong tech stocks and BTC within the same market picture. BiyaPay, described as a global one-stop asset allocation platform, uses USDT as a core funding entry point and connects U.S. and Hong Kong equities, cryptocurrencies, commodity futures and cross-border remittances through a single account system. It also supports exchange between more than 30 fiat currencies and more than 200 digital currencies, and allows users to trade real U.S. and Hong Kong stocks with USDT at zero commission. For investors who track several markets at once, that kind of centralized access can make it easier to compare how different assets respond to rates and risk appetite. Corporate operating quality, however, still has to be judged through earnings reports, especially revenue, profit, cash flow and capital expenditure, rather than inferred from price moves alone.
There is also no single template for the technology sector. Chipmakers need to show whether orders can turn into revenue and whether gross margins can hold. Cloud service providers need to show the relationship between AI service revenue and data center spending. Platform companies need to explain whether AI is improving advertising, recommendations or user efficiency. The investment cycle is still in place, but the market is shifting its valuation lens from how much is being spent to what verifiable return that spending is producing.
Hong Kong market review: Hang Seng nearly flat, Hang Seng Tech weaker
Hong Kong stocks traded normally from Oct. 12 to Oct. 16 and were closed on Oct. 19 for a holiday. The Hang Seng Index closed at 24,186.83 on Oct. 16, down about 0.1% from the Oct. 9 close. The Hang Seng Tech Index ended at 4,108.24, down about 2.1% for the week. The headline index looked almost unchanged, but pressure on technology shares was more visible.
Volatility in the Hang Seng Tech Index increased during the week. It fell 2.72% on Oct. 15 and rebounded 1.16% on Oct. 16. A one-day rebound did not erase the weekly decline, nor does it justify calling a trend reversal. Hong Kong tech stocks are being influenced at the same time by U.S. tech valuations, expectations for the mainland economy and their own earnings performance. Those three forces do not necessarily move in the same direction.
Data released on Oct. 19 showed that China’s GDP grew 4.6% year over year in the first three quarters of 2026. Third-quarter GDP rose 4.5% from a year earlier and 1.2% quarter over quarter. September retail sales increased 1.1% year over year, while the producer price index for industrial goods rose 5.0% in the same month. Aggregate growth, consumer demand and upstream pricing are not telling exactly the same story. Once Hong Kong trading resumes after the holiday, the market’s interpretation of those figures will still need to be read alongside sector earnings and company guidance rather than through GDP growth alone.
Bitcoin offered another reference point for risk appetite. On a UTC daily close basis, BTC moved from about $82,564 on Oct. 9 to about $81,524 on Oct. 16, a decline of about 1.3%, while remaining range-bound overall. Because crypto trades continuously, its statistical cutoffs differ from those used in stock markets, so direct comparisons between a single daily close in BTC and a weekly move in U.S. equities require care.
What to watch this week: economic data and earnings
China data and Hong Kong’s market reopening
China’s National Bureau of Statistics released first-three-quarter data on the morning of Oct. 19. Hong Kong markets were closed that day and will resume trading on Oct. 20. The more useful question now is how investors map consumption, industrial prices and trade data onto sectors such as internet, consumer and manufacturing, rather than treating headline GDP as a direct proxy for company revenue growth.
Netflix, Meta and Tesla earnings
Netflix is scheduled to report third-quarter results on Oct. 20 U.S. Pacific Time. Meta has confirmed that it will release earnings after the close on Oct. 21 U.S. Eastern Time and hold a conference call the same day. Tesla is also scheduled to report on Oct. 21. For Beijing time, the Meta and Tesla reports fall in the early hours of Oct. 22.
Meta will need to address whether ad revenue growth can cover AI-related capital expenditure, and whether investors can see operating evidence in recommendations, ad conversion and cost efficiency. Tesla has already reported third-quarter deliveries of 486,532 vehicles, production of 464,391 vehicles and energy storage deployments of 13.7 GWh. The earnings report will need to fill in what those volumes mean for per-vehicle profit, cash flow, the profit contribution from energy storage and progress in newer businesses such as autonomous driving. Deliveries and deployments are volume indicators. They do not directly substitute for profit.
For Netflix, the focus is on the quality of growth in subscriptions and advertising, the efficiency of content spending and free cash flow. Streaming, social media and auto manufacturing run on different business models, so earnings comparisons should be based on the operating metrics that matter most to each company, not only on whether earnings per share beat expectations.
U.S. housing, labor and business surveys
The U.S. economic calendar is relatively light this week. Sept. housing starts and building permits are due on Oct. 20. Weekly initial jobless claims are due on Oct. 22. On Oct. 23, markets will get preliminary S&P Global manufacturing and services PMI readings, along with the final University of Michigan consumer sentiment reading. Federal Reserve officials are in the pre-meeting blackout period, so with no fresh policy remarks to parse, economic data and earnings carry more weight.
Housing data shows how high rates are affecting mortgages and new construction demand. Initial jobless claims offer a weekly read on the labor market. PMI and consumer sentiment help frame business activity, costs and demand expectations. Any single release can be affected by seasonality and sampling. Market reaction will also depend on how far the data deviates from expectations and whether it changes views on inflation, growth and rates.
The week’s framework: can earnings catch up with costs?
Last week did not produce a verdict that the AI trade is over. It did push financing and return questions to the front. CPI and PPI showed that price pressure remains. Volatility in chip stocks and AI infrastructure-linked assets reminded investors that funding conditions can affect the pace of expansion. Hong Kong’s broader market was nearly flat, but technology shares were weaker. BTC stayed range-bound and did not break into a clear one-way move.
The key test this week comes from corporate earnings and macro data. If companies such as Meta and Tesla can explain spending through revenue, profit and cash flow, the market may start to separate capital expenditure that is generating returns from longer-dated narratives that still need proof. If the data continues to show rising costs, softer demand or weaker earnings guidance, valuation pressure may persist. The central question is not whether a headline is dramatic enough. It is whether operating data can support the growth expectations already embedded in prices.
