Nonfarm Payrolls “Scare Off” US Stock Market, But AI Hardware Soars Against the Trend! Semiconductor Index Jumps Over 3%, SanDisk Surges 12%

The unexpectedly strong U.S. August non-farm employment report put pressure on the major U.S. stock indices on Friday, but funds did not fully exit tech stocks; instead, they distinctly flowed into the hardware chain of the AI infrastructure sector.

The three major U.S. stock indices all ended a two-day winning streak. By Friday’s close, the Dow Jones fell 0.51%, S&P 500 slipped 0.38%, and the Nasdaq dropped 0.29%. At the same time, the Philadelphia Semiconductor Index closed up about 3.4%, making it one of the standout performers of the day.

Storage chips, semiconductor equipment, and AI optical communications all moved higher as segments. Sandisk (SNDK) closed up nearly 12%, becoming the biggest gainer in the S&P 500 on Friday; Marvell (MRVL) rose more than 7%, Applied Materials (AMAT) climbed over 4%, and AMD increased about 4.7%.

SK Hynix (SKHY), Seagate Technology (STX), Micron Technology (MU), and Western Digital (WDC)—these storage and data storage stocks gained about 8.1%, 6.3%, 6.1%, and 5.9% respectively. Hardware stocks related to AI also surged, with Super Micro Computer (SMCI) rising more than 4.5%.

Even more notable is that this contrarian rally isn’t just limited to traditional semiconductors. Optical communication and optical hardware stocks, which benefit from AI data center expansion, also strengthened against the tide. Among them, Coherent (COHR), Corning (GLW), and Lumentum (LITE) closed up about 6.6%, 5.7%, and 4%, respectively.

The market is expressing an increasingly clear judgement with real capital inflows: While the Fed might hike rates—thus suppressing overall valuations and some software growth stocks—the industry logic for AI computing power, storage, and data center interconnection demand has not yet been disrupted by rate expectations.

The Index Fears Rate Hikes, But Chips Don’t: Funds Are “Rotating from Software to Hardware”

The most noteworthy aspect of Friday’s U.S. stocks wasn’t just the drop in the indices, but a very clear divergence emerged within the tech sector itself.

The Philadelphia Semiconductor Index once surged more than 3% intraday, Roundhill Memory ETF also moved up noticeably, while software became a significant drag on the tech sector.

Barron’s noted that Thomas Martin, Senior Portfolio Manager at Globalt Investments, bluntly stated that the chip sector’s strength on Friday had “zero relevance” to the non-farm payrolls, and mainly reflected a continued rotation between semiconductors and software. That day, iShares Expanded Tech-Software ETF dropped about 2.4%, showcasing the stark contrast.

This actually explains an apparently paradoxical phenomenon:

Stronger non-farm payrolls → Higher Fed hike probability → Theoretically, high-valued tech stocks feel the pressure; yet funds are moving out of certain software and into AI hardware.

The reason is that the pricing logic for AI trades is shifting in the market.

In past years, the AI investment boom concentrated significantly on software, models, and application layers; but post-2026, investors are increasingly focused on the “pick-and-shovel” providers that truly benefit from AI capital expenditure—namely GPUs, storage, network equipment, optical modules, fiber optics, and data center infrastructure.

This means, for these companies, the key variable driving profit growth isn’t just the risk-free rate, but whether large-scale data center capital spending keeps expanding, whether the demand for AI computing power continues to rise, and whether there are supply constraints on relevant hardware.

This week’s market focus on OpenAI’s latest GPT-6 Astra launch has further reinforced this industry narrative. The release has signaled continuous upgrading of AI capability, and the massive AI infrastructure investment plans between OpenAI and giants like Microsoft, Oracle, CoreWeave, AMD, and Broadcom make the sustainability of AI hardware demand once again a focal trading theme.

Sandisk Soars 12%: Storage Chips Become the “Hottest” AI Hardware Segment

If the Philadelphia Semiconductor Index’s 3.4% gain is eye-catching, then Sandisk’s (SNDK) nearly 12% jump stands out even more.

Sandisk became one of the biggest gainers in the S&P 500 on Friday; Micron rose around 6%, Western Digital and Seagate Technology both advanced about 6%. The overall storage segment strengthened, with the Roundhill Memory ETF once up nearly 7%.

The focus on storage chips by capital is no accident.

The rapid expansion of AI data centers not only drives GPU demand, but simultaneously increases the need for HBM, DRAM, NAND, and enterprise SSDs. Especially as AI model sizes grow, inference tasks multiply, and data centers continuously expand, storage and data transmission are becoming ever more critical parts of AI infrastructure.

The market has recently formed a clear “AI storage super cycle” narrative. The Financial Times cited TechInsights executive Dan Kim, noting that AI data centers are fueling a surge in DRAM and NAND demand, sending storage chip prices up by even over 200%. And new capacity is not expected in scale until around 2028, further strengthening storage vendors’ pricing power due to supply constraints.

Therefore, unlike many software stocks that face “whether AI will disrupt their own business model,” storage chip companies stand directly on the supply side of AI capital expenditure.

This also means that in a rising-rate, reduced-risk exposure environment, investors remain willing to bet on higher profit expectations for these hardware firms.

From “Computing Power” to “Connectivity”: Optical Communication Stocks Under the Spotlight

Another clear uptrend was seen in the optical communication and interconnection infrastructure of AI data centers.

Corning, Coherent, Lumentum, and other optical communication-related stocks all outperformed on Friday. While their business specifics differ, the shared label given by the market is very clear: AI data center expansion needs ever more, and higher speed, optical interconnection.

Deutsche Bank just released a report this week pointing out that the AI sector is entering a new stage—the investment focus is shifting from purely increasing computing power to connections and networking. The bank believes that, as data centers continue to expand, companies involved in “manufacturing and transmitting optical signals” will become key beneficiaries of AI infrastructure, listing Lumentum and Coherent as prime picks.

This is also why optical communication stocks could rally against the “rate hike trade” backdrop on Friday.

For the market, advances in GPU performance boost computing ability, but as more GPUs are deployed within or even across data centers, data transmission between chips becomes the new bottleneck.

Thus, the importance of optical modules, lasers, fiber, and high-speed network equipment keeps rising.

Already, earnings reports from Lumentum and Coherent have shown that AI demand is swiftly entering their business flows. Lumentum’s latest quarter revenue grew more than 100% year-over-year, and Coherent set new records for quarterly income.

As a result, the market is no longer just trading “AI chips” but a more complete AI infrastructure chain: compute chips → storage → networking → optical communication → data center power and cooling.

AMD, Applied Materials Also Advance: Hardware Rally Spreads Through the Supply Chain

This capital preference is not limited to memory and optical communications.

AMD jumped 4.69% on Friday, while Applied Materials rose 4.31%; AI hardware stocks like Marvell and Super Micro Computer also moved higher. Reuters reported that the semiconductor sector rose 3.4% that day, notably outperforming other industries.

Applied Materials’ rally is particularly worth noting. As a leader in semiconductor equipment, its share price gains show that capital attention is now extending beyond just AI chips to the equipment needed to manufacture AI chips.

In other words, the market is not just trading good news for a single company—it’s about a more macro logic: As long as AI capital expenditure doesn’t cool off significantly, the entire hardware supply chain—from GPU, storage, to semiconductor equipment, optical communications, and network infrastructure—could continue to benefit.

This also explains why the strong non-farm data’s rates impact didn’t hit all high-valuation tech shares equally, as in the past.

The Moat of AI Hardware: Profit Growth Temporarily Trumps Rate Pressure

Of course, stronger non-farm data still poses an unavoidable issue for AI hardware stocks: valuation.

On Friday, U.S. August non-farm payrolls added 162,000 jobs, far outpacing the market's expected 56,000. July jobs data was also revised from a former loss of 23,000 to an increase of 21,000. After the data, the probability of a Fed September rate hike jumped from about 49.4% to 58.4%; the 2-year U.S. Treasury yield rose to about 4.37%.

Usually, a higher risk-free rate means a higher discount on future cash flows for growth stocks, so AI high-valuation stocks should theoretically be under pressure.

But this time, a different pattern appeared: Investors are distinguishing “valuation stories” from “profit stories.”

Some software companies are facing competitive, business model, and valuation digestion issues arising from AI; meanwhile, the AI hardware supply chain has more direct order growth and capital expenditure backing.

Barron’s, citing Globalt’s Martin, said the strength in chip stocks that day came mainly from sector rotation, while continued AI enthusiasm remains a key driver for the tech sector's resilience. Meanwhile, signs of a rotation from software to hardware already appeared previously in the market.

More importantly, some AI hardware leaders—after their previous considerable gains—still remain a fair distance from historical highs. Storage stocks like Sandisk and Micron have rallied sharply this year, but are not all near record highs, leaving room for capital to rotate back in. Micron rose about 4.4% to near $1,000 on Friday, Sandisk soared over 11%, but Micron remains about 20% below its June peak.

On One Side the “Rate Hike Trade,” on the Other the “AI Trade”: U.S. Stocks Are Accelerating Divergence Internally

Ultimately, the market on Friday actually saw two simultaneous trades.

The first was a macro trade: Strong non-farm payrolls → Higher probability of a Fed September hike → Rising U.S. Treasury yields → Pressure on the overall market and some growth stocks.

The second was an industry trade: AI model upgrades continue → Data center capital expenditure extends → Growing demand for chips, storage, networking, and optical communication → AI hardware becomes a safe haven for capital.

This is why, while all three major indices ended lower, the Philadelphia Semiconductor Index jumped 3.4%, and Sandisk could surge about 12%. Reuters data showed the semiconductor sector was one of the day’s most obvious winners; Barron's noted the semiconductor index rose about 3% while the software ETF fell more than 2%.

For AI hardware, the real test isn’t a single non-farm report, but whether a high-rate environment will eventually hit corporate capital expenditure.

If future data shows the economy is still strong and AI investment continues to expand, these hardware companies may actually benefit from a “strong economy + high interest rates” environment; but if rates keep rising and eventually start to constrain corporate financing and data center capital spending, then this strong sector could face a double test of valuation and demand.

At least on September 4th, the market made its view very clear: The broader market is trading rate hikes, AI hardware is trading AI.