Are SMH, SOXX, and Hardware Leaders Absorbing Outflows from Software Stocks?

SMH, SOXX, and hardware leaders are indeed receiving stronger capital preference, but “absorbing outflows from software stocks” should not be interpreted as a one-for-one transfer. A more accurate conclusion is that AI compute, semiconductor equipment, storage, memory, networking, and data-center infrastructure are attracting incremental allocation, while software stocks are entering a new screening phase based on valuation and AI monetization capability. To judge whether this is a short-term crowded trade or a durable migration into AI hardware, you need to look at ETF flows, relative returns, holding concentration, earnings revisions, and cloud-provider capital expenditure together.

Key Takeaways

• SOXX and SMH have seen strong inflows, showing allocation demand for semiconductor ETFs.
• IGV has seen recent monthly outflows, but still retains year-to-date inflows, so software has not been fully abandoned.
• Hardware leaders benefit from AI compute, storage, networking, and data-center capital expenditure.
• Software stocks are under pressure from valuation, unclear AI revenue, and changing enterprise budget priorities.
• To judge capital absorption, watch ETF flows, relative returns, earnings revisions, and holding concentration.
• Semiconductor ETFs are more volatile, and inflows do not mean risk has declined.

What Do Recent Inflows into SMH and SOXX Tell Us?

Inflows into SMH and SOXX show that the market is increasing exposure to semiconductors and AI infrastructure, but this looks more like “incremental allocation to AI hardware” than “all outflows from software stocks being absorbed by semiconductor ETFs.” Only if software ETFs continue to see outflows, semiconductor ETFs continue to see inflows, and hardware leaders receive upward earnings revisions at the same time can the conclusion move closer to a clear capital migration. Otherwise, the better interpretation is that market preference is tilting toward the hardware chain.

SOXX has become a recent focal point because the intensity of inflows has been unusually strong. ETF.com reported that the iShares Semiconductor ETF saw more than $5.4 billion of inflows in a single day on July 8, 2026, showing that semiconductor ETFs had drawn attention from both short-term traders and institutional allocators. ETF Central data as of July 20, 2026, also showed that SOXX had about $46.38 billion in assets, with about $8.44 billion of net inflows over the past month and $15.42 billion year to date. These figures are enough to show that semiconductor ETFs have become an important vehicle for AI hardware capital flows.

SOXX’s advantage is that it covers the semiconductor value chain rather than simply betting on one company. BlackRock positions the iShares Semiconductor ETF as a fund that seeks to track the performance of an index composed of U.S.-listed semiconductor stocks, with holdings typically covering chip design, manufacturing, equipment, and related supply-chain companies. Therefore, SOXX is better suited for observing “semiconductor-sector capital flows” than for tracking Nvidia alone.

SMH works differently. ETF Central data as of July 17, 2026, showed that SMH had about $67.42 billion in assets, with about $2.17 billion of net inflows over the past month and $8.28 billion year to date. VanEck describes SMH as tracking the performance of U.S.-listed companies involved in semiconductor production and equipment. Because SMH is more sensitive to leaders such as Nvidia, TSMC, and Broadcom, it is more suitable for observing capital preference toward concentrated AI hardware leaders.

Summary: Inflows into SMH and SOXX prove that semiconductor and AI hardware assets are receiving stronger allocation preference. SOXX’s large one-day inflow and strong monthly net inflow, in particular, show that investors are willing to increase semiconductor exposure quickly through ETFs. But this does not prove that outflows from software stocks have been fully absorbed, because ETF flows only show fund-level creations and redemptions, not the exact source of the money. A more prudent conclusion is that AI hardware is attracting incremental capital, software funding pressure is rising, but the direct link between the two still needs more evidence.

Are Software Stocks Really Seeing Outflows? How Should You Read IGV, Software Leaders, and Valuation Pressure?

Software stocks are indeed facing short-term funding pressure, but this is not a full-scale exit. IGV has seen net outflows over the past month, and software ETFs have underperformed semiconductor ETFs year to date. Yet IGV still shows year-to-date net inflows, which means the market has not completely abandoned software assets. Instead, investors are re-screening companies based on whether they can convert AI features into renewals, upsells, RPO, and margin improvement. The problem for software is not “no value,” but “slower monetization than hardware.”

IGV is an important representative for observing capital flows in North American software. ETF Central data as of July 17, 2026, showed that IGV recorded about $399 million of net outflows over the past month, but $6.08 billion of net inflows year to date. This combination is crucial: recent monthly outflows show that software is under short-term pressure, while year-to-date net inflows show that institutions have not fully abandoned the sector. BlackRock positions the iShares Expanded Tech-Software Sector ETF as covering software, cloud, and some digital media companies, so IGV can reflect overall risk appetite toward SaaS, cloud software, cybersecurity, and enterprise applications.

Software stocks are under pressure from several directions. First, enterprise AI budgets are prioritizing servers, GPUs, HBM, storage, networking, and data centers, while application software purchases can be delayed. Second, SaaS seat expansion is slowing, and customers are paying more attention to actual usage and AI ROI. Third, high-valuation software stocks are more vulnerable to valuation compression when earnings forecasts are revised down. Fourth, some AI tools and agents may reduce the functional scarcity of traditional software, giving customers more room to renegotiate.

Still, software stocks are not moving in one single direction internally. Investopedia reported that Morgan Stanley analysts believed sentiment toward software stocks had become too negative, while highlighting certain software companies that may still have upside potential. MarketWatch also noted that after IGV had fallen significantly from earlier highs, some indicators suggested software could deserve a second look. This suggests that the market is not abandoning software, but moving from “buying the AI narrative” to “buying verifiable cash flow.”

To assess software funding pressure, focus on five indicators:

• Continuous flows in software ETFs such as IGV, WCLD, and XSW;
• RPO, net retention, and renewal rates among software leaders;
• Whether AI features generate standalone revenue or higher-tier pricing;
• The direction of analyst revisions to software EPS and revenue forecasts;
• Whether the relative performance gap between IGV and SMH/SOXX continues to widen.

Summary: Software stocks are under funding pressure, but a narrative of “complete capital exit from software” is not supported. IGV’s recent monthly outflow shows weaker short-term allocation preference, while year-to-date inflows show that capital is still searching for opportunities inside the sector. The companies being punished are those with unclear AI monetization, high valuations, and weak support from renewals and upsells. Cybersecurity, databases, cloud platforms, automation, and data governance may still benefit from AI implementation. You should treat software as a divergent asset class, not as an abandoned one.

Why Are Hardware Leaders More Able to Attract AI-Themed Capital?

Hardware leaders are attracting more capital because the early stage of AI investment is most constrained by GPUs, HBM, advanced processes, networking equipment, storage, and power capacity. These links can convert cloud-provider and enterprise capital expenditure into orders, deliveries, and revenue more quickly. Software companies usually need more time to prove paid conversion from AI features. In an uncertain environment, the market prefers companies with “orders already visible” over software stories that “may monetize later.”

The transmission path of AI capital expenditure is straightforward: cloud providers raise data-center budgets, which drives demand for servers, GPUs, ASICs, HBM, switches, optical modules, storage, power equipment, and semiconductor equipment. Reuters analysis noted that capital expenditure expectations around AI among major technology companies have risen significantly. This can pressure free cash flow, but it also sends more spending into the AI infrastructure chain. For capital markets, the closer a company is to a hardware bottleneck, the more easily it can receive a valuation premium.

The monetization path for hardware companies is easier to verify. Nvidia, Broadcom, ASML, TSMC, Micron, Arista, and Vertiv correspond to chips, custom ASICs, lithography equipment, foundry capacity, memory, networking, and power infrastructure respectively. Investors can assess demand strength through orders, capacity, gross margin, delivery cycles, and customer capital expenditure. Software AI monetization has to pass through more steps: whether customers activate AI features, whether they are willing to pay, whether they expand seats, whether renewal prices rise, and whether margins improve.

Hardware leaders are also amplified by ETF mechanics. When semiconductor ETFs such as SMH and SOXX receive inflows, they passively buy the underlying stocks according to their index weights. After leaders rise, ETF performance becomes stronger, which can attract more inflows. This mechanism can reinforce gains during a trend, but it can also amplify volatility during a pullback. When a small number of leaders have high weights, ETF performance can resemble a “leader stock basket” more than a fully diversified industry fund.

Summary: Hardware leaders are attracting capital not because software has no value, but because AI infrastructure investment has a shorter order path, clearer budget attributes, and more visible supply bottlenecks that can support pricing and profit. Semiconductor ETFs amplify this preference further, allowing investors to express an AI hardware view quickly through SMH and SOXX. But hardware does not equal low risk. Leader concentration, valuation expansion, order cycles, and export restrictions can all affect returns. You need to look at both order delivery and valuation levels, not just ETF inflows.

Are SMH and SOXX Truly Absorbing Outflows from Software Stocks? What Evidence Is Needed?

To prove that SMH and SOXX are absorbing outflows from software stocks, at least three pieces of evidence need to appear together: software ETFs must continue to see outflows, semiconductor ETFs must continue to see inflows, and hardware leaders must receive upward earnings revisions. If SOXX and SMH are seeing inflows while IGV still has year-to-date net inflows, you can only say that capital preference is changing, not that a clear migration has taken place. ETF flows are important clues, but they are not a complete map of capital movement.

The first layer of evidence is ETF flows. SOXX has seen strong recent inflows, SMH has also seen monthly net inflows, and IGV has seen recent outflows, showing short-term pressure on software. But these figures do not tell you whether money invested in SOXX came directly from selling IGV. The capital may have come from cash, broad technology ETFs, active fund rebalancing, options hedging, short-term trading, or cross-sector rotation. ETF flows show “where money went,” but not directly “where money came from.”

The second layer is relative performance. If SMH and SOXX continue to outperform IGV over several weeks or months, it indicates that the market is willing to pay a higher premium for semiconductors. ETF Central comparison data has shown that SOXX’s year-to-date performance through mid-July 2026 was meaningfully stronger than some technology and software ETFs; for example, SOXX had clearly outperformed SOXQ year to date, reflecting capital concentration in more representative semiconductor tools. Still, relative performance requires earnings estimate support; otherwise, it may just be valuation expansion.

The third layer is fundamental revision. Hardware leaders need to show upward EPS revisions, rising orders, improving gross margins, or continued capital expenditure expansion. Software leaders, by contrast, would need to show slowing RPO, renewal pressure, or AI revenue falling short of expectations. Only when fundamentals move in the same direction as capital flows does the rotation become more reliable. Otherwise, inflows into semiconductor ETFs may simply represent short-term momentum buying, while software outflows may just be a temporary stop-loss move after valuation compression.

Summary: It is fair to say that “semiconductor ETFs are receiving capital preference,” but not that “outflows from software have been fully absorbed by SMH and SOXX.” True capital absorption requires ETF flows, relative returns, earnings revisions, and cloud capital expenditure to confirm one another. Inflows into SOXX and SMH show that AI hardware is currently a more popular expression vehicle. Recent IGV outflows show that software is under short-term pressure. But as long as IGV still has year-to-date inflows and there remain strong software subsectors, the market should not be described as a one-way exit from software.

What Are the Main Risks for SMH, SOXX, and Hardware Leaders?

The main risk for SMH and SOXX is not whether the AI theme exists, but holding concentration, valuation expansion, semiconductor cycles, export restrictions, and flow reversal. Rapid inflows can push prices higher, but they can also magnify drawdowns. When the market realizes that orders, capital expenditure, or earnings estimates are not keeping up with stock prices, semiconductor ETFs may become more volatile than software ETFs. The risk profile is shifting from “uncertain demand” to “overcrowded expectations.”

Holding concentration is the first layer of risk. SMH has higher exposure to a small number of leaders, so if core companies such as Nvidia, TSMC, Broadcom, or ASML lower guidance, the ETF may pull back quickly. SOXX is relatively more diversified, but it is still concentrated in the semiconductor value chain and cannot avoid sector-wide valuation compression. The “diversification” of a semiconductor ETF is not cross-sector diversification; it is diversification within the same industry chain.

The second layer is cyclicality. Semiconductor equipment, storage, GPUs, servers, and networking equipment can all experience inventory cycles. If cloud-provider capital expenditure slows, orders may be delayed. If storage and memory prices weaken, earnings estimates may be revised down. If equipment orders have pulled forward demand, later deliveries may fall short of expectations. Reuters reported on July 17, 2026, that as chipmakers and other high-flying stocks fell, the market began worrying about the intensity of the AI trade and leveraged positioning, showing that semiconductors do not only move upward.

The third layer is geopolitics and tail risk. Advanced chips, Taiwan supply chains, export controls, and equipment delivery can all affect expectations. Research on Taiwan-related ETF risk noted that semiconductor-concentrated ETFs may show higher tail risk and asymmetric volatility under technology concentration, geopolitical uncertainty, and supply-chain disruption. These risks do not appear every day, but when they occur, ETF-level drawdowns can be highly concentrated.

Key risk signals to monitor include:

• The top-ten holding weights of SMH and SOXX continue rising;
• EPS or order expectations for semiconductor leaders are revised down;
• Cloud-provider capital expenditure growth slows;
• Storage, memory, and equipment orders weaken;
• ETF flows shift from sustained inflows to sustained outflows;
• Software ETFs see capital return and relative performance recovery at low levels.

Summary: Capital inflows do not mean risk has disappeared; they mean risk has changed shape. SMH and SOXX are attractive because of AI hardware orders, semiconductor supply bottlenecks, and cloud capital expenditure, but they are also exposed to concentration, valuation, cycles, and policy. Semiconductor ETFs can amplify the hardware theme quickly during an uptrend, but they can also expose crowded positioning quickly during a pullback. You should not look only at “inflows.” You should also ask whether those inflows are supported by earnings revisions and whether capital has become overly concentrated in a small number of leaders.

How Should You Track AI Hardware Capital Flows Instead of Chasing Semiconductor ETFs?

A more prudent approach is not to chase SMH or SOXX just because you see inflows, but to build a tracking framework based on “flows—relative returns—earnings revisions—capital expenditure—trading costs.” Only when ETF flows, hardware leader earnings, and cloud capital expenditure confirm one another can AI hardware capital flows be considered more durable. Otherwise, large inflows into semiconductor ETFs may simply be a short-term crowded trade rather than a long-term allocation migration.

You can divide your watchlist into four groups. The first group is semiconductor ETFs: SMH, SOXX, SOXQ, and SOXL, which help observe the AI hardware theme and leveraged sentiment. The second group is software ETFs: IGV, WCLD, and XSW, which help observe funding pressure on SaaS, cloud software, and the application layer. The third group is broad technology ETFs: QQQ, XLK, and VGT, which help assess overall risk appetite toward the technology sector. The fourth group is hardware and software leaders: Nvidia, Broadcom, ASML, TSMC, Micron, and Arista on the hardware side, and Microsoft, Oracle, ServiceNow, Salesforce, and Adobe on the software side.

Trading costs also need to be included in the framework. When you track U.S. stocks or ETFs such as SMH, SOXX, IBM, ASML, Nvidia, and Broadcom, you should look not only at flows, price changes, and volume, but also commissions, platform fees, external agency fees, bid-ask spreads, exchange rates, and order types. If you use Biya to observe U.S. and Hong Kong stock opportunities, you can use U.S. stock search to compare semiconductor, software, and AI infrastructure-related stocks. Biya charges $0 commission for U.S. stock trading, while platform fees, external agency fees, and other costs are subject to U.S. stock trading fees and the order display. Service availability also depends on the user’s location, identity verification result, platform rules, and applicable laws and regulations.

Summary: To properly track AI hardware capital flows, you need to place ETF flows, relative returns, earnings revisions, cloud capital expenditure, and trading costs in the same table. Inflows into SMH and SOXX show that the semiconductor theme is more popular, but that does not mean every entry point is suitable for chasing. Recent IGV outflows show software is under pressure, but they do not mean software has lost long-term allocation value. A better method is to track flows weekly, verify earnings and capital expenditure quarterly, and then decide whether to adjust your watchlist based on costs and risk tolerance.

If you are watching capital rotation among SMH, SOXX, AI hardware leaders, and software stocks, you should not rely only on headlines. You need to put ETF flows, holding structure, price volatility, fees, and order data into the same tracking table. Through Biya, you can observe multi-asset market information across U.S. stocks, Hong Kong stocks, and digital assets, and build a watchlist around semiconductor, enterprise software, and AI infrastructure-related companies. After you register an account, you should still assess trading feasibility based on order displays, liquidity, fee structures, and applicable local rules. Public market information, ETF flows, and fee disclosures are useful for improving decision quality, but they do not constitute investment advice or imply any return guarantee.

FAQ

Which Is Better for Observing AI Hardware Capital Flows, SMH or SOXX?

Both SMH and SOXX can be used to observe AI hardware capital flows, but they emphasize different exposures. SMH is more concentrated in semiconductor leaders and is more sensitive to companies such as Nvidia, TSMC, and Broadcom. SOXX offers broader semiconductor value-chain exposure. When choosing an observation tool, you should compare holdings, concentration, fees, liquidity, and your own risk tolerance.

Does a Large One-Day Inflow into SOXX Indicate a Long-Term Rotation?

A large one-day inflow into SOXX does not, by itself, indicate a long-term capital rotation. Large inflows may come from institutional allocation, short-term trading, rebalancing, options hedging, or arbitrage demand. To judge a longer-term rotation, you need to combine continuous flows, relative returns, hardware leader earnings revisions, and cloud-provider capital expenditure, rather than relying only on one-day creation volume.

Do Outflows from Software ETFs Mean SaaS Stocks Are No Longer Worth Watching?

Outflows from software ETFs do not mean SaaS stocks are no longer worth watching. Recent outflows show that software is under short-term pressure, but the sector will still diverge internally. Cybersecurity, databases, cloud platforms, and AI automation software may regain capital attention. The key is whether renewal rates, RPO, AI revenue, and earnings estimates improve.

How Can Beginners Judge Whether a Semiconductor ETF Is Too Concentrated?

Beginners can check a semiconductor ETF’s top-ten holdings, single-stock weight, industry distribution, and rebalancing rules. If a few leaders account for too much of the portfolio, the ETF’s volatility may resemble a basket of leading stocks rather than a fully diversified sector fund. The higher the holding concentration, the more attention you need to pay to leader earnings and valuation risk.

What Costs Should Investors Watch When Trading U.S.-Listed ETFs Such as SMH and SOXX?

When trading U.S.-listed ETFs such as SMH and SOXX, investors should watch commissions, platform fees, external agency fees, bid-ask spreads, exchange rates, and order types. Fee structures vary by platform, and final costs should be based on fee disclosures, account statements, and order displays. Fees are only one part of the decision; liquidity, valuation, and risk tolerance also matter.

What Signals May Appear Before AI Hardware Capital Flows Reverse?

Before AI hardware capital flows reverse, common signals include downward revisions to cloud capital expenditure, slowing semiconductor orders, weakening storage prices, downward earnings revisions for hardware companies, consecutive outflows from SMH and SOXX, and renewed inflows into software ETFs. A single signal may not be enough; multiple indicators should be verified together.