AMD is experiencing a dramatic shift in its business model as GPU sales for AI infrastructure approach the scale of its traditional CPU business, CEO Lisa Su announced during the company’s third-quarter earnings call. The chipmaker has revised its 2024 data center GPU sales forecast upward to $5 billion or more, a significant increase from the $4.5 billion projected in July and more than double the $2 billion expected at the start of the year.
The company generated over $1.5 billion in GPU revenue during Q3 alone, reflecting the explosive demand for AI computing infrastructure. This growing parity between CPU and GPU sales represents a fundamental shift in the computing landscape, though analysts caution that the picture is nuanced. The balance may reflect both the rising importance of GPUs for AI workloads and potential weakness in the traditional CPU market, with multiple Wall Street analysts expecting data center CPU spending to cool in Q1.
Bank of America estimates AMD holds approximately 5% of the AI chip market, with demand projected to grow 50-100% next year. This positions AMD for substantial growth even without significantly challenging Nvidia’s dominant market position. However, Morgan Stanley analysts suggest AMD’s current revenue primarily comes from inference applications, which would put their inference market share above 10% - a notably high figure given their relatively small customer base.
A critical challenge for AMD remains the immaturity of its GPU ecosystem. While Nvidia has spent 18 years building its software stack on the CUDA framework, AMD is only one year into its AI journey. Developing robust software tools and developer communities takes considerable time, which may limit AMD’s ability to capture market share quickly.
Analysts predict 2025 will be another “investment year” for AI data centers, with UBS forecasting higher GPU sales in 2026 than 2025. Morgan Stanley projects AMD’s GPU sales will reach approximately $8 billion next year, up from $5 billion in 2024. Despite the positive growth trajectory, AMD’s share price dropped more than 10% following the earnings announcement, suggesting investors had even higher expectations for the company’s AI infrastructure growth.
Key Quotes
We’re actually seeing now our GPU business really approaching the scale of our CPU business
AMD CEO Lisa Su made this statement during the company’s Q3 earnings call, highlighting the dramatic shift in AMD’s business composition as AI infrastructure demands reshape the computing landscape.
We think that AMD’s current revenue is mostly coming from inference applications, which would put their inference market share at above 10%, which seems a high bar, given a relatively small customer set
Morgan Stanley analysts provided this assessment after AMD’s earnings call, suggesting the company has achieved significant market penetration in AI inference workloads despite having fewer customers than competitors.
2025 may be yet another transition year in which AI data centers and GPU sales don’t reach maturity
UBS analysts offered this perspective, indicating that the AI infrastructure buildout will continue to evolve over multiple years, with 2026 potentially seeing higher GPU sales than 2025 as the ecosystem matures.
Our Take
AMD’s GPU growth story reveals both the enormous opportunity and significant challenges in the AI chip market. While the company is experiencing explosive growth - more than doubling initial forecasts - the 10% stock drop demonstrates how difficult it is to meet inflated investor expectations in the AI sector. The real competitive moat isn’t just hardware performance but software ecosystems, where Nvidia’s 18-year head start with CUDA presents a formidable barrier. AMD’s success in capturing even 5% of the market is notable, but the path to significantly higher market share will require sustained investment in developer tools and community building. The prediction that 2025 will be another “investment year” suggests the AI infrastructure market is still in early innings, with companies prioritizing capability building over profitability. This creates opportunities for AMD to close the gap, but also means continued volatility as the market sorts out long-term winners.
Why This Matters
This development signals a fundamental transformation in computing architecture driven by artificial intelligence demands. The fact that GPU sales are approaching CPU sales at a major chipmaker like AMD demonstrates how AI workloads are reshaping the entire semiconductor industry and data center infrastructure investments.
The story matters because it highlights the massive capital flowing into AI infrastructure and the competitive dynamics between AMD and Nvidia in the lucrative AI chip market. With demand projected to grow 50-100% annually, companies building AI capabilities face critical decisions about which hardware platforms to adopt.
For businesses investing in AI, AMD’s progress represents an important alternative to Nvidia’s dominant position, potentially offering more competitive pricing and supply availability. However, the immaturity of AMD’s software ecosystem underscores the importance of considering not just hardware performance but the entire development environment when making infrastructure decisions. The market’s reaction - a 10% stock drop despite strong growth - also reveals how extraordinarily high investor expectations have become for AI-related companies, creating volatility even when fundamentals appear strong.
Recommended Reading
For those interested in learning more about artificial intelligence, machine learning, and effective AI communication, here are some excellent resources:
Recommended Reading
Related Stories
- Jensen Huang: TSMC Helped Fix Design Flaw with Nvidia’s Blackwell AI Chip
- Apple Q4 Earnings Preview: Wall Street Sees AI Fueling iPhone Demand in 2024
- Wall Street Asks Big Tech: Will AI Ever Make Money?
- EnCharge AI Secures $100M Series B to Revolutionize Energy-Efficient AI Chips
- Pitch Deck: TensorWave raises $10M to build safer AI compute chips for Nvidia and AMD
Source: https://www.businessinsider.com/amd-gpu-cpu-ai-data-center-boom-2024-10