NVIDIA Strategic Analysis: Focus on AI Chips and the "Non-Cloud" Niche

Core Conclusions (Based on Verified Data):

• NVIDIA’s long-term growth remains driven by data center/AI chips, with revenue highly concentrated here (data center accounts for ~88.3% in FY2025) [0]
• The company clearly positions itself as a “chip and platform supplier” rather than directly competing with public clouds (AWS, Azure, GCP). This “non-cloud” positioning reduces customer conflict risks and is more conducive to ecosystem expansion
• However, there are valuation and structural risks: The current price is significantly higher than the DCF intrinsic value range under multiple scenarios (median ~$86.33 vs current $190.53) [0]; meanwhile, self-developed alternatives by key customers, trade policies, gross margin, and profit sustainability constitute medium-to-long-term uncertainties

Note: Regarding the claim that “NVIDIA planned to build a cloud computing service competing with AWS and abandoned it last week, restructuring its cloud computing team”, this search failed to obtain verifiable sources; therefore, the following will not infer this specific event, but analyze the “non-cloud” positioning and recent verifiable strategic moves.

I. Strategic Positioning and the Choice of “Non-Cloud” (vs Common Misconceptions)

• Media and industry analyses widely point out: NVIDIA insists on not operating its own public cloud services, but instead sells GPUs and platform solutions to cloud vendors and enterprise customers, avoiding direct competition with customers and strengthening ecological collaboration [1][2]
• Industry reports have mentioned that the company explores “extending to the upper end of the value chain” (e.g., DGX Cloud hosting/joint operation models for enterprise users), which is fundamentally different from “being a direct competitor of AWS” [1][2]
• Recent strategic moves focus on enhancing inference and system capabilities: In December 2025, it reached a ~$20 billion technology licensing deal (not acquisition) with Groq, acquiring its LPU technology and absorbing key talents to enhance inference-side capabilities and system-level competitiveness [2][3]

II. Reassessment of Long-Term Growth Logic

• Revenue Structure and Concentration:
  • FY2025 data center revenue is ~$115.19B, accounting for ~88.3%; Gaming, Professional Visualization, Auto, etc., account for ~11.7% in total [0]
  • Net profit margin ~53.01%, operating profit margin ~58.84%, with extremely strong profitability, but it also attracts industry discussions about “unsustainable high gross margins” [0][1]
• Growth Drivers (Positive):
  • Cloud computing and AI infrastructure investment are still in a high boom cycle, with training + inference demand continuing to expand
  • The Groq deal and others help fill the gap in “low-latency inference” and increase customer stickiness to NVIDIA’s solutions on the inference side [2][3]
  • System-level products and ecosystem (NVLink/networking, software stack, and toolchain) constitute significant migration costs [2][4]
• Growth Drivers (Risks):
  • Self-developed and multi-sourcing by key customers: Large cloud vendors and enterprises are increasing their efforts in self-developed ASICs to reduce reliance on a single GPU supplier [1][2][4]
  • Inference pattern restructuring: When large model deployment shifts to high concurrency and low latency scenarios, dedicated ASICs and customized solutions may have cost/efficiency advantages in specific workloads [2][4]
  • Policy and geopolitical variables: Export controls, market access, and compliance costs continue to bring revenue fluctuations and regional uncertainties [1][4]

III. Impact on Investment Valuation

• Current Valuation Perspective:
  • Latest share price $190.53, market cap ~$4.64 trillion, P/E TTM ~46.7x–47.2x [0]
  • DCF multi-scenario (conservative/neutral/optimistic) intrinsic value median ~$86.33, lower than current price; indicating that the valuation implies strong expectations for long-term high growth and sustained high profitability [0]
  • Analysts’ median target price ~$257.5, showing the market remains optimistic in the short term [0]
• Valuation Implications of Strategic Adjustments (Focus on Core + Strengthen Inference):
  • Positive: Resource concentration helps consolidate the moat of the core chip and system business; filling the inference capability gap through deals like Groq increases the premium space and customer lock-in for end-to-end AI solutions [2][3]
  • Still to be observed: The penetration pace of customers’ self-developed and alternative solutions, and the evolution path of gross margin against the backdrop of intensified competition
• Risk Factors Summary:
  • Industry and Competition: Accelerated self-developed ASICs by key customers, increased competitors on the inference side (e.g., TPU, various inference-dedicated chips), which may erode market share/profits [1][2][4]
  • Policy and Supply Chain: Export controls, localization alternatives, and compliance requirements exert pressure on the global revenue structure [1][4]
  • Valuation and Cycle: Current valuation has factored in long-term high growth expectations; if industry growth or company gross margin falls short of expectations, there is a risk of valuation correction [0]

IV. Key Uncertainties and Scenario Analysis

• Baseline Scenario (Probability-weighted neutral): AI training and inference demand continues to grow, but intensified competition leads to a moderate decline in gross margin; the company maintains leading advantages through system-level products and ecosystem
• Upside Scenario: Accelerated integration of Groq and others, significant improvement in inference market share and profitability; cloud/enterprise customers still regard NVIDIA’s solutions as the preferred platform
• Downside Scenario: Rapid expansion of self-developed and alternative solutions by key customers, accelerated decline in gross margin; or geopolitical and regulatory constraints on access to key markets

V. Key Takeaways for Investors

• Short-term: Technicals show volatility/no clear trend; key support ~$181.31, resistance ~$193.39; news (e.g., Groq deal and regulatory developments) may increase volatility [0][2][3]
• Medium-term: Growth momentum remains strong, but need to focus on tracking changes in gross margin, penetration rate of self-developed solutions by key customers, and evolution of inference market share
• Long-term: Under the positioning of “non-cloud, focus on chips and platforms”, its long-term moat depends on: 1) Iteration pace of GPU/system-level technology; 2) Integration effect of inference side and ecosystem; 3) Ability to respond to customers’ self-developed solutions and geopolitical risks
• Valuation Advice: The current share price has a large gap relative to DCF implied value; need to carefully balance between strong growth expectations and potential retracement; if pursuing a margin of safety, wait for a better entry point or choose to participate in batches, and closely track competition and policy variables [0][1][2][3][4]

Risk Disclaimer: The above conclusions do not constitute investment advice; please make independent decisions based on your own risk tolerance and investment objectives.

References
[0] 金灵API数据（NVDA实时报价、公司概览、财务分析、DCF估值、日线价格与技术分析）
[1] The Motley Fool – Meta Negotiates Billions in Google TPU Deals, Nvidia Stock Drops 4%（https://www.fool.com/investing/2025/12/27/meta-negotiates-billions-in-google-tpu-deals-nvidia-stock-drops-4/）
[2] WebProNews – Nvidia Strikes $20B Deal with Groq for AI Chip Tech and Key Hires（https://www.webpronews.com/nvidia-strikes-20b-deal-with-groq-for-ai-chip-tech-and-key-hires/）
[3] TechSpot – Nvidia’s $20 billion Groq deal looks a lot like an acquisition in disguise（https://www.techspot.com/news/110723-nvidia-20-billion-groq-deal-looks-lot-like.html）
[4] 经济观察报 – 英伟达真正的对手是谁（http://www.eeo.com.cn/2025/1222/774452.shtml）