AMD: Server CPU Revenue Surpasses Intel for the First Time, TAM Doubles to $120 Billion, but the Real GPU Surge Awaits in H2

① Data Center Revenue of $5.8 Billion Surpasses Intel for the First Time: Year-over-year growth of 57% to $5.8 billion, with server CPU revenue increasing by over 50% year-over-year to hit a record high. During the same period, Intel's DCAI revenue was $5.1 billion, with a growth rate of 22%—AMD's revenue scale in the data center sector has surpassed Intel for the first time.

② Server CPU TAM Doubles from $60 Billion to $120 Billion: The Total Addressable Market (TAM) for 2030, estimated at approximately $60 billion during the analyst day last November, has doubled in just five months. The driver is Agentic AI significantly boosting CPU demand—the ratio of CPUs to GPUs is evolving from 1:4/1:8 towards 1:1 or even higher CPU proportions.

③ Meta 6GW + OpenAI Multi-Generation Contracts Lock in Long-Term Visibility: Meta plans to deploy up to 6 gigawatts of AMD Instinct GPUs, including custom accelerators based on the MI450 architecture; cooperation with OpenAI continues to advance. Management stated that data center AI revenue will reach the "tens of billions of dollars" level by 2027.

④ GPU Revenue Declines Quarter-over-Quarter in Q1 Due China Transition: Data center AI business saw a slight quarter-over-quarter decline due to a significant reduction in China-related revenue in Q1 compared to Q4. The real volume ramp-up for GPUs will begin in Q3 (initial Helios capacity), with significant scaling in Q4 continuing into Q1 2027.

⑤ Free Cash Flow Triples to a Record $2.6 Billion: FCF margin stands at 25%, with operating cash flow jumping from $940 million in the same period last year to $3 billion. However, the MI450 ramp-up in the second half will dilute gross margins—as this product's margin is below the company average.

⑥ H2 Consumer Sector Under Pressure: Memory Price Hikes Impact PC and Gaming: Management expects gaming revenue in the second half to drop by more than 20% compared to the first half, and PC shipments will also be affected by rising memory and component costs.

The most striking aspect of this earnings report is not the +38% revenue growth, but a structural crossover point: AMD's quarterly data center revenue of $5.8 billion exceeded Intel's DCAI revenue of $5.1 billion for the first time. From the debut of the Zen architecture in the server market in 2017 to today, it took AMD nearly 10 years to achieve this historic overtaking.

However, the other side of this earnings report is equally worth reading closely. The data center AI business (i.e., Instinct GPU) actually saw a slight quarter-over-quarter decline in Q1—due to substantial China-related revenue in Q4 that decreased significantly in Q1. This means AMD's "true explosion period" in the AI accelerator market has not yet arrived. Lisa Su clearly defined the mass production timeline for Helios (the full rack solution combining MI450 + Venice) as starting in Q3, with significant volume ramp-up in Q4—investors need to patiently wait for delivery proofs in the second half.

Meanwhile, the server CPU story is becoming increasingly compelling. The doubling of TAM from $60 billion to $120 billion is itself a figure worth pausing to digest. Management's logical chain is clear: The proliferation of Agentic AI → Explosion in inference volume → Each agent requires CPUs for orchestration and data processing → The CPU-to-GPU ratio increases from 1:4 to 1:1 or even higher → CPU TAM doubles. Q2 server CPU revenue guidance indicates year-over-year growth of over 70%, with the full-year growth trajectory still accelerating.

Below is a detailed analysis of the earnings content

AMD's first-quarter revenue was $10.3 billion, a 38% year-over-year increase, exceeding the upper end of guidance. Non-GAAP gross margin was 55%, up 1 percentage point year-over-year; operating income was $2.54 billion, up 43% year-over-year, with an operating margin of 25%. On a GAAP basis, operating income was $1.5 billion (+83% YoY), net income was $1.4 billion (+95% YoY), and EPS was $0.84 (+91% YoY). The main differences between GAAP and Non-GAAP figures stem from amortization of acquisition-related intangible assets ($551 million) and stock-based compensation expenses ($487 million).

Data Center: CPU Reigns Supreme, GPU Gathers Momentum

Data center revenue was $5.8 billion, up 57% year-over-year and 7% quarter-over-quarter, setting consecutive record highs. Operating income was $1.6 billion, with a margin of 28%, expanding by 3 percentage points year-over-year.

The server CPU business was the biggest highlight—setting revenue records for the fourth consecutive quarter, with year-over-year growth exceeding 50%. Both cloud and enterprise customers grew by over 50%, and the number of EPYC-driven cloud instances increased by nearly 50% year-over-year to over 1,600. The ramp-up of the 5th generation EPYC Turin and continued shipments of the 4th generation Genoa jointly drove growth. Lisa Su specifically pointed out that growth was primarily driven by shipment volumes rather than price hikes, with ASP improvements largely coming from product mix and increased core counts.

More importantly, the redefinition of TAM. The server CPU TAM for 2030 was estimated at approximately $60 billion (18% CAGR) during the analyst day last November, but just five months later, management doubled this figure to over $120 billion (>35% CAGR). The logic is that Agentic AI has reconstructed the CPU demand equation—previously, CPUs were merely "supporting actors" (head nodes) to GPUs, with ratios of 1:4 or 1:8; now, agentic workflows require substantial CPU power for orchestration, data processing, and parallel tasks, moving the ratio towards 1:1 or even higher.

Data Center AI (Instinct GPU) presents a state of "gathering momentum." Q1 saw a slight quarter-over-quarter decline, mainly because Q4 had significant China-related revenue which decreased in Q1. Year-over-year growth remained in the "significant double digits." The MI355X demonstrated comprehensive competitiveness in the latest MLPerf tests. The MI450 series GPUs have begun sampling to top-tier customers, with an H2 mass production timeline: initial capacity in Q3, significant ramp-up in Q4, and continued acceleration in Q1 2027. Management stated that demand forecasts from top-tier customers have exceeded initial plans, and the pipeline for large-scale deployments by new customers is also expanding, including multiple "multi-gigawatt" opportunities.

Horizontal comparisons reinforce AMD's momentum. Intel's DCAI revenue for the same period was $5.1 billion, with a growth rate of 22%—this marks the first time AMD's data center revenue has surpassed Intel, representing a milestone victory for AMD in the server market since the introduction of the Zen architecture in 2017. Intel's counterpoints lie in Xeon 6 being selected as the host CPU for NVIDIA DGX Rubin NVL8, and multi-year cooperation agreements with Google. However, in terms of growth rate and market share trends, AMD's advantage continues to expand, with management reiterating its goal of over 50% market share in server CPUs.

In the AI accelerator market, NVIDIA still dominates with approximately 75-80% share, but AMD is carving out a position as the second-largest player through strategic contracts with Meta (6GW) and OpenAI.

Client and Gaming: Commercial PC is a Highlight, H2 Faces Pressure

Client and Gaming revenue was $3.6 billion, up 23% year-over-year and down 9% quarter-over-quarter (seasonal). Operating income was $575 million, with a margin of 16%.

Client segment revenue was $2.9 billion, up 26% year-over-year, driven by strong Ryzen processor shipments and continuous market share growth. Commercial PCs were the core highlight of the quarter—Ryzen PRO PC sales increased by over 50% year-over-year, with Dell, HP, and Lenovo all expanding their AMD product lines. New enterprise customer wins cover financial, healthcare, industrial, and digital infrastructure sectors.

Gaming segment revenue was $720 million, up 11% year-over-year, primarily driven by demand for Radeon 9000 series GPUs, partially offset by a decline in semi-custom (console) revenue.

However, management issued clear warning signals for the second half: Due to rising memory and component costs, gaming revenue in the second half is expected to drop by over 20% compared to the first half, and PC shipments will also be impacted. The root of this dynamic is that AI's huge demand for HBM and DDR5 has pushed up prices for all memory types, indirectly squeezing consumer electronics. Despite this, management still expects the client business to grow year-over-year for the full year and outperform the market.

Embedded: Moderate Recovery

Embedded revenue was $873 million, up 6% year-over-year and down 8% quarter-over-quarter. Operating income was $338 million, with a margin of 39%. Demand improvement in sectors such as test and measurement, aerospace, and communications drove growth. Design wins increased by double digits year-over-year, valued in the billions, reflecting the embedded business's expansion from an FPGA-centric model to a broader portfolio of adaptive embedded x86 and semi-custom solutions.

Financial Quality: FCF Surge, but R&D Investment Accelerates

Free cash flow was $2.6 billion, more than triple that of the same period last year, with an FCF margin of 25%. Operating cash flow was $3 billion (vs. $940 million in the same period last year), with improvements stemming from higher net income and working capital efficiency.

Capital expenditures were $389 million (vs. $212 million in the same period last year), an 83% increase, but the absolute amount remains low—AMD's fabless model means capacity investments are primarily on the supply chain partner side. R&D expenses were $2.4 billion, up 39% year-over-year, accounting for 23% of revenue, reflecting accelerated investment in the AI roadmap.

Regarding the balance sheet, cash and short-term investments stood at $12.3 billion, inventory at $8 billion (basically flat), and long-term debt at $2.4 billion. In Q1, 1.1 million shares were repurchased ($221 million), leaving $9.2 billion in remaining repurchase authorization.

Notable one-time items: Q4 '25 included $280 million in long-term investment gains (which boosted Q4 EPS), while Q1 '26 had only $66 million, which is one reason Non-GAAP EPS dropped from $1.53 (Q4) to $1.37 (Q1). Additionally, the ZT Systems manufacturing business was divested in Q4, resulting in an $11 million gain from discontinued operations in Q1.

Management's Strategic Roadmap

Lisa Su outlined a clear growth path during the conference call:

Short-term (Q2 2026): Revenue guidance of $11.2 billion (+46% YoY, +9% QoQ), with gross margin improving to 56%. Server CPU revenue to grow by over 70% year-over-year, with both data center AI and server segments seeing double-digit quarter-over-quarter growth.

Medium-term (H2 2026-2027): Helios mass production is the core catalyst. MI450 series GPUs will start shipping in Q3, with significant volume ramp-up in Q4. Venice (6th generation EPYC) will be released within the year, with more customers in validation and ramp-up phases than any previous generation. Management expressed "strong and increasing confidence" that data center AI revenue will reach the "tens of billions of dollars" level by 2027.

Long-term: AI GPU CAGR target raised from >80% (will exceed previous targets). Long-term EPS target of $20+ (current annualized approx. $5.5, implying nearly 4x growth potential). Long-term gross margin target of 55%-58%.

Outlook: Two Main Lines Determine Medium-Term Direction

Execution Risk and Timeline for Helios Mass Production. The MI450 + Venice full rack solution is AMD's bet in the AI accelerator market, and the mass production progress in Q3-Q4 will directly determine the slope of 2027 revenue. Contracts with Meta and OpenAI provide demand anchors, but there are still execution risks between "sampling" and "scaled shipments." Key observation indicators: Quarter-over-quarter changes in DC AI revenue in Q3 (should rebound significantly), and whether there are announcements of new large-scale customers. If Helios shipments in Q3/Q4 meet expectations, AMD is poised to raise DC AI revenue from approximately $10 billion to the $20-30 billion range by 2027.

Where is the Ceiling for CPU Market Share? A 50% market share target under a $120 billion TAM implies $60 billion in annual revenue—this would make server CPU a larger business than the entire AMD of today. Intel is fighting back with Xeon 6 and Granite Rapids, and the Arm camp (AWS Graviton, Ampere, various in-house chips) is also rising. Lisa Su's response is: The market is large enough, different workloads require different CPUs, and AMD's "full spectrum" portfolio (general purpose, head node, agentic optimized) has a combination advantage over Arm's "single-point products." Whether this argument holds true, the progress of Venice customer validation will be the key test.

The impact of rising memory prices on the consumer business is a secondary line that requires continuous monitoring. The >20% drop in gaming revenue in the second half and pressure on PC shipments may suppress overall revenue growth, but given that data center accounts for over 56% of total revenue and is growing much faster than the consumer sector, the drag effect is structurally diminishing.