NVIDIA Corporation (NVDA), the AI darling, recently hit a rough patch. A report from The Information revealed that Nvidia’s highly anticipated Blackwell series chips are delayed due to design flaws, causing a sharp 15{3da602ca2e5ba97d747a870ebcce8c95d74f6ad8c291505a4dfd45401c18df38} drop in the stock over the past week. Even with this dip, the stock is still up more than 170{3da602ca2e5ba97d747a870ebcce8c95d74f6ad8c291505a4dfd45401c18df38} over the past year, but as we know, past performance isn’t a guarantee of future returns.

So, what’s going on with Nvidia? And more importantly, is it time to consider alternatives?

Dark Clouds Are Looming Over the Future of Nvidia

Back in March, NVDA announced its Blackwell series, boasting capabilities that promised to build and operate real-time generative AI on trillion-parameter large language models at a fraction of the cost and energy consumption of its predecessor. But fast forward a few months, and the picture isn’t as rosy.

According to the report, the company has informed major customers, including tech giants like Alphabet Inc. (GOOGL) and Microsoft Corporation (MSFT), that shipments of its Blackwell AI accelerator will be delayed by at least three months due to design flaws. It appears to involve Taiwan Semiconductor Manufacturing’s new packaging technology, which NVDA is one of the first to use, and issues with the placement of bridge dies connecting two GPUs.

This isn’t just a minor hiccup. The delay could throw off the plans of customers such as Microsoft and Meta Platforms, Inc. (META), who have invested billions in Nvidia’s new GPUs to drive their AI services. The worry is that these delays might prevent these companies from deploying large clusters of the new chips in their data centers by the first quarter of 2025, as they had hoped.

Design flaws aren’t something that can be fixed overnight, which explains the significant delay. Nvidia, for its part, hasn’t outright confirmed or denied the delays but did say that “production is on track to ramp later in 2024.” However, with only a few months left in the year, this sounds more like an early 2025 release.

The delay has led tech companies to look for alternatives from NVDA’s competitors, such as Advanced Micro Devices, Inc. (AMD). MSFT and GOOGL, for example, are already working on next-generation products with AMD.

While Nvidia still dominates the data center GPU market, the Blackwell delay could weigh on its stock price and reputation. It’s arguably the most significant setback NVDA has faced since the AI boom began, and it might just be the moment for AMD to shine.

The Future of Advanced Micro Devices

With a market cap of $3.18 trillion, NVDA’s growth prospects seem more limited compared to AMD, which could see its valuation double from its current $250 billion as it gains momentum in the data center space.

In the second quarter, AMD’s data center revenue surged 115{3da602ca2e5ba97d747a870ebcce8c95d74f6ad8c291505a4dfd45401c18df38} year-over-year to $2.83 billion, accounting for nearly half of its total revenue. The Mi300 series brought in over $1 billion in quarterly revenue for the first time, with its customer base expanding as Microsoft became the first cloud provider to offer general availability for the Instinct Mi300X.

The significant increase in AMD’s data center sales, driven by AI applications, is expected to boost profits further, as this segment typically yields higher margins. Additionally, the company’s recent acquisition of Silo AI, Europe’s largest private AI lab, will enhance its capabilities in generative AI, including inference, training, and large language models.

Furthermore, Advanced Micro Devices’ client revenue rose 49{3da602ca2e5ba97d747a870ebcce8c95d74f6ad8c291505a4dfd45401c18df38} year-over-year to $1.49 billion, though with slimmer margins than its data center business. The recent drop in the gaming and embedded segments will likely bottom out soon, potentially lifting overall results. Even modest gains could significantly boost AMD’s bottom line. The company reported net income of $265 million or $0.16 per share, up from $27 million or $0.20 per share recorded last year.

Investors are keen to see AMD challenge NVDA with its MI300X AI chip and demonstrate growth in its data center AI business. On the other hand, Street expects its revenue and EPS for the current year (ending December 2024) to increase 12.9{3da602ca2e5ba97d747a870ebcce8c95d74f6ad8c291505a4dfd45401c18df38} and 27.6{3da602ca2e5ba97d747a870ebcce8c95d74f6ad8c291505a4dfd45401c18df38} year-over-year to $25.62 billion and $3.38, respectively. If AMD can exceed expectations, the stock could experience significant gains in the coming months. Earlier this year, the company projected $4 billion in AI chip sales for 2024, representing about 15{3da602ca2e5ba97d747a870ebcce8c95d74f6ad8c291505a4dfd45401c18df38} of its expected revenue.

Is It Time to Ditch NVDA and Buy AMD?

Delays in Blackwell chip could impact NVDA’s market share and growth. If the delay is short, the stock might have minimal impact on its fiscal 2025 results. However, if it extends beyond three months, it could weigh heavily on the stock, especially as some analysts were anticipating a quicker resolution.

Additionally, concerns about whether the design flaw could lead to chip failures or affect production yields add to the uncertainty. Nvidia’s decision to pause production and address the issue is a smart move, but it highlights the risks of its aggressive development timeline, which has been shortened from two years to one. While this strategy could pay off, it also increases the risk of errors or delays.

On the other hand, AMD is well-positioned to benefit from NVDA’s ongoing headwinds. With its MI300X AI chip gaining traction and strong data center growth, Advanced Micro Devices could capture some market share from Nvidia. Given this backdrop, it might be the right time to consider rotating out of NVDA and into AMD, especially for investors looking to capitalize on the AI-driven growth in the semiconductor sector.



Source link

By admin