To The Moon

$Advanced Micro Devices(AMD)$ is gearing up for a major AI event on October 10th, and the focus should be on product development rather than just hunting for new financial metrics.

Analysts remain optimistic about chip stocks, believing AMD has significant long-term growth potential in AI.

1.Product Launch and Market Competition

This event marks AMD's first major product launch, but the key challenge lies in sales and functionality. Investors are aware of $NVIDIA Corp(NVDA)$ CEO Jensen Huang’s emphasis on the "incredible demand" for the upcoming Blackwell GPU, which raises the stakes for AMD.

The real question is whether AMD’s roadmap can keep pace with Nvidia's developments. Ahead of the event, AMD unveiled its AI GPU roadmap: the MI300 is available now, followed by the MI325X in Q4 2024, the MI350X in 2025, and the MI400 in 2026—AMD’s next-generation CDNA architecture.

2.Growth Expectations and Market Share

At Computex, AMD refreshed its AI GPU roadmap, making some bold claims. CEO Lisa Su believes the Instinct MI350 series will outperform the Blackwell Ultra version expected in 2025. The market will undoubtedly keep a close eye on product updates and comparisons, but favorable comparisons need to translate into stronger growth opportunities.

Analysts estimate AMD's AI revenue for the quarter will reach around $1 billion. While Nvidia consistently reports quarterly revenues surpassing $2 billion, AMD struggles to break that threshold. Estimates for AMD’s annual AI GPU sales show $5.1 billion in 2024, $9.7 billion in 2025, and $12.8 billion in 2026.

3.Challenges Ahead

It's remarkable for a chip company to more than double its sales in two years, yet AMD’s growth appears limited, and it risks losing market share. The company might not even capture 5% of the projected $400 billion AI market by 2027, falling short of the $20 billion sales mark.

Bank of America analyst Vivek Arya suggests that achieving a 10% market share by 2026 could mean an additional $5 billion in AI sales, potentially pushing AMD’s revenue to $17.8 billion by then—a substantial jump from 2024's forecast.

4.Revenue Expectations and Market Dynamics

AMD’s data center revenue saw steady growth but not explosive increases like Nvidia, which reported nearly $26.3 billion for the July quarter—a $16 billion year-over-year boost—while AMD's sales were only $2.8 billion, up $1.5 billion.

The second half of the year will reveal whether AMD can replicate Nvidia's substantial growth. To meet its $5.1 billion AI data center sales target for 2024, AMD needs to achieve $3.5 billion in Instinct GPU sales, expecting $1.5 billion in Q3 and $2 billion in Q4.

If successful, AMD could reach $8 billion in AI GPU sales for 2025, with market predictions suggesting they might slightly fall short of $10 billion next year. If AMD only needs a 10% market share to achieve earnings per share of $8 to $9 by 2026, the stock could be undervalued at 20 times that estimated EPS.

5.Risks and Considerations

The biggest risk lies in failing to meet the AI chip demand for a $400 billion market by 2027. There's also a secondary risk from new startups like Cerebras Systems, whose potential IPO could capture more market share.

Even if AMD meets its revenue targets through additional AI chip sales, it may still fall short of achieving a 60% gross margin compared to Nvidia. Rising foundry costs, including $Taiwan Semiconductor Manufacturing(TSM)$ 's N2 wafer costs hitting $30,000, could pressure profits for chip designers like AMD.

The October 10th AI event is crucial for investors to monitor AMD's product roadmap closely. If AMD can produce competitive AI GPUs, its revenue metrics could shift favorably over time, positioning it as a solid player in the AI chip market.