I’ve Covered Semiconductors for 5 Years. Nvidia Is My No. 1 Pick Heading Into 2027.

Nvidia is rapidly gaining ground in these newer businesses. According to IDC, Nvidia’s data center Ethernet switching revenue grew 192.7% year over year to $2.1 billion in the first quarter of calendar year 2026. Nvidia accounted for 21.5% share of the data center Ethernet switch market. The company’s networking revenue also rose 18% sequentially, with Spectrum-X Ethernet revenue increasing 2.6 times year over year in the second quarter.

Nvidia’s server CPU business is also gaining momentum. The company’s Grace CPU revenue exceeded $5 billion on a trailing-12-month basis, while the next-generation Vera CPU was already in full production at the end of the second quarter. Management continues to see around $20 billion in total server CPU demand and preliminarily expects CPU revenue to more than double in fiscal 2028.

Advanced Micro Devices and major cloud providers are developing competing AI accelerators, so Nvidia may not retain its current share of the accelerator market. But that does not necessarily mean its AI revenue opportunity will shrink. Hence, I think investors should focus less on Nvidia’s GPU market share alone and more on how much of each AI data center the company can monetize.

Nvidia’s growth is impressive for its size

Nvidia is growing revenues at an impressive pace despite its already massive size. The company is guiding for third-quarter revenue of $108 billion, plus or minus 2%, implying year-over-year growth of roughly 89% at the midpoint.

Additionally, Nvidia’s longer-term growth outlook is also limited by supply rather than demand. Management said its fiscal 2028 revenue outlook is supply-constrained and that customer forecasts indicate even faster growth. Nvidia expects supply to remain a bottleneck through the end of fiscal 2028.

Nvidia is also moving quickly to its next product generation. The company began production shipments of Vera Rubin systems in August 2026. The company has also received purchase orders from every major hyperscaler, AI cloud, and system manufacturer. Management expects Vera Rubin to account for around 20% of data center revenue in the third quarter.

I think this matters because major product transitions can create execution risks for semiconductor companies. Nvidia, however, is ramping up Vera Rubin while demand for Blackwell remains strong among hyperscalers. As a result, the company’s size does not yet appear to be slowing its growth.

Nvidia’s valuation appears reasonable

I like several other semiconductor stocks, including Broadcom, Taiwan Semiconductor Manufacturing, and Advanced Micro Devices. However, Nvidia still offers the combination of scale, growth, and valuation I prefer heading into 2027.

Nvidia was trading at around 23.2 times analysts’ expected fiscal 2027 adjusted earnings per share (EPS) of $9.05 and nearly 16 times expected fiscal 2028 EPS of $13.13, as of Aug. 26. While the stock is not exactly cheap, I still find it difficult to call it excessively expensive when management expects revenue to grow by around 70% in fiscal 2028 and demand continues to exceed available supply.

A risk that cannot be ignored

Nvidia has made significant commitments to secure primarily memory and manufacturing capacity for current and future data center products. The company’s supply and capacity commitments increased from $119 billion at the end of the first quarter to $279 billion at the end of the second quarter of fiscal 2027. However, some of these agreements can be canceled, rescheduled, or adjusted before firm orders are placed.

The strategy makes sense while customers want more AI infrastructure than Nvidia can supply. However, it also makes correctly forecasting future demand increasingly important. If AI spending slows unexpectedly, Nvidia could end up with more supply commitments than it needs.

But for now, I don’t see evidence of that slowdown. Nvidia does not need to win every part of the semiconductor market for the stock to work. It needs to keep increasing the revenue it generates from each AI data center while converting that growth into profits.