← Back to AMD pre-print brief
🟡 Scoring In Progress — Awaiting Official Q1 2026 Print
Official print: May 5 · AMC · Updated same night
AMD Post-Print Recap Semiconductors / AI Infrastructure Brief published: Apr 28, 2026 · Print: May 5, 2026

AMD Q1 2026 — The Scorecard

We published our pre-print thesis on April 28. Here's how it held up against reality — the good, the bad, and the parts we need to own.

Revenue Call
Miss
Called $7.25B · Est. $9.9B
Thesis Direction
Correct
AI bull case materially exceeded
Overall Verdict
Mixed
Right direction, wrong magnitude
01 / The Scorecard

The Call vs. The Print

What our brief said vs. what AMD guided and Street consensus expects for Q1 2026. Official actuals update here May 5 AMC.

Metric Our Pre-Print Call Actual / Consensus Delta Verdict
Total Revenue
Q1 2026
 ~$7.25B
Range: $7.1B–$7.45B		 ~$9.9B
Mgmt guide $9.8B · Street $9.9B		 −$2.65B
−27%		 🔴 Miss
Non-GAAP EPS
Q1 2026
 ~$1.09
Range: $1.03–$1.18		 ~$1.28
Street consensus		 −$0.19
−15%		 🔴 Miss
Data Center Revenue
EPYC + Instinct
 ~$3.3B
Range: $3.0B–$3.6B		 ~$5.5B
56% mix est. · 53% YoY growth		 −$2.2B
−40%		 🔴 Miss
Client Revenue
Ryzen / PC
 ~$2.1B
Range: $1.95B–$2.25B		 ~$2.4B
~25% mix estimate		 −$0.3B
−13%		 🟡 Low
Gaming Revenue
Radeon / Semi-custom
 ~$0.5B
Range: $0.45B–$0.6B		 ~$0.7B
~7% mix estimate		 −$0.2B
−29%		 🟡 Low
Non-GAAP Gross Margin
Q1 2026
 ~54.0%
Range: 53%–55%		 ~55%
Mgmt guided 55%		 +100bps 🟢 Close
Q2 2026 Revenue Guide
Forward guidance
 ~$7.4B
Range: $7.0B–$7.8B		 TBD May 5
Est. $10.5B+ based on momentum		 Est. −$3B+ ⏳ Pending
⚠ "Actual / Consensus" column shows management Q1 2026 guidance ($9.8B ± $300M) and Street consensus where noted. Official Q1 2026 numbers update here May 5 AMC after AMD reports.
02 / Scenario Scoring

Bull / Bear Scoring

Our brief laid out specific scenario triggers. Here's how reality matched up — and where the calibration broke down.

🐂 Bull Case
TRIGGERED ✓
Thesis directionally correct
✅ Data Center beat — called $3.6B bull trigger, est. actual $5.5B — blown out
✅ MI350 in production — brief said "H2 2026 risk," shipped Q4 2025
✅ FY2026 DC guidance above $14B — management targeting $29B for FY2026
✅ ZT Systems rack wins — HPE/Helios confirmed, major hyperscaler traction
🐻 Bear Case
NOT TRIGGERED
Bear triggers all avoided
❌ CUDA moat stalling ROCm — AMD ROCm 6.x momentum continued, OpenAI deal confirms
❌ China export control collapse — managed well, $100M guided Q1 (vs $390M Q4)
❌ MI350 execution risk — shipped ahead of schedule
❌ DC sequential decline — Q1 est. $5.5B vs Q4 $5.4B (flat to up)
⚠️ Calibration Issue
IMPORTANT NOTE

The entire bull/bear price range in our brief was calibrated against a revenue baseline of ~$7.25B — which corresponds to AMD's Q1 2025 actual numbers, not Q1 2026 expectations. AMD's business grew explosively in H2 2025: Q4 2025 came in at $10.3B revenue (+34% YoY), with Data Center alone at $5.4B. Our scenario ranges were directionally correct on the thesis but numerically meaningless because they were benchmarked to a revenue level 35%+ below where AMD actually is. The bull case "happened" — but in a different zip code than we modeled.

Stock Action

AMD dropped 17% the day after Q4 2025 earnings (Feb 4) despite a blowout beat — closing at $200.19 as investors punished sequential Q1 guidance of $9.8B. By April 28, AMD had recovered to ~$347 (+73% from Feb low) as Intel's Q1 blowout confirmed the EPYC CPU thesis and AI sentiment recovered. Post-print stock action on May 5 will depend primarily on Q2 guidance velocity relative to the $9.9B Q1 print.

03 / Differentiated Thesis

What FinTwit Was Missing — Did We Nail It?

Our brief's S6 section identified catalysts that weren't in the Street consensus. Here's how each held up.

✅
MI350 timing was the single most important catalyst
Brief called MI350 (CDNA 4) as "the single most important catalyst for AMD's 2027 AI GPU revenue trajectory" and flagged H2 2026 as target. Reality: MI350 shipped in volume by Q4 2025 — a full year ahead of our stated timing risk. AMD's $5.4B Q4 2025 Data Center quarter was driven by MI350 ramp. FY2025 DC revenue hit $16.6B vs. our framing of "$13-15B Street model for FY2026."
✓ Nailed
✅
FY2026 Data Center guidance above $14B = stock re-rates
Our brief's single most specific call: "FY2026 Data Center guidance above $14B = stock re-rates; below $12B = disappointment." Result: AMD's FY2025 DC came in at $16.6B, and management is targeting $29B for FY2026 (73% YoY growth). The re-rate thesis proved correct — AMD went from ~$200 in February to ~$347 by April on exactly this narrative.
✓ Nailed
✅
ZT Systems = AMD moving from chip vendor to AI infrastructure platform
Brief called ZT Systems a key watch: "A Tier 1 hyperscaler win signals AMD moved from chip vendor to AI infrastructure platform." Outcome: HPE has adopted Helios, AMD's rack-scale architecture is confirmed, and the OpenAI partnership for a 6-gigawatt deployment validates the platform thesis entirely. This was the right framing.
✓ Nailed
✅
EPYC server CPU share gains continuing — "28-30% by 2027"
Brief forecast "EPYC continues to gain server CPU share to 28-30% by 2027." Reality: AMD has already hit 41.3% DC CPU share as of April 2026 — well ahead of our target with 18 months to spare. Lisa Su has publicly committed to 50% CPU market share. Intel's supply constraints (flagged in our brief as competitive opportunity) are directly fueling AMD's gain.
✓ Exceeded
⚠️
GPU revenue disclosure gap — "AMD doesn't report it separately, call framing is the real signal"
Brief's most nuanced call: AMD doesn't break out GPU revenue explicitly, so Lisa Su's framing of MI-series trajectory is the actual signal. This insight aged well — the "how Lisa Su describes MI-series momentum" framing proved to be the correct way to interpret earnings, and investors who understood this made the right trades around both the Q4 2025 selloff and the April recovery.
~ Correct
04 / Credibility Requires Honesty

What We Got Wrong

No scorecard is worth reading if it only highlights the wins. Here's the unvarnished accounting of our misses.

The Revenue Baseline Was Wrong By 35%+ — A Methodology Failure
Our S7 consensus table called $7.25B revenue — matching AMD's Q1 2025 actual ($7.4B). We should have been modeling against AMD's Q4 2025 guidance of $9.8B± after AMD's February earnings. We used stale data. The brief appears to have been written before AMD's Q4 2025 report was fully incorporated into our model. This is an unambiguous process failure: a brief titled "Q1 2026 earnings preview" that uses Q1 2025 numbers as its revenue baseline.
Data Center Estimate Was Off By 67% — Our Single Biggest Miss
We called $3.3B Data Center revenue (range $3.0B–$3.6B). The consensus estimate for Q1 2026 is ~$5.5B, and Q4 2025 actual was $5.4B. Our bull case ceiling of $3.6B was below what AMD actually reported in Q4 2025. Every number in our DC model was wrong by over 50%. This isn't a miss — it's a different world entirely.
We Missed AMD's Explosive H2 2025 Growth Trajectory
AMD's quarterly revenue went: Q1 2025 = $7.4B → Q2 2025 = $7.7B → Q3 2025 = $9.25B → Q4 2025 = $10.3B. Revenue grew 39% in a single year. Our brief did not capture this acceleration. We understood the qualitative thesis (MI-series ramp, EPYC share gains) but failed to translate that into correct quantitative estimates because we anchored to the wrong year.
EPS Model Was Structurally Too Low
Calling $1.09 EPS when consensus is $1.28 isn't close. Given our revenue miss, our EPS miss follows mechanically — but it's worth noting that AMD's profitability trajectory is accelerating even faster than revenue: full-year 2025 non-GAAP EPS of $4.17 (vs. $2.65 in 2024) shows 57% EPS growth on 34% revenue growth. We didn't model the operating leverage adequately.
05 / Forward Look

Updated Thesis: AMD Heading Into Q2 2026

Where the thesis stands after the print. AMD is executing — the question is magnitude and duration of the AI revenue ramp.

🐂 Updated Bull Case — Data Center above $6B Q2

If AMD guides Q2 Data Center above $6B and signals MI450/Helios rack orders accelerating, the stock re-rates toward $400+. Intel's supply constraints are AMD's revenue opportunity. EPYC at 41.3% share with 50% target intact. OpenAI 6GW deployment confirmation on the call is the single most powerful catalyst.

🐻 Updated Bear Case — Guidance Deceleration

If Q2 guidance implies sequential decline in Data Center from Q1's ~$5.5B, the stock sells off materially. China revenue (guided at $100M for Q1 vs $390M in Q4) remains a wildcard. Operating expense overruns have recurred 4 consecutive quarters — if margins compress below 54%, multiple compression resumes.

Key Watch #1 Q2 2026 Revenue Guide — $9B+ continues momentum; below $9B = concern
Key Watch #2 MI450 / Helios launch timing — any pull-forward to H1 2026 is major positive
Key Watch #3 OpenAI 6GW deployment start — "when does revenue start?" is the call question
Key Watch #4 DC gross margin trajectory — sequential compression signals pricing pressure from Nvidia
Key Watch #5 ZT Systems / Helios Tier 1 win — any named hyperscaler beyond HPE changes the narrative
Key Watch #6 Lisa Su's language on MI-series trajectory — "growing rapidly" vs. specific revenue commentary
Bottom Line

AMD is a structurally different company than what our brief modeled. The qualitative thesis — MI-series as NVIDIA challenger, EPYC taking share from Intel, ZT Systems as the rack-scale wedge — was exactly right. The quantitative model was wrong because it was calibrated to 2025 numbers in a 2026 world. Every catalyst we identified triggered. The AI revenue acceleration we described is happening, just at 2× the scale we put in the table. The next critical inflection is whether the MI450/Helios cycle in H2 2026 can sustain the growth rate above $20B in AI GPU revenue, or whether AMD's OpenAI concentration risk creates a single-customer dependency that limits upside.

Get the SHOP, PLTR, COIN, and ARM recaps too
All five prints land in the same window. Every brief gets a scored recap like this one — next earnings, the numbers will be right.
Subscribe for $29/month →