NVIDIA's latest earnings tell two very different stories depending on where you look. The headline numbers are undeniably impressive: $68.13 billion in Q4 revenue, 75% year-over-year growth in data center business, and a 75.2% gross margin that would make most technology companies envious. CEO Jensen Huang is calling this an "AI industrial revolution" and saying "Blackwell sales are off the charts." But beneath those spectacular numbers lies a concentration of risk that seasoned investors are beginning to scrutinize more carefully.

Why Is NVIDIA's China Revenue Suddenly Invisible in Forward Guidance?

The most telling detail in NVIDIA's Q1 guidance may be what's missing. The company is projecting roughly $78 billion in revenue for the next quarter, but that figure explicitly excludes any data center compute revenue from China. When a company zeroes out an entire major geography in forward guidance, management is essentially telling investors that the regulatory picture is too uncertain to model. Add to that an $8 billion hit from H20 revenue in Q2, a $4.5 billion charge before it, and you have a stock trading at 46x trailing price-to-earnings with concentrated geopolitical exposure.

The production concentration makes this vulnerability even sharper. The bulk of Blackwell production remains tied to TSMC (Taiwan Semiconductor Manufacturing Company), a single foundry in Taiwan. This means NVIDIA's growth story depends not just on sustained AI spending, but on Washington's Taiwan posture and the export-controls regime remaining stable. Until that clarity arrives, investors are essentially paying for a narrative the company cannot fully underwrite.

What Makes NVIDIA's Competitive Position More Fragile Than It Appears?

NVIDIA's dominance in AI infrastructure GPUs remains real, and the company has exceeded revenue estimates for 28 consecutive quarters. Hyperscaler spending is massive; Amazon, Microsoft, Alphabet, and Meta collectively plan to spend over $700 billion on AI infrastructure this year, with NVIDIA remaining the primary beneficiary. Next-generation chips like Blackwell and Rubin are expected to drive the next wave of data center growth.

But competition is intensifying in ways that matter. AMD, Alphabet's internally developed AI chips, and emerging companies like Cerebras Systems are all chipping away at NVIDIA's moat. Cerebras stock surged nearly 70% following its IPO last week, giving the company a valuation approaching $70 billion and raising questions about whether specialized AI compute platforms could eventually erode NVIDIA's dominance. NVIDIA's software ecosystem, CUDA platform, and scale advantages remain significant competitive moats, but the window for that advantage to persist is narrowing.

NVIDIA also carries roughly $117 billion in non-cancelable supply commitments, which could compress margins if hyperscale spending slows. The company has a history of falling after earnings, even when results are excellent, a dynamic CEO Jensen Huang has publicly acknowledged.

How to Evaluate NVIDIA Against More Diversified Alternatives

• Revenue Durability: NVIDIA's revenue must be re-won every product cycle, whereas Microsoft's commercial remaining performance obligation hit $627 billion, up 99% year-over-year, representing signed customer money that provides a more durable foundation.
• Business Diversification: NVIDIA derives roughly 91% of revenue from one segment, while Microsoft has Azure growing 40% year-over-year, an AI business at $37 billion annual run rate, and Microsoft 365 Copilot at over 20 million paid seats, plus Office, LinkedIn, GitHub, Dynamics, and Xbox generating steady cash flow.
• Geopolitical Insulation: Microsoft holds a stake in OpenAI worth roughly $135 billion, has a contracted $250 billion incremental Azure commitment from OpenAI, and secured a "frontier model royalty-free with all the IP rights" through 2032, providing AI exposure without Taiwan dependency.

Bill Ackman's Pershing Square has reportedly been building a position in Microsoft, and the setup looks more durable for long-duration investors. Microsoft also pays you to wait, returning real capital through a $3.56 annual dividend versus NVIDIA's nominal $0.04. Operating margins of 46.3%, return on equity of 34%, and a debt-to-equity ratio of 0.33 describe a fortress balance sheet.

NVIDIA's stock has pulled back roughly 6% from its recent record high of $236 per share, reached last week amid optimism that global AI spending continues to accelerate while restrictions on chip sales to China stabilize. Consensus forecasts call for NVIDIA's Q1 revenue to reach $78.75 billion, representing 79% year-over-year growth, with gross margins projected above 74%. On the bottom line, analysts expect Q1 earnings to surge 118% to $1.77 per share.

Despite these risks, bullish sentiment around NVIDIA remains intact. The company continues to dominate one of the fastest-growing areas of technology, while hyperscaler AI spending appears to be accelerating rather than slowing. NVIDIA stock currently carries a Zacks Rank #2 (Buy), supported by positive earnings estimate revisions for its current fiscal year 2027 and 2028. Even if NVIDIA experiences a post-earnings pullback, long-term investors may ultimately view any weakness as a buying opportunity.

The real question for investors is not whether NVIDIA can deliver strong results, which Wall Street largely expects, but whether management can convince investors that AI infrastructure demand remains strong enough to offset geopolitical risks and intensifying competition. NVIDIA offers exposure to the AI infrastructure boom, but Microsoft offers a more diversified setup for that same exposure, while NVIDIA's risk-reward may look clearer once Washington's Taiwan posture is no longer a guessing game.