Key Takeaways

• AI Server Market Expansion: The AI server market is projected to grow nearly 4x from $55.2B in 2023 to $208.4B by 2027, with AI servers increasing from 9.5% to 19.5% of total server shipments. This provides a massive growth runway for NVIDIA’s core AI business.
• GPU Revenue Growth: Server GPU revenue is expected to grow from $35.1B in 2023 to $99.8B by 2027. With NVIDIA maintaining the vast majority of market share, this translates to a huge revenue opportunity in AI compute alone by 2027, not including software and networking revenue streams.
• Increasing GPU Intensity: The number of GPUs per AI server is trending up from 3.06 in 2022 to 4.4 by 2027, while GPU ASPs are maintaining around $7,000 (a la carte estimate), indicating both growing compute demands and strong pricing power for high-performance AI accelerators.
• Near-Term Momentum: Despite concerns about a potential plateau, Q3 results ($35.1B revenue, +94% YoY) and Q4 guidance ($37.5B) exceeded expectations, with management describing Blackwell demand as “staggering” and emphasizing no slowdown in AI model scaling – suggesting continued strong growth through the Blackwell transition.

NVIDIA delivered another impressive quarter with revenue of $35.1B (+94% YoY, +17% QoQ), beating consensus estimates of $33.0B. The outperformance was driven primarily by Data Center revenue of $30.8B (+112% YoY, +17% QoQ), which continues to see robust demand for AI computing solutions. Gaming and Auto segments also exceeded expectations, with Gaming at $3.28B (+15% YoY) and Automotive at $449M (+72% YoY).

Looking forward, NVIDIA guided Q4 revenue to $37.5B (+7% QoQ), ahead of consensus at $36.5B. While this represents a moderation in sequential growth, the underlying market dynamics strongly support continued momentum. Market data shows AI server shipments are expected to grow from 1,103,000 units in 2023 to 3,207,000 units by 2027, representing a 28.5% CAGR. More importantly, AI-optimized server revenue is projected to surge from $55.2B in 2023 to $208.4B by 2027, highlighting the dramatic increase in value per system.

Chart: Projected growth of AI servers as a percentage of total servers annually. 

The above chart highlights the upside for Nvidia as the datacenter goes through a transformational shift from general-purpose servers to AI servers. The GPU opportunity within these AI servers is particularly compelling for NVIDIA. Server GPU revenue (a la carte) is forecast to grow from $35.1B in 2023 to $99.8B by 2027. With NVIDIA commanding the vast majority of the AI GPU market. These estimates are GPU only, where Nvidia has revenue upside from total systems like DGX, etc.

Several structural factors support this growth trajectory. First, the number of GPUs per AI server is increasing, from 3.06 in 2022 to an estimated 4.4 by 2027, indicating more compute-intensive workloads.

While gross margins are expected to temporarily decline to the low 70s in early 2025 due to Blackwell’s initial manufacturing complexity, management expressed confidence in returning to mid-70s levels in the second half as production scales. This is supported by the increasing mix of high-value AI products – AI servers are expected to grow from 9.5% of total server shipments in 2023 to 19.5% by 2027.

The company’s competitive position remains strong, with recent concerns about potential Blackwell overheating issues appearing unfounded according to management. The fundamental thesis centers on NVIDIA’s role in enabling the estimated $1 trillion of existing data center infrastructure that needs to be upgraded for AI and accelerated computing workloads. Additionally, AI compute demands are expanding beyond just pre-training to include more test-time compute and post-training optimization.

The main risks to monitor include potential digestion periods as customers absorb AI infrastructure investments, manufacturing constraints that could limit Blackwell supply through 2025, and growing competition in AI chips. However, given the massive scale of AI computing demand, NVIDIA’s entrenched ecosystem advantages, and the company’s dominant market share in AI GPUs, the path to sustained growth appears well-supported by market fundamentals.

Wall Street analysts maintain NVIDIA as a top pick in semiconductors with price targets ranging from $168-185, citing strong visibility into the Blackwell ramp and the company’s leading position in an expanding AI computing market. The market data suggests this optimism is well-founded, with AI server and GPU growth trajectories supporting significant revenue expansion through 2027.

For more context, see my videos on Yahoo Finance and Schwab Network.

Join the newsletter and stay up to date