AMD Q1 2026 Earnings Preview: MI300X Gaining Ground

Overview

Advanced Micro Devices (AMD) is scheduled to report its Q1 2026 financial results on May 6, 2026, with Wall Street analysts expecting earnings per share (EPS) of $0.94 on revenue of approximately $7.11 billion — representing roughly 30% year-over-year growth (FactSet consensus, as of April 17, 2026). The focal point of investor attention remains the MI300X AI accelerator, which AMD management indicated surpassed $5 billion in annualized revenue run-rate by late 2025, and which analysts believe is gaining meaningful share against NVIDIA's dominant H100 and H200 product lines. This preview examines what the numbers signal, why trajectory matters more than a single quarter's beat, and where the key risks lie heading into the print.

Sources: FactSet consensus estimates (April 17, 2026), AMD Investor Relations (Q4 2025 Earnings Call, January 28, 2026)

---

Key Metrics (as of April 17, 2026)

Metric	Consensus Estimate (Q1 2026)	Q1 2025 Actual	YoY Change
Revenue	$7.11B	$5.47B	+30.0%
EPS (Non-GAAP)	$0.94	$0.62	+51.6%
Data Center Revenue	$3.80B	$2.34B	+62.4%
Client (PC) Revenue	$1.55B	$1.37B	+13.1%
Gaming Revenue	$0.48B	$0.922B	-47.9%
Embedded Revenue	$0.58B	$0.46B	+26.1%
Gross Margin (Non-GAAP)	54.2%	52.0%	+220 bps
Operating Margin (Non-GAAP)	25.5%	22.5%	+300 bps

Source: FactSet consensus estimates as of April 17, 2026; Q1 2025 actuals per AMD Q1 2025 earnings release (SEC Form 8-K, April 30, 2025)

---

Data Center: The MI300X Inflection Story

The Data Center segment is the single most important driver of AMD's re-rating thesis heading into Q1 2026, and the consensus estimate of $3.80 billion for the segment alone illustrates how dramatically the business has shifted. For context, AMD's entire company revenue in Q1 2024 was $5.47 billion — meaning Data Center alone is now expected to approach that figure in a single quarter.

The MI300X GPU, AMD's flagship AI accelerator built on a chiplet-based architecture combining CDNA 3 compute dies with HBM3 memory, has positioned AMD as the credible alternative to NVIDIA in large language model (LLM) inference workloads. AMD CEO Dr. Lisa Su emphasized on the Q4 2025 earnings call (January 28, 2026) that hyperscaler customers including Microsoft Azure and Meta are deploying MI300X at scale, and that the successor MI350X is on track for volume production in the second half of 2026.

Why does this matter beyond the headline number? The gross margin expansion implied in the consensus — 54.2% non-GAAP, up 220 basis points year-over-year — suggests AMD is not simply winning share by discounting. Data Center GPUs carry structurally higher margins than PC client CPUs, and a revenue mix shift toward Data Center lifts blended margins organically. If AMD delivers at or above the $3.80 billion Data Center consensus in Q1 2026, it would signal that MI300X demand is durable rather than a one-time pull-forward from hyperscalers diversifying supply chains.

EPYC server CPUs also contribute meaningfully to Data Center revenue. AMD's EPYC Genoa (4th Gen) and the newer Turin (5th Gen) continue to take share from Intel Xeon, with AMD now reportedly holding approximately 24% of the x86 server CPU market as of Q4 2025, according to Mercury Research data cited by Bloomberg (February 2026). This dual engine — CPUs and GPUs — within Data Center provides revenue diversification that pure-play GPU competitors lack.

---

Forward Outlook: MI350X, ROCm Maturity, and Guidance Watch

Beyond Q1 2026 results, the forward guidance AMD provides for Q2 2026 will likely carry more weight with investors than the quarterly beat or miss itself. Analysts at Morgan Stanley (as of March 2026) project AMD full-year 2026 Data Center revenue could reach $17–$19 billion if MI350X ramps on schedule, implying sequential quarterly growth must accelerate through the second half of the year.

A critical near-term catalyst is the maturity of AMD's ROCm software ecosystem. Historically, NVIDIA's CUDA platform has maintained a significant software moat — developers build models optimized for CUDA, creating switching costs. AMD has invested substantially in ROCm compatibility, and management indicated on the Q4 2025 call that ROCm 7.0 (released February 2026) achieves parity with CUDA on the majority of PyTorch and TensorFlow workloads. Independent benchmarks from MLCommons' MLPerf v4.0 (November 2025) showed MI300X achieving competitive performance with H100 on inference tasks, a notable improvement from prior test cycles.

For Q2 2026 guidance, the Street consensus anticipates revenue of approximately $7.75 billion. A guide in line with or above this level would reinforce the bull case, while guidance below $7.4 billion could trigger a meaningful re-rating lower given AMD's current valuation — the stock trades at approximately 35x forward 12-month non-GAAP EPS as of April 17, 2026, according to FactSet, a premium that demands consistent execution.

The Client and Gaming segments are worth monitoring as well. PC CPU demand has been stable, supported by the AI PC upgrade cycle in which AMD's Ryzen AI 300 series chips are competing directly with Intel Lunar Lake and Qualcomm Snapdragon X. Gaming revenue, however, continues to decline as the current console cycle (PlayStation 5 and Xbox Series X/S) ages, and analysts expect this headwind to persist through at least mid-2026. Embedded recovery — tied to industrial and automotive customers de-stocking — appears gradual but improving, with sequential growth expected for the third consecutive quarter.

---

Risk Factors

• Export Control Escalation: The U.S. Department of Commerce has progressively tightened restrictions on AI chip exports to China and other restricted entities. AMD's MI300X sales to Chinese hyperscalers and cloud providers are already constrained under existing rules, and any further tightening before or after the Q1 report could reduce the total addressable market estimate for AMD's Data Center GPU business, potentially forcing downward revisions to full-year 2026 guidance.

• NVIDIA Competitive Response: NVIDIA's Blackwell (B100/B200) GPU platform began volume shipments in late 2025 and represents a significant performance leap, particularly for training workloads. If Blackwell supply normalizes faster than expected through 2026, hyperscalers may reduce MI300X orders in favor of NVIDIA's flagship products, pressuring AMD's Data Center growth trajectory and the $17–$19 billion full-year revenue thesis.

• Execution Risk on MI350X Ramp: AMD's forward multiple is partially pricing in a successful MI350X launch in H2 2026. Any manufacturing delays at TSMC (the contract foundry producing AMD's chips on N3 process node) or yield challenges with the advanced chiplet packaging could delay revenue recognition and create a guidance miss in Q3 or Q4 2026, exposing the stock to multiple compression from its current elevated valuation.

---

Investment Outlook

AMD enters its Q1 2026 earnings report with credible momentum in its most important segment: Data Center GPU and CPU combined suggest a business in genuine structural growth, not cyclical recovery. The MI300X has demonstrated that AMD can win at scale with hyperscalers, and the ROCm software improvements address what had historically been the primary competitive disadvantage.

That said, at approximately 35x forward non-GAAP EPS as of April 17, 2026, AMD is priced for continued outperformance. The risk-reward is constructive for long-term investors who believe MI350X ramps successfully and the ROCm ecosystem continues closing the gap with CUDA — but the margin for disappointment on guidance is narrow. Analysts broadly maintain a positive bias: Bloomberg consensus as of April 17, 2026 shows 35 Buy ratings, 9 Hold ratings, and 2 Sell ratings on AMD, with a median 12-month price target of $148. Investors should size positions accordingly and monitor the Q2 2026 guidance closely on May 6.

Disclaimer: This content is for informational purposes only and was produced with AI assistance. It does not constitute financial advice. All investment decisions carry risk and are solely your own responsibility. Past performance is not indicative of future results.