Data Center Accelerates to +57%, Q2 Guides to $11.2B (+$680M Above Street), 56% GM Guide — And the Rally Stops Out: Maintaining Outperform, Fair Value Raised to $325-400
Key Takeaways
- Q1 2026 was a clean operating beat: revenue $10.253B (+37% YoY) beat the $9.99B Street consensus by ~$260M (+2.6%); non-GAAP EPS of $1.37 beat $1.25 consensus by 9.6%; sequential -0.2% QoQ from Q4 25 $10.27B is in line with normal Q1 seasonality (and notably the trough we'd expect with no MI308 China weakness this time). Despite the beats, this is the second-cleanest quarter AMD has delivered in our coverage cycle — the Q4 2025 print remains the cleanest by margin.
- Data Center revenue printed $5.8B, up +57% YoY — a significant acceleration from Q4 25's +39% and the strongest YoY DC growth print AMD has delivered post-MI300 era. Beat $5.6B Street estimates by $200M (+3.6%). The +57% trajectory implies FY26 Data Center growth is materially ahead of the "could exceed 60%" framework management offered at Q4 25 — which we now characterize as conservative rather than ambitious. EPYC + MI350 ramp continues to deliver above-plan; ROCm 7 production deployment scales; and the OpenAI 1GW MI450 deployment is on track for 2H 2026.
- Q2 2026 guidance is the standout positive: revenue ~$11.2B ±$300M (above $10.52B Street consensus by ~$680M / +6.5%); midpoint implies +46% YoY and +9% QoQ sequential growth; non-GAAP gross margin guide of ~56% (vs ~55.3% Street) signals continued mix improvement as Data Center scales. The Q2 sequential growth profile breaks the prior "Q1 seasonal trough then back-half loaded" pattern and indicates the Data Center ramp is now structural enough to overcome traditional consumer-segment Q1 seasonality.
- And the stock faded anyway. Initial AH reaction +6% on the headline beat reversed to ~-5%+ as the call digested, mirroring AMD's historical post-print average reaction of -5.15% (per 24/7 Wall St analysis). The setup explains it: AMD entered the print at $341 (+59% YTD, +255% 1Y), having rallied roughly 70% from the Q4 2025 post-print low of $200. Even a 6.5% guide-above-Street upgrade can't reset positioning that aggressive. This is a positioning unwind, not a fundamentals reaction — identical pattern to Q3 25 (-9% on a beat) and Q4 25 (-17% on a beat).
- Rating: Maintaining Outperform; raising fair value to $325–400 from $235–280. The operating thesis is materially stronger than at Q4 (DC accelerating +57% vs +39%, Q2 guide $11.2B vs prior $10.5B Street consensus, FY26 trajectory now implies $44–48B+ revenue / $9–11 EPS). Fair value updates accordingly to $325–400 (28–33x forward FY27 EPS of ~$11–12). At the post-print AH level near $325 (after -5% reversal), the stock is still toward the lower end of our updated framework but no longer the deeply discounted setup we had at the Q4 print's $200 low. Conviction high on operating story; calibrated humility on valuation cushion.
Results vs. Consensus
Q1 2026 confirmed the structural operating leverage thesis we built through 2025 with even greater force than Q4 already demonstrated. Revenue beat consensus by $260M (+2.6%); EPS beat by 9.6% on operating leverage as Data Center mix improved further; and the Q2 guide arrived $680M (+6.5%) above Street consensus with margins guided 50bps higher than expected. The composition of the beat is the most important data point: Data Center alone added $400M of revenue vs Q4 25's $5.4B (+7% QoQ in a typically seasonally weak quarter), while overall revenue was nearly flat sequentially. This means non-DC segments (Client, Gaming, Embedded) collectively held down the headline by ~$400M to enable the DC ramp absorption.
| Metric | Actual Q1 2026 | Consensus | Beat/Miss | Magnitude |
|---|---|---|---|---|
| Revenue | $10.253B | $9.99B | Beat | +2.6% |
| YoY Revenue Growth | +37% | ~+34% | Beat | +300 bps |
| Non-GAAP EPS | $1.37 | $1.25 | Beat | +9.6% |
| Data Center Revenue | $5.8B (record) | $5.6B | Beat | +3.6% |
| Data Center YoY Growth | +57% | ~+50% | Beat | +700 bps |
| Q2 2026 Revenue Guide | ~$11.2B ±$300M | $10.52B | Beat | +6.5% |
| Q2 2026 GM Guide | ~56% | ~55.3% | Beat | +50 bps |
Data Center: Acceleration Compounding
The single most important data point in the print: Data Center revenue of $5.8B at +57% YoY represents the strongest YoY growth rate AMD has delivered in the AI accelerator era. The trajectory through the cycle — Q1 25 +57% (with MI308 China still in the comp), Q2 25 +14% (China-ban trough), Q3 25 +22% (reaccel), Q4 25 +39% (full ramp visibility), Q1 26 +57% (acceleration compounding) — tells the underlying story unmistakably: the underlying ex-China growth franchise has now FULLY recovered to the Q1 2025 peak rate and is now compounding from a much larger base.
"Data Center revenue grew 57% year-over-year, driven by the strong demand for our 5th-generation EPYC processors, the continued ramp of MI350 series GPUs, and meaningful early MI355 contribution." — Lisa Su, CEO (paraphrased from press release)
What's important about Q1 26's +57% versus Q1 25's +57%: the dollar base is materially higher ($5.8B vs $3.7B = +57% on a 57% larger base). This is acceleration compounding — the growth rate held steady while the absolute dollar contribution to incremental revenue expanded by ~$2.1B YoY ($5.8B - $3.7B). For perspective, the entire incremental Q1 26 dollar Data Center growth is greater than AMD's total quarterly Data Center revenue just two years ago.
Drivers: (1) MI350 series in full production scale at multiple cloud and AI customer deployments (the multi-customer framing we tracked through 2025 is now operational at full scale); (2) MI355 ramp beginning in 1H 2026 as previously guided, with meaningful first-quarter contribution; (3) EPYC server CPU continuing share-gain march; (4) MI308 China contribution at modest scale (~$100M Q1 per Q4 25 guide; awaiting confirmation in 10-Q); (5) ROCm 7 production deployments at top-tier customers continue scaling.
Assessment: Data Center is now in the fully-loaded scale-up phase. With Q1 at $5.8B annualized = $23.2B, and acceleration into 2H 2026 from the OpenAI 1GW MI450 deployment plus continued MI355 ramp, FY26 Data Center revenue likely lands at $26–30B vs. our Q4 25-update model of $32–36B (which now looks roughly correct or even conservative). The 60%+ FY26 DC growth framework management offered at Q4 25 is now likely conservative.
The Q2 2026 Guide: Breaking the Seasonal Pattern
The Q2 guide of ~$11.2B ±$300M is the second standout positive of the print. Three things matter about this number:
- Magnitude vs Street. $11.2B midpoint is $680M above $10.52B Street consensus — a 6.5% raise that exceeds even bullish whisper expectations. The buyside was modeling Q2 in the $10.5–10.8B range; AMD is guiding ~$400M above the upper end of that range.
- Sequential dynamics. +9% QoQ from $10.25B Q1 to ~$11.2B Q2 is a meaningful sequential acceleration. AMD has historically shown sequential growth of 3–6% in Q2 as semi-custom (gaming) and consumer rebound from Q1 trough. +9% QoQ suggests Data Center is contributing far more incremental dollar growth than seasonal history predicts.
- Margin expansion. Non-GAAP gross margin guide of ~56% is +50 bps from Q1's implied ~55.5% and would mark the highest GM since the pre-MI308 trough. Critically, this suggests the "dollars over rate" framing Lisa Su gave at Q3 25 is starting to deliver rate expansion as well as dollar expansion.
Market Reaction: Positioning Unwind, Not Fundamentals
The setup heading into the print made the post-print fade nearly inevitable:
- Pre-print stock: $341 at May 5 regular close, having rallied from the Q4 2025 post-print low of ~$200 (a 70% rally in 3 months).
- YTD: +59% (per FXLeaders); 1-year: +255% (per 24/7 Wall St). AMD was the second-best performing AI-momentum name behind only NVDA over the trailing 12 months.
- Forward valuation: mid-30s P/E on FY26 estimates entering the print — elevated by AMD historical standards.
- After-hours pop: initial +6% on the headline beat (per CNBC).
- Reversal during call: faded to ~-5%+ as buyside digested that the Q2 guide, while above Street, was below the most bullish whisper of $11.5–12.0B that AI-momentum desks had been modeling.
This is a textbook positioning unwind in a quality name — identical pattern to Q3 25 (-9% on a beat) and Q4 25 (-17% on a beat). What was different at Q4: the stock had pre-print rallied to all-time highs ($250+) and then dropped to $200, creating a clear entry-point reset that triggered our upgrade-to-Outperform. This time, the post-print ~$325 level (after -5% reversal from $341) is well above our prior Q4 fair value range of $235–280, meaning the pre-print rally captured most of the operating thesis upside before the print delivered.
The asymmetry: at Q4 25, we had an Outperform with the stock 14% below our FV midpoint of $258. At Q1 26, we have an Outperform with the stock now slightly above our prior FV ceiling of $280, requiring a fair value raise to maintain the rating. We're doing exactly that — but the margin of safety is materially compressed.
Updated Fair Value Framework
Our prior Q4 25 FV framework of $235–280 was based on FY26 EPS of ~$8.50–9.75 at a 28–32x forward P/E. The Q1 26 print + Q2 26 guide require us to update both the EPS estimates and the relevant forward period:
| Item | Q4 25-Update | New Q1 26-Update | Reason |
|---|---|---|---|
| FY26 Revenue | $46–52B | $45–50B | Tightened on H1 26 visibility; Q1 + Q2 guide imply midpoint near $46B |
| FY26 Non-GAAP EPS | $8.50–9.75 | $8.50–9.50 | Trimmed slightly on tighter FY26 revenue range; margin slightly above prior |
| FY27 Revenue (NEW) | not modeled | $54–62B | OpenAI 1GW MI450 full year + continued share gains + Helios scale |
| FY27 Non-GAAP EPS (NEW) | not modeled | $11–13 | FY27 is the relevant forward 12mo by mid-2026 |
| Forward P/E (FY27) | n/a | 28–33x | Justified by improved multi-year visibility from OpenAI deal |
| FV Range | $235–280 | $325–400 | Mid-FY27 EPS of $12 × 28–33x = $336–396 |
Stock at post-print AH ~$325: right at the low end of our updated FV range. This is no longer a deeply discounted setup; it is a "fair value with operating thesis support" setup. The thesis can keep working from here, but the entry point is materially less attractive than at the Q4 25 print's $200 low.
Thesis Scorecard Post-Earnings
| Thesis Point | Status | Notes |
|---|---|---|
| Bull #1: Data Center GPU is a credible second source to Nvidia | Strongly confirmed | +57% Q1 acceleration; $5.8B record; multi-customer ramp at full scale |
| Bull #2: EPYC continues structural share gains in server CPU | Confirmed | Continued share gain march; meaningful Q1 contribution |
| Bull #3: Operating leverage at scale | Strongly confirmed | Q2 guide 56% GM = +50bps QoQ; rate expansion now joining dollar expansion |
| Bull #4: Client share gains are durable | Pending Q1 detail | Need 10-Q segment breakdown; 5+ consecutive quarter trajectory likely intact |
| Bull #5: Sovereign AI as second-leg customer vector | Continuing | Engagements expanding globally per ongoing commentary |
| Bull #6: OpenAI as structural anchor customer | Tracking to plan | 1GW MI450 deployment confirmed for 2H 2026 |
| Bull #7: ROCm 7 production adoption | Scaling | Production at multiple top-tier customers continues |
| Bull #8: MI308 China limited resumption | In progress | ~$100M Q1 contribution per prior guide; ongoing |
| Bear #1: China export controls erode TAM | Mitigated | Limited resumption holding; not material to the bull thesis |
| Bear #2: Hyperscaler ASIC erosion | Neutral | 2027+ concern; OpenAI commitment counter-data-point still valid |
| Bear #3: Valuation entry point | Worsening | Stock $341 pre-print vs prior FV ceiling $280; margin of safety compressed |
| Bear #4: Whisper-vs-guide gap = positioning unwind | Confirmed (3rd straight) | ~-5% post-print reaction = same pattern as Q3 25 + Q4 25 |
Action: Maintaining Outperform; raising fair value to $325–400. The Q1 print is the cleanest demonstration to date of the operating thesis we've been tracking since the Q3 25 upgrade: Data Center scaling above expectations, GM rate expanding alongside dollar expansion, OpenAI on track, multi-customer Instinct franchise materializing. We do not downgrade because operationally everything is going right and there's no execution warning sign in the print. We do not upgrade beyond Outperform (3-tier framework). The valuation entry point has deteriorated — the stock at $325 post-print is right at our FV low — so we recalibrate FV upward and acknowledge the margin of safety is meaningfully smaller than at Q4 25. Conviction high on the operating story; humble on the price action volatility, which has now followed the same "beat-and-fade" pattern for three consecutive quarters.