AMD (NASDAQ:AMD) is on the brink of a major AI-driven expansion, projecting double-digit growth in 2025 as it ramps up data center dominance and AI accelerator adoption.
AMD's Record-Breaking AI Surge: Data Center Dominance & the Road to MI350AMD delivered a record-breaking 2024 with $25.8B in revenues, a 14% year-over-year growth. The Data Center segment took the spotlight with nearly doubling at $12.6 billion in annual revenues, as a testament to accelerating cloud adoption, AI accelerator growth, and EPYC CPU market-share expansion.
In the meanwhile, the Client segment grew 52% year-over-year at $7.1 billion with the support of AI-driven Ryzen processors as well as the recovery of PCs. Meanwhile, the Gaming as well as Embedded segments took a hit with gaming revenues falling 58% year-over-year at $2.6 billion as a result of cyclical declines in the sale of consoles, as well as Embedded revenues falling 33% at $3.6 billion largely because of normalization of inventories in the industrials. Amidst these headwinds, the rising margins at AMD as well as disciplined capital allocation saw the operating efficiency continue.
In Q4 2024, revenues at AMD amounted to $7.7B, 24% year-over-year growth, 12% sequential growth. $3.9 billion in revenues came from the Data Center segment, 69% year-over-year growth, reinforcing the dominance of AMD in the realm of AI as well as cloud computing. MI300X was installed at Meta (NASDAQ:META) for inference workloads successfully, a breakthrough in the AI GPU roadmap of AMD, as ROCm adoption sped up in the hyperscalers, reinforcing the position of the AMD GPUs as a viable alternative to the CUDA ecosystem of NVIDIA (NASDAQ:NVDA).
Client delivered strong results at $2.3 billion, indicating the adoption of AI PCs as well as continuing market-share growth in the CPU market, while revenues from the Gaming sector down 59% YoY, as well as Embedded revenues down 13% YoY due to weak industrial demand. Despite these results, the stock market reacted cautiously when the $7.1 billion revenues (+30% YoY but -7% sequentially) guidance for Q1 2025 from AMD arrived.
Analysts expressed worry regarding the lumpy second-half revenues from the data center, but investors took reassurance from the second-half ramp-up in AI accelerators, notably with the introduction of the MI350 later in the year. CEO Lisa Su cleverly postponed the introduction of the MI350 from the end of 2025 to the middle of the year, setting the stage for a fierce growth in the AI accelerator market, targeted at $120 billion in 2030.
AMD's AI Expansion: The Road to Billions & The Battle for Market ShareLooking ahead, AMD is projecting double-digit growth in 2025 revenues as well as EPS, with AI being the growth driver. Its long-term data center AI roadmap is impressive, with Lisa Su indicating the company is poised to scale AI accelerator revenues from $5 billion in 2024 levels to tens of billions per annum in the next few years.
Growing AI GPU market share is a priority area, with the aim of taking down NVIDIA's market dominance in the AI GPU market through competitive AI training capabilities, ROCm adoption growth, as well as prices. In enterprise AI workloads as well as inference computing beyond cloud hyperscalers, the sustainability of the market penetration is going to be dictated.
AI-powered Ryzen CPUs also power the AI PC revolution, cementing the diversity of revenues at AMD. With investments in AI infrastructure, core cloud partnerships, as well as a maturing software ecosystem, the long-term upside is set at AMD even as there is near-term uncertainty. Growth in the MI350 launch, ROCm advancement, as well as the growth in AI chip demand is going to be the drivers shaping the long-term trajectory at AMD.
On the margin side, the 54% non-GAAP margin at AMD is a considerable expansion from the 30%-plus margins of the early 2010s, indicating improved pricing power as well as a more valuable mix of data center. Finally, the $2.4 billion free cash flow let the company continue its aggressive R&D spend as well as engage in repurchases. But the $1 billion repurchases at AMD fall considerably short of the $15 billion repurchase program at NVIDIA, indicating a more conservative capital allocation approach.
Valuation Analysis: An In-Depth Look at Its Position in the Market and Future OutlookAMD is currently trading at $110.84, with a market capitalization of $179.61 billion. Despite its 3,463.99% surge over the past decade, the stock remains 47.56% below its all-time high, signaling a potential mispricing of its AI-driven growth.
One of the most telling valuation indicators is the forward price-earnings (P/E) ratio of 23.92, significantly lower than its trailing P/E of 111.96. This sharp compression suggests that Wall Street is still underestimating AMD's future earnings potential, despite its accelerating AI expansion. AMD's price-to-sales (P/S) ratio of 8.65 also lags behind NVIDIA's premium multiple, reflecting an opportunity for a market re-rating as AI revenue scales.
From a cash flow perspective, AMD is demonstrating improving efficiency. The company generated $2.4 billion in free cash flow (FCF) in 2024, marking a 35% YoY increase. However, its EV/FCF ratio of 73.47 suggests that investors have yet to fully price in the AI-led margin expansion that will come with MI350's rollout and data center acceleration.
Lastly, the GF Value estimate of $129.61 implies that AMD is currently ~17% undervalued. This model considers earnings growth, profitability, and industry trends, reinforcing the notion that AI tailwinds and expanding hyperscaler partnerships are not yet fully priced into AMD's stock price.
TakeawayAMD's AI-driven growth is accelerating, yet the market hasn't fully priced in its long-term potential. With Meta deploying MI300X for inference, ROCm adoption expanding, and the MI350 launch pulled forward, AMD is rapidly closing the gap with NVIDIA. Despite short-term concerns over cyclical headwinds, its AI accelerator revenue is set to scale from $5 billion to tens of billions, reinforcing its position as a long-term AI powerhouse.
This content was originally published on Gurufocus.com
