Advanced Micro Devices

AMD reported second-quarter revenue of $11.5 billion, up 50% year over year and ahead of guidance. AMD expects September-quarter revenue of $13.0 billion, representing 41% year-over-year growth and ahead of FactSet consensus estimates of $12.5 billion.

Why it matters: We’re pleased with AMD’s results and forecasts as the firm is positioned to deliver its first artificial intelligence rack-scale solution, Helios, to customers in the fourth quarter. AMD stated that its early estimates for Helios unit shipments are ahead of the company’s initial expectations. Data center revenue was $6.7 billion, up 107% year over year. We're encouraged by such strong growth. We estimate AMD’s server CPU revenue grew about 75% to $4.0 billion while its AI GPU revenue came in at $2.7 billion, flattish sequentially but at a low point ahead of the Helios ramp.

The bottom line: We maintain our $530 fair value estimate for narrow-moat AMD. Shares fell about 7% after hours, as investors were looking for even stronger results, but we view it as a modest buying opportunity. We recently raised our fair value estimate to $530 from $450, driven by AMD’s forecast for a much larger server CPU market than estimated just a few months ago. AMD’s results support this higher forecast with strong upcoming growth and tight supply as AMD sells everything it can make.

Coming up: In 2027, AMD expects its server CPU business to grow 70%-plus and its total data center segment to grow over 100%. AMD commented that its AI GPU business should grow “well over 100%,” and we estimate 150% growth to $36 billion, driven by the Helios ramp and key partnerships. We think these stellar growth rates were largely baked into expectations, which might explain the stock selloff. However, we also suspect these growth rates may prove conservative over the next four to six quarters as Helios and future AI racks come to market.

AMD expects its server CPU business to grow by 80%-plus in the second half of 2026 versus the second half of 2025, driven by the rapid rise of agentic AI. We think this forecast implies modest sequential growth off AMD’s strong second-quarter revenue of $4.0 billion, rising to $4.4 billion and $4.5 billion in the third and fourth quarters, respectively.

AMD forecasts adjusted gross margin of 56% in the third quarter, flattish sequentially. We model flattish gross margins for each quarter through the end of 2027, recognizing that the Helios launch should be modestly dilutive on a gross margin basis, offsetting any gross margin improvements from other products, such as embedded chips or server CPUs.

In total, AMD indicated that its revenue growth should exceed its prior target (outlined in November 2025) of a 35% CAGR over the next three to five years. Similarly, AMD believes its prior target of $20 of adjusted earnings per share by 2030 is now conservative. We agree and model a 39% revenue CAGR for AMD and think the firm can get close to $20 of EPS as soon as 2028.

Our fair value estimate is $530 per share, which implies a 2026 adjusted price/earnings ratio of 80 times but a 2028 and 2030 adjusted price/earnings ratio of 27 times and 16 times, respectively.

We expect AMD to achieve a top-line compound annual growth rate of 40% from 2026 to 2030. We attribute our optimism to the rise of agentic AI, which should require exponentially more server CPUs for agentic orchestration workloads.

We model 37% growth in 2026, thanks to the rise of agentic AI driving tremendous demand for server CPUs. We also think AMD will increase its AI GPU business nicely toward the end of 2026 when it brings its Helios rack-scale solutions to market.

We model growth of 72% and 37.5% in 2027 and 2028, respectively, as partnerships with OpenAI, Meta, and Anthropic ramp up and AMD starts to earn tens of billions of dollars in revenue from AI GPUs. Agentic AI should again support strong server CPU growth too. We model longer-term growth rates of 31% in 2029 and 26% in 2030. We bake in a pause in demand at some point and assume only 3% growth in 2031, but nearly 10% of revenue growth per year thereafter as AMD cements its place as a prominent AI vendor.

We are most bullish on the data center segment, in which AMD forecasts a total addressable market of $1.4 trillion by 2030. This TAM includes GPUs, CPUs, high-bandwidth memory, and AI networking chips.

For AMD’s data center segment, which includes both x86 server CPUs and AI GPUs, we model a 58% revenue CAGR over the next five years. Within this segment, we think AMD can carve out a decent portion of the AI accelerator market over time.

We estimate AMD’s AI GPU revenue reached $7.4 billion in 2025 versus nearly zero in 2023 and prior.

We model AMD reaching $14.5 billion of AI GPU revenue in 2026, boosted by the ramp of its MI450 rack-scale products into OpenAI and Meta in the fourth quarter of 2026. These two partnerships lead us to double our AI GPU revenue estimates thereafter.

We model 144% growth to $35 billion in 2027, 30% growth to $46 billion in 2028, 25% growth to $57 billion in 2029, and 20% growth to $69 billion in 2030. This equates to an 83% CAGR from 2026 to 2030, in line with AMD’s prior forecast of an 80% CAGR.

In client revenue (PC CPUs), we model a 6% CAGR over the next five years as AMD gains market share from Intel, although high memory chip prices may stunt PC production and demand for AMD’s PC CPUs. We model an 11% CAGR for AMD’s embedded business, albeit off of a lower-than-usual base after a couple of years of revenue declines.

Adjusted gross margin expanded from 45% in 2020 to 52% in 2025. We model an increase to 56% in 2026 and 57% in 2030, within management’s long-term target of 55%-58. Similarly, adjusted operating margin was 22% in 2025, but data center growth should lead to operating margin expansion to just under 40% in 2030.

We assign AMD a Narrow Morningstar Economic Moat Rating based on intangible assets around a variety of chip designs. We think it is more likely than not that AMD will generate excess returns on capital over the next 10 years, and perhaps even longer.

AMD is perhaps best known for its CPUs for PC desktops and notebooks within its client business segment. We think that AMD has a moat in this business, not only because of chip design expertise honed over decades, but also because it is one of two prominent firms to hold an x86 instruction set architecture license. AMD is also emerging as a prominent GPU vendor in artificial intelligence and holds valuable expertise in hardware, software, and rack-scale solutions, as well.

In PCs, Intel and AMD are the primary licensees of x86 and owners of x86 intellectual property, as they have a cross-licensing agreement that covers the x86 patent portfolio and would be terminated if either firm were acquired or had a change of control. The exclusivity of the x86 license serves as a significant barrier to entry for other chipmakers that cannot make an x86 processor without a license.

For many years, effectively all PC software was designed for the x86 architecture. Today, however, Apple has already done the heavy lifting of converting its Mac software away from x86 and toward ARM’s instruction set. Microsoft is now supporting Windows on ARM-based devices and may increase these investments over time. Thus, we don’t think the x86 architecture is as moaty as it once was, although we still foresee x86-based processors from Intel and AMD as making up a significant portion of the PC market for the next several years.

Given Intel’s manufacturing stumbles, AMD has gained market share from Intel in PC CPUs in recent years. If Intel recovers, AMD’s recent market share gains may flatline, but certainly, we no longer view AMD as an also-ran in CPUs.

In server CPUs within AMD’s data center segment, we see similar dynamics where most of the software (particularly in the enterprise) is based on x86. Intel’s stumbles again enabled AMD to gain market share in x86. However, we see leading cloud vendors support ARM-based server CPUs, if not develop their own, such as Amazon’s Graviton. We anticipate strong growth in server CPUs in the years ahead, thanks to agentic AI, but we also foresee more ARM-based competition from Nvidia, ARM, Qualcomm and others that might weigh on x86 expansion.

In AI, we think AMD’s GPU expertise is becoming increasingly valuable in AI applications. Nvidia has a clear lead in AI GPUs, thanks to its superior GPU designs, incumbency in software (CUDA), and strong networking and interconnectivity products. Still, GPUs offer flexibility, and we see leading tech companies supporting AMD to emerge as a second GPU alternative for programmable, flexible AI workloads.

AI offers a massive opportunity to GPU makers, and while AMD lags industry leader Nvidia, we see plenty of room in the AI market for AMD.

AMD has gained share in the PC CPU market as Intel’s manufacturing prowess has hit several bumps in recent years.

AMD’s partnership with chip manufacturing leader TSMC, plus its adoption of a chiplet manufacturing strategy, has allowed the company to come to market with more formidable products and greater flexibility to bring new products to market quickly.

AMD will need to improve its software capabilities to make a dent in Nvidia’s AI dominance, as Nvidia is strong in not only GPUs but also associated AI software tools.

Despite AMD’s recent share gains, Intel remains the industry leader in PC CPUs and might recapture most of the market if it can deliver industry-leading manufacturing capabilities once again.

AMD’s gaming semicustom chip business is beholden to the design cycles and launches of new gaming consoles, and it might be a couple of more years until next-generation consoles arrive.