To The Moon

Early this morning, $NVIDIA(NVDA)$ released its Q4 FY26 earnings report covering the three months ending January 25.

Compared to analyst expectations, the fourth-quarter performance was impressive:

Looking ahead to the first quarter of fiscal year 2027, NVIDIA expects revenue to reach approximately $78 billion, a massive 77% year-over-year increase that significantly surpasses analysts' projections of $72.8 billion.

Despite the explosive earnings report, NVIDIA's stock remained largely unchanged in after-hours trading, edging up just 0.18%!

Following this, semiconductor ETFs experienced minor corrections after hitting record highs overnight. The largest semiconductor ETF, $VanEck Semiconductor ETF(SMH)$ , fell 0.43% in after-hours trading, $iShares Semiconductor ETF(SOXX)$ dropped 0.35%, and FTXL declined 0.66%.

With robust earnings colliding with a lukewarm stock price, where do NVIDIA and semiconductors go from here? Is there still opportunity in related ETFs?

According to NVIDIA's Q4 report, quarterly revenue reached $68.1 billion, significantly exceeding analysts' expectations of $65.9 billion and surging 73.2% year-over-year:

By segment, data center revenue reached $62.3 billion, accounting for 91.5% of total revenue, up 75% year over year. The traditional gaming segment generated $3.7 billion, rising 46.5% year over year, but its share of total revenue has declined to 5%.

Professional visualization, automotive, and other segments delivered mixed results in the fourth quarter. Given their minimal revenue contribution and limited impact on the stock price, they will not be analyzed in detail here:

In terms of profitability, NVIDIA's gross margin reached 75% in the fourth quarter, adjusted to 75.2%, exceeding analysts' expectations of 74.7%:

From this perspective, NVIDIA's latest earnings report outperformed expectations both in current performance and future guidance, maintaining its consistent track record of excellence.

In terms of valuation, NVIDIA's price-to-sales ratio stands at 22 times, positioning it near the lower end of its range over the past five years:

From a price-to-earnings ratio perspective, NVIDIA trades at 45 times earnings. While the absolute value is relatively high, NVIDIA's rapid growth is expected to bring its P/E ratio down to 26 times by year-end, making it reasonably priced.

Given this, why hasn't NVIDIA sparked a new rally? Instead, its stock price has fluctuated near peak levels for nearly half a year.

The underlying reason is primarily investor concern that NVIDIA's strong performance may not last.

On one hand, tech giants' capital expenditures are projected to reach $650 billion this year, a substantial year-over-year increase of approximately 60%:

Tech giants account for roughly 50% of Nvidia’s total revenue. Their aggressive spending has clearly supported Nvidia’s strong earnings growth this year. However, pressure is emerging on free cash flow at some of these companies. Analysts expect Amazon’s free cash flow to turn negative this year. That raises questions about how long the current pace of capital expenditure can be sustained, especially as AI investments have yet to generate clearly visible returns at scale.

At the same time, the scramble by large tech firms to secure AI chips has created supply shortages for Nvidia’s GPUs. With demand exceeding supply, Nvidia has gained significant pricing power, pushing its net margin to an extraordinary 63%. This profitability has also accelerated efforts by major customers to develop in-house chips and ASICs (custom-designed chips), prompting investor concerns that such high margins may not be sustainable over the long term.

As a result, investor sentiment toward Nvidia has become increasingly polarized. The key to restoring balance lies in whether Nvidia can ease concerns about growth prospects in 2027. That will require clearer signals over the coming quarters regarding demand durability and competitive dynamics.

Importantly, large tech firms are not Nvidia’s only growth engine. Startups and smaller AI companies contribute the remaining 50% of revenue. During the earnings call, Nvidia indicated that this segment is also growing rapidly. Demand is not limited to hyperscalers; emerging AI players are expanding quickly and competing for scarce compute resources.

Founder Jensen Huang highlighted a central theme: “compute is revenue.” He noted that over the past two to three months, agentic AI has reached a genuine inflection point. Products such as Cloud Code, OpenAI Codex and Cursor have seen exponential growth in demand for code-generating agents. Companies like Anthropic have reportedly grown revenue tenfold within a year, yet remain constrained by compute bottlenecks. In this new paradigm, inference equals revenue, and compute equals revenue.

Looking ahead, Huang suggested that after agentic AI, the next wave will be physical AI—bringing artificial intelligence into manufacturing and robotics. This could unlock a significantly larger, trillion-dollar market opportunity. In that context, the market’s focus may not need to rest solely on hyperscaler capital spending, as broader AI adoption continues to expand the addressable demand base.

As long as AI technology continues to advance, demand for computing power is likely to grow exponentially. In AI chips, Nvidia still maintains a dominant position. Reflecting strong market demand, Nvidia’s CFO stated that the company expects to exceed its previously set $500 billion revenue target.

The earnings report continues to show robust AI momentum, with upstream and downstream supply chains benefiting directly. Following the results, SK Hynix, which supplies HBM memory to Nvidia, rose 7.9%, while Samsung Electronics gained 7.1%. Taiwan Semiconductor Manufacturing Co., Nvidia’s key foundry partner, climbed nearly 1% in late trading.

Against this backdrop, semiconductor ETFs remain worth monitoring. The largest is $VanEck Semiconductor ETF(SMH)$ , with assets of approximately $48.6 billion. It offers strong liquidity and a relatively moderate expense ratio of 0.35%. $NVIDIA(NVDA)$ and $Taiwan Semiconductor Manufacturing(TSM)$ are its top holdings, accounting for 18.67% and 11.3% of net assets respectively, and the fund has gained more than 18% year to date.

$iShares Semiconductor ETF(SOXX)$ ranks second only to SMH in size, currently at $22.9 billion. Its management fee of 0.34% is slightly lower than SMH's. Its top holding is $Micron Technology(MU)$ , which has gained 22% year-to-date:

$SPDR S&P Semiconductor ETF(XSD)$ is the third-largest semiconductor ETF by assets under management. While its management fees are relatively low, its top holdings differ from other semiconductor ETFs, favoring smaller semiconductor companies such as $ON Semiconductor(ON)$ and $Analog Devices(ADI)$ . It has not benefited significantly from AI trends, resulting in a year-to-date gain of only 13%.

$First Trust Nasdaq Semiconductor ETF(FTXL)$ has a total scale of $1.7 billion with a slightly higher management fee of 0.6%. However, its heavy weighting in $Micron Technology(MU)$ —a single stock accounting for 17% of its net asset value—has allowed it to fully capitalize on the epic price surge in memory chips, delivering a year-to-date gain of 26%.

$INVESCO SEMICONDUCTORS ETF(PSI)$ has assets under management of 1.4 billion, charges a management fee of 0.56%, and has achieved a year-to-date gain of 33%. Its holdings are concentrated in semiconductor equipment companies such as $Micron Technology(MU)$ and $Lam Research(LRCX)$ :

$Invesco PHLX Semiconductor ETF(SOXQ)$ has $1.1 billion in assets under management with a management fee of 0.19%, positioning it as a relatively low-cost semiconductor ETF. It has gained 20% year-to-date, with holdings heavily weighted toward chip design companies. Its top three holdings are $NVIDIA(NVDA)$ , $Broadcom(AVGO)$ , and $Advanced Micro Devices(AMD)$ . Therefore, if the market re-embraces NVIDIA, SOXQ could see even greater gains.

$XTRACKERS SEMICONDUCTOR SELECT EQUITY ETF(CHPS)$ has a relatively small scale of $44.29 million, but its management fee is only 0.15%, making it the lowest-cost semiconductor ETF! Moreover, CHPS's top holdings lean toward memory chips and semiconductor equipment—its largest holding is $SK Hynix, Inc.(HXSCL)$ , followed by $Applied Materials(AMAT)$ and $Micron Technology(MU)$ . This ETF has surged over 28% year-to-date, delivering outstanding performance: