Missed Nvidia in 2009? This Rare Signal Is Flashing Again.\u00a0In 2009, a “Double Down” signal flashed for a little-known chipmaker called Nvidia.\u00a0For the first time in years, that same “Total Conviction” signal is flashing for a company 1\/100th the size of Nvidia.\u00a0Continue \u00bb<\/a> <\/p>\n AMD is set to ride two of the biggest waves in AI right now in inference<\/a> and agentic AI. While it was never able to challenge Nvidia in AI model training, the company is better positioned for inference, which isn’t as technically complex and tends to be much more about quick access to memory than about raw computing power. On this end, the company’s chiplet design allows its chips to be packaged with more memory, reducing latency. <\/p>\n The company has also made two major acquisitions in the memory space. Memory optimization company MEXT provides an AI solution that can direct less frequently accessed data from DRAM (dynamic random access memory) to flash memory and then retrieve it before it’s needed, helping reduce costs without sacrificing performance. <\/p>\n Its acquisition of Taalas, meanwhile, gives it inference chips hardwired with specific AI models that can significantly reduce latency. It’s also formed a partnership with Cerebras, in which its Helios system will be used in the pre-fill phase of inference, and Cerebras’ high-end solution will be used for the decode phase. <\/p>\n AMD already has two large $100 billion inference deals in place with Meta Platforms and OpenAI, while Anthropic and Microsoft have also recently become customers. This should be a huge area of growth for the company over the next five years. <\/p>\n AMD also sees a $220 billion opportunity in server central processing units (CPUs). The company is a leader in this rapidly growing market, where demand has surged due to the need for more CPUs for agentic AI. <\/p>\n Marvell: An optics leader <\/p>\n Like AMD, Marvell is riding two powerful trends: optical networking and custom chips. Both businesses are growing quickly. <\/p>\n The company’s strongest segment has been optical interconnects. AI data centers are beginning to move away from copper wiring to optical components, which generate less heat and offer higher speeds. Marvell, meanwhile, is the market leader in optical DSPs (digital signal processors), the primary chips that convert electrical signals from AI chip clusters into light, enabling data to be transmitted through optical systems. This business is booming, with the company expecting its interconnect revenue to climb 70% this fiscal year. <\/p>\n Story Continues <\/p>\n At the same time, Marvell is seeing strong growth in its custom chip business. It provides some intellectual property (IP) for Amazon’s Trainium chips, which have been growing rapidly. The company also helped Microsoft develop its new Maia chip, and it will provide Alphabet with several components for its TPU architecture, including network interface controllers (NICs), AI inference accelerators, storage controllers, and near-memory compute. In addition, it has strong IP around SRAM (static random-access memory), which is used in Nvidia’s LPUs (language processing units) and Cerebras’ chips. <\/p>\n On the downside, Marvell may play a less prominent role in future iterations of Trainium, which could affect its growth. The stock is also expensive, trading at a forward P\/E of 37 times fiscal 2028 (ending January 2028) analyst estimates. <\/p>\n The verdict <\/p>\n While both stocks look like potential long-term winners riding great trends, I think AMD should outperform over the next five years. It is slightly cheaper, trading at 30 times 2027 EPS, and doesn’t face the potential Trainium headwind that Marvell does. <\/p>\n Should you buy stock in Advanced Micro Devices right now? <\/p>\n Before you buy stock in Advanced Micro Devices, consider this: <\/p>\n The Motley Fool Stock Advisor analyst team just identified what they believe are the\u00a010 best stocks<\/a> for investors to buy now\u2026 and Advanced Micro Devices wasn’t one of them. The 10 stocks that made the cut could produce monster returns in the coming years. <\/p>\n Consider when Netflix made this list on December 17, 2004… if you invested $1,000 at the time of our recommendation,\u00a0you’d have $443,461!* Or when Nvidia made this list on April 15, 2005… if you invested $1,000 at the time of our recommendation, you’d have $1,307,633!* <\/p>\n Now, it’s worth noting\u00a0Stock Advisor’s total average return is 973% \u2014 a market-crushing outperformance compared to 213% for the S&P 500.\u00a0Don’t miss the latest top 10 list, available with\u00a0Stock Advisor, and join an investing community built by individual investors for individual investors. <\/p>\n See the 10 stocks \u00bb<\/a> <\/p>\n *Stock Advisor returns as of August 26, 2026. <\/p>\n Geoffrey Seiler<\/a> has positions in Advanced Micro Devices, Alphabet, Amazon, and Meta Platforms. The Motley Fool has positions in and recommends Advanced Micro Devices, Alphabet, Amazon, Marvell Technology, Meta Platforms, Microsoft, and Nvidia. The Motley Fool has a disclosure policy<\/a>. <\/p>\n<\/a> Image source: The Motley Fool. AMD: An inference and agentic AI winner <\/p>\n