Photo by Glsun Mall on Unsplash
Photo by Unsplash
What Just Happened
99%. As of June 29, 2026, the Invesco PHLX Semiconductor ETF (SOXQ) has returned exactly that much year-to-date — a figure that would read like a misprint if it weren't showing up in real brokerage accounts. Its larger counterpart, the iShares Semiconductor ETF (SOXX), is up 54.7% over the same first five months of the year, according to publicly available market data. Both moves trace back to a single source: an AI infrastructure spending wave with no modern precedent.
Wolfe Research, as of June 29, 2026, is leaning further into that bet. The firm reaffirmed its bullish stance on U.S. equities for the second half of the year, naming technology — and semiconductor companies in particular — as the principal driver of earnings growth going forward. According to Investing.com, Wolfe expects markets to "grind higher through the summer as lower energy costs ease pressure on consumers," adding that any pullbacks are likely to be brief as long as retail fund inflows remain strong. The firm identifies U.S.-Iran tensions as the primary geopolitical risk to that outlook.
In specific terms: Wolfe raised its price target on Micron Technology to $1,500, up from $1,250. That target is anchored to roughly 9x the firm's fiscal year 2027 earnings-per-share estimate of approximately $173 for Micron. The firm also forecasts that AI-capable PCs will account for 60% of Windows notebook shipments in H2 2026 — roughly double the 30% share recorded one year prior.
The $725 Billion Machine — and What It Means for Your Portfolio
Here is where the numbers get large enough to require a translation. As of 2026, the four major hyperscalers — Amazon (approximately $200 billion), Microsoft (approximately $190 billion), Google ($175–185 billion), and Meta ($115–135 billion) — are collectively projected to spend $725 billion on AI infrastructure, according to a breakdown compiled by ValueAddVC. In 2025, that combined figure stood at $410 billion. The math works out to a 77% year-over-year increase — roughly $315 billion in additional chip orders, server buildouts, and networking gear landing in a single calendar year.
Chart: Combined AI infrastructure capex from Amazon, Microsoft, Google, and Meta — 2025 actual vs. 2026 projected. Source: ValueAddVC analysis, as of June 29, 2026.
For a beginner investor, here is the kitchen-table version: imagine a city that decided to expand its entire road network by 77% in twelve months. Every concrete supplier, steel mill, and equipment company in that supply chain suddenly has a two-year backlog. Semiconductor companies are the concrete and steel of the AI buildout. When hyperscalers commit $725 billion, the chips powering those data centers have to come from somewhere — and the companies making them are reporting earnings that reflect exactly that demand surge.
Gartner adds texture from the consumer side. As of June 2026, the research firm projects that AI-capable PCs will reach 143 million units shipped this year, representing 55% of the total PC market — up from 31% (77.8 million units) in 2025. Gartner analyst Ranjit Atwal noted that "AI PCs are reshaping the market," though he flagged that adoption in 2025 was slowed by tariffs and buying pauses driven by market uncertainty. The projected 2026 acceleration suggests those obstacles have largely cleared. Wolfe's own estimate goes slightly further, specifically projecting 60% of Windows notebook shipments will carry AI capability in H2 2026.
At the industry level, the global semiconductor market is forecast to reach between $975 billion and $1 trillion in 2026, with AI-specific chips approaching $500 billion — roughly 50% of total industry revenue. Wolfe projects 30% growth in the CPU market driven by AI through 2028, framing this as a multi-year structural shift rather than a one-cycle phenomenon. The firm also identifies NVIDIA, Micron, Broadcom, Microsoft, Alphabet, Meta, Amazon, and Apple as the names contributing disproportionately to overall earnings growth — which means that for most broad index funds, a significant share of returns is already correlated to this cluster. As Smart Wealth AI recently highlighted in its breakdown of 401(k) performance data, the semiconductor bull case is already embedded in millions of retirement accounts, whether or not their holders realize it.
Where the Bull Case Could Break
Wolfe's optimism is not universal. Deloitte has issued a pointed caution: "If monetization timelines slip or constraints in data center buildouts hamper growth, the entire industry could suffer" — a concern rooted in the widening gap between capital deployed and revenue generated from AI products. Microsoft itself has acknowledged it expects to remain capacity-constrained through at least 2026 as GPU, CPU, and storage infrastructure deployment continues to lag actual demand. That is the kind of operational friction that should temper uncritical enthusiasm.
Goldman Sachs offers a counterpoint worth taking seriously. The firm's strategists argue that stock price gains among large-cap AI companies are "backed by actual profit growth," and that forward P/E ratios (the stock price divided by anticipated earnings per share) remain well below the levels seen during the dot-com era of the late 1990s. Goldman's structural comparison to 1999 doesn't hold up, in the firm's view, because this cycle's earnings are real.
The geopolitical layer adds further complexity. South Korea announced a major AI and chip investment push in June 2026 to compete in the global semiconductor race. Chinese AI and chip firms are fueling an onshore IPO rebound. And in late 2025, the Dutch government seized Nexperia — splitting it into two entities — prompting China to respond with export controls on Nexperia components. The U.S. semiconductor sector is operating in an increasingly contested global market, where policy shifts can disrupt supply chains faster than earnings models can adjust.
In my analysis, Goldman's earnings-backed argument holds the stronger near-term foundation — the comparison to 1999 simply does not fit when you look at actual profit margins. But Deloitte's ROI-timeline warning deserves serious weight for anyone building a concentrated semiconductor position. Some industry observers place the timeline for AI's truly transformative commercial applications at around 2030, which means this thesis demands a patience horizon that most retail investors underestimate going in.
Three Moves to Make This Week
Pull up your 401(k) or brokerage account and check the top ten holdings of any S&P 500 or total-market fund you own. The AI-semiconductor cluster — Nvidia, Microsoft, Apple, Alphabet, Meta, Amazon, Broadcom — likely represents a substantial slice. This is not necessarily a problem for your personal finance goals, but knowing your actual sector exposure is the first step in any honest financial planning conversation.
Both are semiconductor ETFs, but their year-to-date returns through June 2026 — 54.7% for SOXX versus 99% for SOXQ — diverged sharply because of different weighting methodologies. Before adding either to your investment portfolio, review the top five holdings by weight. If the same three companies dominate, your apparent diversification is largely cosmetic, and a single-name correction will hit harder than the ETF wrapper implies.
Rather than chasing semiconductor stocks after a 99% run, decide now what portfolio weight you are comfortable holding in tech and chips combined. If that weight gets exceeded — say, a trigger you set at 25% of your total investment portfolio — rebalance mechanically regardless of what the stock market today is doing. Rules made in calm markets protect you from decisions made in euphoric ones.
Frequently Asked Questions
How much are tech companies spending on AI in 2026?
As of June 29, 2026, Amazon, Microsoft, Google, and Meta are collectively projected to spend approximately $725 billion on AI infrastructure — broken down as roughly $200 billion (Amazon), $190 billion (Microsoft), $175–185 billion (Google), and $115–135 billion (Meta), according to data compiled by ValueAddVC. That total is up 77% from the estimated $410 billion those same four companies spent in 2025.
Are semiconductor stocks still a good investment after such large gains in 2026?
As of June 29, 2026, Wolfe Research maintains a bullish outlook on semiconductors and names technology as the primary earnings-growth driver for H2. Goldman Sachs strategists have argued that valuations are supported by real earnings — not speculation — and that forward P/E ratios (stock price divided by anticipated earnings) remain well below dot-com era levels. Deloitte, however, has warned that a slip in AI monetization timelines could expose the sector to broad losses. This is editorial analysis only, not financial advice — consult a licensed advisor before making any investment decisions based on this material.
What are the biggest risks to semiconductor stocks in 2026?
Four risks stand out across current reporting: (1) AI return-on-investment timelines — Deloitte warns the industry could suffer broadly if monetization slips; (2) supply chain constraints — Microsoft expects to remain capacity-constrained through at least 2026; (3) geopolitical friction — export controls, tariffs, and moves like the Dutch government's late-2025 seizure of Nexperia show how quickly policy can disrupt supply chains; and (4) intensifying global competition — South Korea and China are both accelerating domestic semiconductor investment programs as of June 2026.
- SOXQ is up 99% year-to-date; SOXX is up 54.7% — both reflecting the largest coordinated AI infrastructure investment cycle in history.
- Wolfe Research is bullish on equities for H2 2026, with semiconductors named as the principal earnings engine; Micron's price target raised to $1,500 from $1,250.
- Combined hyperscaler AI capex hits $725 billion in 2026 — a 77% jump from $410 billion in 2025 — flowing directly into the hardware that chip companies produce.
- Concentration risk is real: Deloitte warns of sector-wide exposure if AI monetization timelines slip, even as Goldman Sachs argues that earnings growth validates current valuations.
Disclaimer: This article is for informational purposes only and does not constitute financial advice. It represents original editorial commentary on publicly reported information. Research based on publicly available sources current as of June 29, 2026.