NVDA vs TSM: How NVIDIA and Taiwan Semiconductor Compare (2026)

Same yardstick, side by side (as of August 2026). Valuation lined up like this is most meaningful for two names in the same corner of the market, which these are. Figures are approximate; verify before investing.

Reading it: NVDA is the cheaper of the two on forward earnings, but cheaper is not the same as better. Pair the valuation with growth (how far the forward P/E sits below the trailing P/E) and risk (beta) before you decide.

The metrics above tell you which is the marginally better business. The bigger risk for most people is not picking the slightly worse stock, it is over-concentrating. NVDA and TSM share themes, so owning both, or adding either to what you already hold, can quietly push a large share of your portfolio into one bet.

This is the part a generic comparison page cannot answer, because it depends on what you own. Connect your brokerage and Walnut shows your real, combined NVDA and TSM exposure, flags overlap with your existing positions, and tells you if adding one would tip you past a concentration you are comfortable with, read-only by default, with your login staying at your broker. Walnut is not an investment adviser.

NVIDIA (NVDA) designs the graphics processing units (GPUs) and the software stack that have become the standard compute platform for modern artificial intelligence. The company operates across four reporting segments. Data Center sells GPUs to the major cloud providers (Microsoft Azure, AWS, Google Cloud, Oracle, Meta) and to AI labs (OpenAI, Anthropic, xAI) for training and running large language models; this is now roughly 85% of revenue. Gaming covers GeForce consumer GPUs, NVIDIA's original core market. Professional Visualization sells workstation GPUs for design and simulation, and Automotive ships the DRIVE platform for assisted and autonomous driving. NVIDIA also builds CUDA, the proprietary software platform that lets developers write code that runs on its GPUs and that AI frameworks like PyTorch, TensorFlow, and JAX target first. Founded in 1993 by Jensen Huang, Chris Malachowsky, and Curtis Priem, headquartered in Santa Clara, California, and led by co-founder and CEO Jensen Huang, NVIDIA is one of the most valuable companies in the world. It designs its chips and outsources manufacturing primarily to TSMC.

Full NVDA guide

Taiwan Semiconductor Manufacturing Company (TSMC) is the largest dedicated semiconductor foundry in the world. TSMC is the company that physically manufactures the chips designed by NVIDIA, AMD, Apple, Broadcom, Marvell, Qualcomm, MediaTek, and most other fabless semiconductor companies. The company operates the most advanced semiconductor manufacturing technology in the world, including leading-edge process nodes at 3nm and 2nm.

Full TSM guide

Both fit overlapping themes, but they are not interchangeable. The useful comparison is which set of drivers and risks you want exposure to.

• NVDA drivers: Continued AI infrastructure dominance; Beyond hyperscalers: sovereign AI and enterprise.
• TSM drivers: AI accelerator capacity is the binding constraint; Geographic diversification.

A faster-growing, richer-valued name usually swings harder, so it suits a longer horizon and a higher tolerance for volatility; a steadier, more cash-generative business suits a more conservative or income-minded investor. The honest test is which set of risks you could hold through a drawdown: Customer concentration is high: the top four or five hyperscalers account for roughly half of revenue, so any slowdown in their AI capex hits results directly. For TSM, geopolitical risk is the single largest factor: any escalation between China and Taiwan would severely impact global chip supply.

Growth-minded investors who believe the theme has years to run tend to accept the richer multiple for more upside; value-minded investors lean toward the cheaper forward earnings and steadier profile. Pick NVDA if you believe its drivers more; TSM if you believe its. Many investors hold both, but since they share themes, that is a concentrated bet, not diversification. Decide deliberately and check overlap. For the full detail, see the NVDA and TSM guides.

NVDA. The P/E of around 50x is the headline most investors focus on, well above the S&P 500's roughly 22x. The market tolerates it because of the combination of triple-digit revenue growth, 60%+ operating margins, and a software moat (CUDA) that competitors keep failing to match. The risk is symmetrical: if hyperscaler capex slows or the AI buildout decelerates, the valuation compresses very quickly. All figures are approximate as of early 2026 and refresh quarterly; verify against NVIDIA's investor relations page or your broker.

TSM. TSMC's P/E of approximately 25x reflects a balance: its AI-driven leading-edge demand justifies a premium, but geopolitical risk caps the multiple compared to fabless designers like NVIDIA (~50x) or Broadcom (~45x). The risk discount is the unusually wide gap with NVIDIA's multiple.

Headline figures (approximate, early 2026): NVDA shows revenue (fy2026 ending jan) ~$130 billion, having grown 100%+ in recent quarters, operating margin ~63%, exceptionally high for a hardware company, net income ~$70 billion, eps (ttm) ~$2.80; TSM shows revenue (ttm) ~$95 billion, operating margin ~45%, net income (ttm) ~$38 billion, eps (ttm) ~$7.50.

NVDA and TSM are related but distinct: same themes, different businesses and risks. Neither wins in the abstract; the right pick is whichever thesis you actually believe, sized so you are not over-concentrated in one theme. Walnut can show your combined NVDA and TSM exposure against your real portfolio. It is not an investment adviser.