In June 2026, the technology sector is undergoing a profound internal "capital restructuring." The one-sided rally driven by AI chips over the past two years began to show clear cracks from late Q1 into early Q2—capital started shifting from overweight AI hardware to underweight software stocks. At the same time, large-cap tech stock holdings plunged from a historic 97th percentile to the 48th percentile, returning to nearly neutral levels, while tech sector funds still recorded a single-week net inflow of $12.3 billion.
This indicates that capital is not leaving the tech sector, but is instead being structurally reallocated within it. To understand the essence of this capital rotation, it’s essential to grasp both macro-level position changes and micro-level fundamentals of individual stocks. This article will explore two dimensions: first, breaking down the data logic behind the rotation between AI chip stocks and software stocks; then, analyzing seven representative, large-cap stocks with clear growth narratives; finally, outlining specific ways to participate in allocation via Gate’s stock trading feature.
Structural Shifts in Capital Flows: From Concentration to Rotation
As of the week ending June 10, 2026, global equity markets attracted $31.5 billion in inflows, marking a two-month high. Of that, tech sector funds saw a single-week net inflow of $12.3 billion—the highest since 2017. This data conveys two key messages: first, capital is not withdrawing from the tech sector as a whole; second, capital is undergoing structural reallocation within the sector.
Position data shows that previous extreme crowding is being rapidly unwound. According to Deutsche Bank Securities’ global asset allocation update released on June 12, as of June 11, large-cap tech stock holdings had dropped sharply from the historic 97th percentile to the 48th percentile, approaching neutral levels. Meanwhile, the S&P 500 retreated about 5%, but most losses were concentrated in the Nasdaq 100 (down 7%) and the "Tech Magnificent Seven" (down 10%), while the S&P 500 Equal Weight Index and small-cap indices actually rose to record highs.
This means the core of this adjustment is not the release of systemic risk, but rather a concentrated restructuring of previously overcrowded large-cap tech positions.
The Logic of Rotation: "Overextension" in AI Chip Stocks and "Revaluation" in Software Stocks
AI Chip Stocks: From "Broad Gains" to "Divergence"
In the first week of June 2026, the overseas AI chip index fell 4.6%. Broadcom and AMD dropped 13.7% and 9.6%, respectively, while Marvell surged 28.5% thanks to Jensen Huang’s remarks at Taipei COMPUTEX about "the next trillion-dollar company." On June 5, AI concept stocks experienced a rare collective plunge—Micron fell nearly 4%, Qualcomm, Intel, and AMD each dropped over 3%, Broadcom and TSMC were down about 2%.
Three core factors drove this adjustment. First, valuations have already priced in future expectations. For example, the Philadelphia Semiconductor Index soared about 95% in just 45 trading days from late March to early June, far outpacing short-term fundamental support. Second, the market is shifting from "narrative-driven" to "delivery verification"—technical roadmaps are no longer the main focus; production and deployment capabilities are now key. Third, the bar for Q2 earnings surprises has risen sharply—even if results beat expectations, the market will demand guidance for the next quarter.
However, this adjustment is more about digesting valuations and easing trading crowding than a reversal of industry trends. The overseas AI chip index’s weekly drop of 4.6% is much smaller than the software sector’s single-quarter maximum drawdown of 30% earlier this year.
Software Stocks: Value Recovery After "AI Replacement Panic"
In the first half of 2026, the global software market underwent a dramatic swing from extreme pessimism to expectation recovery. Early in the year, accelerated rollout of AIAgent technologies sparked widespread concern—would the SaaS seat-based pricing model be disrupted by intelligent agents? The IGV software index saw a Q1 maximum drawdown of nearly 30%, with total sector market cap evaporating by over $1 trillion.
The turning point came during the Q2 earnings season. North American SaaS companies reported results across the board that exceeded market expectations, with many raising full-year guidance. The IGV index jumped over 21% in May alone, marking its best monthly return since inception. Snowflake was the biggest catalyst for this rebound. The data cloud giant’s Q1 revenue grew 33% to $1.39 billion, beating estimates; remaining performance obligations reached $9.21 billion, up 38% year-over-year. Agentforce’s annual recurring revenue surpassed $1.2 billion, up 205% year-over-year. ServiceNow’s total revenue hit $3.77 billion, up 22%, with 16 new large contracts over $5 million each—a nearly 80% increase year-over-year.
These earnings point to one conclusion: AI is not a disruptor for the software industry, but a new engine for growth. As enterprises deploy AI agents, underlying demand for data governance, identity and access management, workflow orchestration, and agent orchestration platforms is actually increasing.
From a valuation perspective, as of late May 2026, the software/semiconductor ratio was more than 40% below its 200-day moving average. BTIG analysts called this level "unprecedented"—signaling that software’s extreme undervaluation relative to hardware has reached historic extremes.
Core Players: Four Companies in the AI Chip Segment
NVIDIA: The "Pricing Anchor" of Compute Infrastructure
NVIDIA is the largest company in AI infrastructure by market cap and one of the most closely watched bellwethers in the current tech sector capital flows. As of June 15, 2026, NVIDIA’s stock price was in the $205–208 range, with a market cap of about $4.97 trillion. In early June, the price briefly hit $224, with a market cap of roughly $5.4 trillion.
The Q1 FY2027 earnings report released in May 2026 showed revenue of $81.6 billion, up 85% year-over-year, far exceeding expectations. Data center business contributed $75.2 billion, accounting for 92% of total revenue. On the product side, Blackwell architecture chips are ramping up mass production, and the next-generation Rubin architecture is expected to launch in the second half of 2026 with a performance boost of about 5x. In terms of valuation, NVIDIA’s static P/E is around 31.8x, and forward P/E is only 20.9–21.6x, well below its three-year average of 56.1x. Risks include the H20 chip export ban leading to zero revenue from China, major clients developing their own ASICs potentially diverting orders, and the market’s expectation that "outperformance is the norm."
AMD: The "Second Pillar" Accelerating in Data Centers
AMD plays the role of "challenger" in the AI chip segment. As of June 2026, AMD’s market cap was between $150–180 billion. Q1 2026 revenue reached $10.25 billion, up 38% year-over-year, beating estimates; non-GAAP EPS was $1.43, up 43%. Data center revenue hit $5.78 billion, up 57%, mainly driven by continued growth in AMD Instinct GPU shipments. Data center GPU revenue is expected to grow 114% year-over-year to $15 billion in 2026.
AMD’s core narrative is that as AI compute demand moves toward "multi-vendor configurations," enterprise clients are motivated to seek alternatives beyond NVIDIA. AMD’s forward P/E is about 68.6x, with a three-year revenue CAGR of roughly 43.6%, leaving limited room for valuation error.
Broadcom: "One of Two Giants" in Custom ASIC Chips
Broadcom is deeply tied to major clients like Google, Meta, Anthropic, and OpenAI in the custom AI chip (ASIC) space. As of June 2026, its market cap was between $700–800 billion. The Q2 FY2026 earnings report released on June 3 showed total revenue of $22.2 billion, up 48% year-over-year, with a record operating margin of 67%. AI semiconductor revenue climbed to $10.8 billion, up 143%, with AI semiconductor orders exceeding $30 billion for the quarter.
Looking ahead, AI semiconductor revenue for FY2026 is projected at $56 billion, up about 180% from FY2025’s $20 billion. Goldman Sachs subsequently raised its AI semiconductor revenue forecasts for FY2026–2028 to $57 billion, $133 billion, and $193 billion, respectively. The early June stock drop was mainly due to the $56 billion full-year guidance being slightly below the market’s $57.6 billion expectation (about a 2.8% gap), which is a difference in expectations rather than fundamentals. The real constraint is supply, not demand—the quarter’s orders exceeded $30 billion, but actual shipments were only $10.8 billion.
Marvell: The "Structural Beneficiary" of Connectivity Infrastructure
Marvell is the most flexible growth story among the four AI chip players, with its core logic centered on "connectivity"—regardless of who supplies the compute chips, Marvell’s optical interconnect products are foundational to AI data center expansion. As of June 15, Marvell’s market cap was about $244.6 billion, with a year-to-date gain of over 200%. FY2027 Q1 net revenue was $2.418 billion, up 28% year-over-year, setting a new quarterly record; data center business revenue was $1.833 billion, up 27%. Operating cash flow reached $639 million, up 91.89%.
Driven by NVIDIA CEO Jensen Huang’s "next trillion-dollar company" comment, MRVL surged over 27% intraday, with a single-day market cap jump of about $62.4 billion. The company raised its FY2027 full-year revenue outlook to about $11.5 billion, up 40% year-over-year, and set an FY2028 target of $16.5 billion. Optical communications business annual growth is over 70%, and custom AI chip business is projected to break $10 billion in revenue by FY2029.
Core Players: Three Companies in the Software SaaS Segment
Snowflake: The "Platform Moat" of AI Data Cloud
Snowflake is one of the biggest catalysts for the software stock rebound. As of June 2026, its market cap was between $40–50 billion. The Q1 FY2027 earnings report showed total revenue of $1.39 billion, up 33% year-over-year. Product revenue (core data platform business) reached $1.33 billion, up 34%, marking the company’s highest single-quarter product revenue dollar growth. Remaining performance obligations hit $9.21 billion, up 38%. Net revenue retention rate was 126%. Over the past 12 months, Snowflake had 779 customers with product revenue over $1 million, up 29% year-over-year.
On the AI front, more than 13,600 accounts are using Snowflake’s AI features; the Snowflake Intelligence agent platform saw quarterly user numbers more than double. Management called Q1 "a clear turning point"—AI is becoming the core driver of data cloud platform growth. The company signed a $6 billion multi-year deal with AWS and deepened its partnership with OpenAI.
Salesforce: The Pioneer of "AI Application Delivery"
Salesforce’s stock dropped about 33% in early 2026 due to AI replacement fears, but the Q1 earnings report quickly corrected the narrative. Q1 FY2027 revenue hit a record $11.13 billion, up 13% year-over-year—the strongest quarterly growth since Q1 FY2023. Adjusted EPS was $3.88, beating the Wall Street consensus of $3.12 by nearly 24%.
The strongest rebuttal to the "AI disruption" thesis comes from Agentforce—Salesforce’s proprietary AI platform. Agentforce achieved an annualized recurring revenue run rate of $1.2 billion in Q1, up 205% year-over-year. Combined with Data Cloud, AI and data businesses now have annualized revenue of $3.4 billion. Management raised FY2026 (ending January 2027) revenue guidance to $41.3 billion. The current stock price is near a 52-week low at about $164, with TD Cowen and others maintaining "Buy" ratings and a target price of $240, implying roughly 46% upside.
ServiceNow: "Platform Expansion" in Workflow Automation
ServiceNow positions itself as the "workflow engine" for cross-enterprise business processes, serving as the "orchestration control layer" for AIAgent to access various enterprise systems in the AI era. As of June 2026, its market cap was between $120–140 billion. Q1 2026 revenue was about $3.77 billion, up 22.1% year-over-year, beating expectations. The company subsequently raised its 2026 full-year subscription revenue guidance to $15.735–15.775 billion, with the midpoint up $205 million and year-over-year growth of about 20.5–21.0%.
ServiceNow’s 2030 goal is to achieve over $30 billion in subscription revenue—nearly double its current scale—mainly through AI-native innovation, autonomous workflows, and horizontal expansion into new areas like security, CRM, and data analytics. The company has approved a $5 billion share repurchase plan.
Gate Stock Trading: One-Click Access to Both Sides of the AI Value Chain
For investors looking to participate in this round of tech sector capital rotation, Gate’s real US stock trading feature, launched in June 2026, offers a convenient allocation channel.
Gate’s stock trading service covers major US markets including NYSE and Nasdaq, supporting over 10,000 stocks and ETFs. The seven core players mentioned above—NVIDIA (NVDA), AMD, Broadcom (AVGO), Marvell (MRVL), Snowflake (SNOW), Salesforce (CRM), ServiceNow (NOW)—can all be settled in USDT through a single account.
Fractional trading allows investments starting from as little as 0.01 shares, with a minimum of $1 to hold any of these stocks, making it easy for investors to flexibly adjust allocations between AI chip and software stocks. Pre-market and after-hours trading extend trading hours to 16×5, covering price discovery opportunities outside regular sessions. Gate partners with licensed broker Alpaca, which handles trade execution and compliance settlement, ensuring fund safety and regulatory compliance. Under the unified VIP system, accounts with $2,000 or more enjoy the lowest stock trading fee rate at 0.023%.
How to operate: Log in to Gate web or app → enter the "Stocks" trading module → search for ticker (e.g., NVDA, SNOW) → choose buy/sell direction → settle in USDT → manage positions with one account.
Conclusion
The essence of this tech sector capital rotation is a shift from "extreme concentration in AI hardware" to "balanced allocation between hardware and software." Financial data from the seven core players shows that NVIDIA, Broadcom, and Marvell on the AI chip side have validated sustained strong AI compute demand in their latest earnings, with pullbacks reflecting crowding relief rather than a fundamental reversal; Snowflake, Salesforce, and ServiceNow on the software side have delivered intensive AI commercialization results in Q2, with forward-looking indicators like remaining performance obligations and annual recurring revenue improving, and software/semiconductor valuation ratios still at historic lows.
According to strategy views from Huatai Securities and Galaxy Securities on June 15, the fundamentals of high-growth tech tracks remain intact, and the current phase adjustment offers a window for repositioning. Overall tech holdings have not reached historic highs—the S&P 500’s non-tech holdings remain flat at low levels—meaning there’s still room for further sector rotation. The next focus will be: whether Q2 earnings season can validate the pace of AI commercialization, and how macro variables like Fed rate path changes and large IPO capital diversion will affect liquidity in the second half.
From an investment strategy perspective, rather than choosing between "chasing chip highs" or "bottom-fishing software," it’s better to base capital allocation strategies on the tech sector as a whole and seek a balanced allocation between the two subsegments. Investors can use Gate’s stock trading feature to build a portfolio among these seven stocks based on their own risk preferences, flexibly participating in this structural rotation within the tech sector.
