NVIDIA’s AI Earnings Explosion

NVIDIA’s earnings trajectory from 2022 to 2024 exemplifies how transformative technology shifts can create unprecedented earnings acceleration. As generative AI demand exploded, the company’s data center revenue—which represented 48% of total sales in fiscal 2022—surged to 84% by fiscal 2025, driving net income growth that defied historical semiconductor cycles and reshaped investor expectations for the entire GPU and semiconductor ecosystem.

Quick Definition

Earnings acceleration occurs when revenue and profit growth rates increase dramatically over successive quarters due to structural changes in demand or market dynamics. For NVIDIA, this wasn’t cyclical recovery; it was structural transformation driven by an entirely new category of workloads (large language model inference and training) that required specialized hardware.

Key Takeaways

• Fiscal Q3 2024 (ended October 29, 2023): Revenue hit $18.1 billion, a 101% year-over-year increase, with data center revenue growing 217% YoY to $14.5 billion.
• Fiscal Q4 2024 (ended January 28, 2024): Revenue reached $22.1 billion (+265% YoY), marking the highest single quarter in company history.
• Gross margin expansion: Despite commodity semiconductor cycles, NVIDIA maintained 70%+ gross margins through 2024, signaling pricing power and supply-constrained demand.
• Earnings per share (EPS): Diluted EPS reached $5.24 in fiscal Q4 2024 (ended January 2024), up from $0.29 in the prior-year quarter—an 1,709% increase.
• Market capitalization: NVIDIA’s market cap surpassed $3 trillion in 2024, making it one of the three most valuable public companies, driven largely by earnings surprise and forward guidance.
• Guidance beat rate: NVIDIA beat revenue guidance 5 consecutive quarters (fiscal Q1 2024 through Q4 2025), signaling supply confidence and execution excellence.

The Pre-AI Baseline: Fiscal 2022–2023

March 2022 – March 2023: NVIDIA’s fiscal 2022 (year ending January 2022) reported $26.9 billion in revenue with $4.5 billion in net income. However, the company faced severe headwinds in fiscal 2023:

• Gaming revenue collapsed as cryptocurrency mining demand evaporated post-2021 and consumers faced a saturated GPU market.
• Fiscal 2023 revenue fell to $60.9 billion (retroactively adjusted for the 4:1 stock split in June 2024), but this masked a 53% revenue decline in gaming.
• Data center revenue remained relatively flat at ~$8 billion annually, representing a smaller and slower-growing segment.
• Stock price volatility: NVIDIA stock traded between $155–$300 in early 2023 as investors questioned whether the company could return to growth.

The gaming collapse and nascent data center business created skepticism: could NVIDIA’s peak profitability be behind it?

The Inflection Point: Fiscal Q1 2024 (April 2023)

April 18, 2023: NVIDIA reported fiscal Q1 2024 (quarter ended April 30, 2023) results that signaled the beginning of the AI boom:

• Revenue: $7.2 billion (down 36% YoY from Q1 2023’s $11.3 billion due to gaming collapse).
• Data center revenue: $2.9 billion, up 70% YoY from $1.7 billion.
• Guidance beat: CFO Colette Kress guided to Q2 2024 revenue of $11 billion, signaling recovery and AI demand acceleration.
• Gross margin: 66%, recovering from fiscal 2023’s 56%.

This quarter marked the inflection. Investors who had written off NVIDIA began reconsidering the company’s competitive moat in AI hardware. H100 and H200 data center GPUs faced nearly infinite demand as major cloud providers (AWS, Google Cloud, Azure) and hyperscalers raced to provision generative AI infrastructure.

Acceleration Phase: Fiscal Q2–Q3 2024 (May–October 2023)

May 24, 2023: Fiscal Q2 2024 results announced:

• Revenue: $13.5 billion (+101% YoY), beating guidance of $11 billion by 23%.
• Data center: $10.3 billion (+171% YoY), representing 76% of total revenue.
• Net income: $6.7 billion (operating margin 59%), up from $2.6 billion in Q2 2023.
• Stock price: Surged 30% in the week following earnings, with Wall Street raising price targets to $500+.

August 23, 2023: Fiscal Q3 2024 results announced:

• Revenue: $18.1 billion (+101% YoY), once again doubling growth.
• Data center: $14.5 billion (+217% YoY), now 80% of revenue.
• Operating margin: 62%, sustaining fortress-like profitability.
• Guidance: Q4 revenue expected to reach $20 billion, signaling 5 consecutive quarters of acceleration.

The Peak: Fiscal Q4 2024 (January 2024) and Normalization

January 30, 2024: NVIDIA reported record fiscal Q4 2024 results:

• Revenue: $22.1 billion (+265% YoY), the highest single quarter in company history.
• Data center: $18.4 billion (+417% YoY), representing 83% of revenue.
• Diluted EPS: $5.24 (GAAP basis), up 1,709% from $0.29 in fiscal Q4 2023.
• Operating margin: 63%, peak profitability for the cycle.
• Stock price: Reached $900+ in February 2024, with market cap exceeding $2.2 trillion.

However, management also signaled that explosive growth rates would moderate:

• Guidance: Fiscal Q1 2025 revenue expected at $26 billion (16% sequential growth, decelerating from prior quarters’ 20–30% growth rates).
• CFO commentary: "We expect growth to be more moderate going forward as we anniversary the initial surge in GPU demand and move into a more normalized adoption curve."

This guidance set up fiscal 2024 and 2025 as a normalization period:

• Fiscal 2025 (year ending January 2025): Full-year revenue expected near $120 billion, still 66% higher than fiscal 2024’s $60.9 billion.
• Long-term growth: Management anticipated data center GPU TAM (total addressable market) at $100+ billion annually by 2030, suggesting NVIDIA could sustain $50–$70 billion annual data center revenue by fiscal 2030.

Margin Expansion Despite Scale

A critical earnings surprise in NVIDIA’s AI boom was gross margin expansion despite unprecedented manufacturing scale:

• Fiscal 2023: 49% gross margin (gaming collapse and clearing inventory).
• Fiscal Q1 2024: 66% gross margin (demand recovery).
• Fiscal Q4 2024: 73% gross margin (peak supply constraints and pricing power).

This margin expansion reflected:

1. Product mix shift: Data center GPUs (70%+ gross margins) replaced gaming GPUs (40–50% margins).
2. Supply scarcity: NVIDIA and TSMC (its manufacturing partner) couldn’t produce enough GPUs, allowing price discipline.
3. High-value applications: AI model training and inference represented far higher-value workloads than gaming, justifying premium pricing.

By fiscal 2025, management expected gross margins to normalize to 68–70% as supply constraints eased and competition (AMD, custom silicon from hyperscalers) intensified.

Real-World Examples: Hyperscaler Spending and Competitive Dynamics

Meta’s AI Infrastructure Investment (2023–2024): Meta announced $30+ billion capital expenditure for 2024–2025, primarily to build AI data centers. Analysis suggested $10+ billion annually would flow to NVIDIA for GPU purchases, driving NVIDIA’s Q2–Q4 2024 revenue beats.

OpenAI and ChatGPT Infrastructure: OpenAI’s $10+ billion training and inference costs were borne by Microsoft’s Azure cloud, which purchased tens of thousands of H100 and A100 GPUs from NVIDIA in 2023–2024, contributing to data center revenue acceleration.

Google’s TPU vs. NVIDIA GPU Strategy: Google invested heavily in custom TPU chips in 2024, signaling competitive threat. However, NVIDIA’s fiscal 2024 guidance and beat rates proved the GPU market expansion outpaced any customer silicon erosion.

Common Mistakes in NVIDIA Earnings Analysis

1. Extrapolating Peak-Cycle Growth Indefinitely Investors who bought NVIDIA at $900 in February 2024 expected 100%+ annual growth to persist. In reality, fiscal Q1 2025 guidance implied 16% sequential growth, a deceleration that caused the stock to fall 30% in March 2024. Lesson: Earnings acceleration curves follow S-curves, not exponentials. Inflections create surprises; peaks create disappointment.

2. Ignoring Gross Margin Normalization Peak gross margins of 73% proved unsustainable. As supply constraints eased (fiscal 2024–2025), competitive pressure and product mix shifts suggested normalization toward 65–68%. Investors who modeled 75%+ sustained margins overestimated future earnings power.

3. Confusing Supply Constraints with Demand Infinity In fiscal Q2–Q4 2024, every GPU NVIDIA produced sold immediately. This created perception of infinite demand. In reality, the TAM for AI infrastructure, while large ($300+ billion by 2030 estimates), was still bounded. By fiscal 2025, supply caught up to demand, requiring more complex competitive and pricing analysis.

4. Overlooking Customer Concentration Risk By fiscal Q4 2024, hyperscalers (Meta, Google, Amazon, Microsoft, Apple) represented 55%+ of data center revenue. Investor analysis often missed this concentration: a capex slowdown from one hyperscaler could reduce NVIDIA revenue 10%+ in a quarter.

5. Assuming EPS Growth = Stock Price Growth NVIDIA’s diluted EPS grew from $0.29 (fiscal Q4 2023) to $5.24 (fiscal Q4 2024)—a 1,709% increase. Yet the stock price grew only 200% over the same period because the earnings multiple compressed as investors repriced risk and normalized growth expectations.

FAQ

Q1: Why did NVIDIA’s data center revenue grow 417% YoY in Q4 2024 when overall revenue only grew 265%? Data center represented an increasingly larger proportion of revenue (83% in Q4 2024 vs. 48% in fiscal 2022). Gaming, professional visualization, and automotive segments collectively declined, so data center growth outpaced blended company growth.

Q2: Could NVIDIA’s gross margins stay at 70%+ permanently? Unlikely. Peak margins of 73% reflected supply scarcity and limited competition (AMD and custom silicon faced multi-year lags). As supply normalized and competitive products arrived, gross margins typically compressed toward 65–68%, in line with semiconductor industry norms.

Q3: What was the primary driver of EPS growth: revenue, gross margin, or share buybacks? All three contributed. Revenue increased 22× from Q4 2023 to Q4 2024, gross margins expanded 4 percentage points (66% to 73%), and NVIDIA repurchased ~$50 billion in stock fiscal 2024–2025, reducing diluted share count by ~5%. EPS benefited from all three levers.

Q4: Did NVIDIA’s stock price peak at $900 in February 2024? No. The stock recovered to $1,100+ by September 2024 and $1,200+ by May 2025 as investors realized the AI cycle would sustain for 5+ years. However, the February 2024 peak represented the peak valuation multiple (60x+ forward P/E), not peak price.

Q5: How much of NVIDIA’s growth came from price increases vs. volume? Fiscal 2024 analysis suggested 70% of data center revenue growth came from volume (GPU unit sales increasing 4–5x) and 30% from pricing. As supply constraints eased in fiscal 2025, the ratio shifted toward 80% volume and 20% pricing.

Q6: What happens if hyperscaler capex slows in 2025? Analyst consensus suggested a 20% slowdown in aggregate AI capex (from $150 billion to $120 billion) could reduce NVIDIA revenue growth to 30–40% in fiscal 2025–2026. This became a key risk monitored in real-time quarterly guidance.

Q7: Could custom silicon from hyperscalers (Google TPU, Amazon Trainium, Meta MTIA) displace NVIDIA GPUs? Partially. Analysis suggested hyperscalers might achieve 20–30% GPU TAM displacement by 2027. However, NVIDIA’s ecosystem (CUDA software, developer tools, backward compatibility) created switching costs that limited displacement. Coexistence was more likely than replacement.

Related Concepts

• Gross Margin Expansion: How product mix and supply dynamics drive profitability growth independent of revenue growth.
• Total Addressable Market (TAM) Expansion: The AI infrastructure TAM doubled from $100 billion (2023 estimate) to $200+ billion (2024 estimate), creating a growing market not just market share theft.
• Customer Concentration Risk: When 50%+ of revenue comes from a handful of customers, single-customer capex decisions can swing earnings materially.
• Earnings Power Index (EPI): NVIDIA’s fiscal 2024 EPI (earnings / price) reached 6–7%, suggesting undervaluation by historical dividend-discount models before the AI boom.
• Multiple Compression and Expansion: NVIDIA’s P/E multiple expanded from 30x (fiscal 2023) to 60x (fiscal Q4 2024), then compressed back to 35–40x (fiscal 2025) as growth decelerated.

Summary

NVIDIA’s fiscal 2024 earnings explosion—with revenue reaching $22.1 billion in a single quarter and EPS growing 1,709% YoY—exemplified how structural technology shifts can create unprecedented earnings acceleration. The company transformed from a cyclical semiconductor player into the primary beneficiary of the generative AI infrastructure buildout, sustaining 70%+ gross margins and 50%+ operating margins through peak demand cycles.

However, the earnings story also illustrated the limits of exponential growth. As fiscal 2025 guidance decelerated to 16% sequential growth from prior quarters’ 20–30% rates, investors confronted the reality that even transformative cycles follow S-shaped adoption curves: explosive early acceleration followed by normalization. The peak earnings multiple and stock price in February 2024 preceded the peak earnings growth by 6–9 months, a pattern classic to earnings-driven bull markets.

For investors, the NVIDIA case study reinforced that earnings acceleration—not absolute earnings level—drives stock returns. The company’s fiscal 2024 earnings growth of 6x year-over-year (the highest in semiconductor history) created returns far exceeding those justified by fundamental valuation alone. Identifying inflection points early, recognizing when acceleration rates peak, and rebalancing before deceleration remains the core challenge in earnings-driven investing.

Next

See JPMorgan’s Fortress Earnings