I'll research AMD before writing the preview. Let me gather recent earnings, filings, news, and price data in parallel.Let me read the most recent earnings release and call, and pull stock price data.Let me read the Q1 2026 earnings call for forward-looking commentary.Let me check recent analyst expectations and any China/MI308 developments via web search.# AMD (Advanced Micro Devices) — Q2 2026 Earnings Preview

Reports after the close Tuesday, August 4, 2026 · Conference call 5:00 p.m. ET


The setup in one paragraph

AMD heads into this print as one of the most consequential AI reports of the season, but with a stock that has cooled sharply into the event. Shares closed at $484.68 on Aug 3, down roughly 12% over the past month and ~17% below the ~$585 high set in June, even though they remain up well over 100% year-to-date (from ~$223 in early January). Q1 was the "inflection" quarter — Data Center became the primary earnings driver — and the story since then has been a torrent of AI dealmaking (OpenAI, Meta, Anthropic, Core Scientific). The debate now shifts from "is the demand real" to "how fast, how profitable, and how much is already in the price." Options are pricing a ~12% post-earnings move, roughly in line with AMD's recent history.


What the numbers need to look like

Company guidance (given May 5): - Revenue ~$11.2B ± $300M → +46% YoY, +9% QoQ at midpoint - Non-GAAP gross margin ~56% - Non-GAAP OpEx ~$3.3B; OI&E gain ~$60M; tax rate 13%; ~1.66B diluted shares

Street consensus (approximate): - Revenue ~$11.3B (Zacks ~$11.32B) — modestly above the guide midpoint - Non-GAAP EPS ~$1.61 (vs. $0.48 in Q2'25, ~235% growth) - Gross margin ~56%

Context: AMD beat on both lines in Q1 ($10.25B rev, $1.37 EPS vs. ~$1.29 consensus) and the stock jumped ~18.6% that day. The company has topped EPS estimates in most recent quarters, though it did narrowly miss in Q2 2025 — a reminder that this quarter is a historically softer one and that the bar is set by AMD's own guide plus a small Street cushion. Given the guide midpoint sits slightly below consensus, an in-line revenue print may not be enough; the market will want the top end (~$11.5B) plus a strong second-half framing.


The three questions that matter most

1. Gross margin is the real "tell," not the revenue beat

This is arguably the single most important line. AMD guided 56%, up from 55% in Q1. The near-term tension: the MI450/Helios ramp carries below-corporate-average margins and begins to bite in Q4, while server CPU, a richer Client mix, and accretive Embedded revenue are the offsets. CFO Jean Hu framed multiple gross-margin tailwinds (server CPU +70% YoY, Gaming rolling off, Embedded accretive) against the MI450 dilution. Watch whether margins hold in the mid-50s (server CPU strength absorbing the GPU drag) or drift toward the low-50s (mix shift biting early). Long-term model target is 55–58%.

2. The server CPU surprise — is the CPU TAM story holding?

The under-appreciated Q1 bombshell wasn't GPUs — it was CPUs. AMD raised its 2030 server CPU TAM to >$120B (>35% CAGR) from ~$60B at the November Analyst Day, driven by agentic AI spawning incremental CPU demand (CPU:GPU ratios moving from 1:4–1:8 toward 1:1 or higher). Management guided server CPU revenue +70% YoY in Q2 with strength continuing through 2H. Note: at the July Advancing AI event AMD reportedly pushed the 2030 server-CPU TAM even higher (~$220B) — so listen for whether Q2 CPU results (units-led, with some ASP/mix and inflation pass-through) validate this rapidly-escalating narrative, and for any read on the >50% share ambition against a resurgent x86 competitor and merchant/custom ARM.

3. Helios / MI450 ramp cadence and 2027 visibility

AMD reiterated: MI450 sampling underway, Helios production shipments ramp in 2H — initial volume in Q3, significant ramp in Q4, continuing into Q1'27. Management now expects tens of billions in Data Center AI revenue in 2027 and confidence to exceed its >80% long-term AI CAGR target, with lead-customer forecasts running above initial plans. The guidance on this ramp — order visibility, supply availability, power/data-center gating — may matter more than the Q2 quarter itself.


The deal flow backdrop (since Q1)

A remarkable string of announcements has reshaped the AMD AI narrative and is the reason the stock re-rated so hard mid-year:

The counter-narrative: much of this is 2027+ revenue, several deals involve AMD funding its own customers, and expectations are now very high.


Segment scorecard to watch

Segment Q1'26 rev Q1 YoY What to watch in Q2
Data Center $5.78B +57% Guided double-digit QoQ growth in both server CPU and Data Center AI; server CPU +70% YoY. Any MI355X/MI450 color.
Client & Gaming $3.61B +23% Client (~$2.9B Q1) expected modest QoQ growth, still YoY growth full-year despite memory/component headwinds.
— Gaming $720M +11% Management guided 2H gaming revenue down >20% vs. 1H on memory/component cost pressure and console cycle.
Embedded $873M +6% Expected to return to double-digit YoY growth; highest-margin segment, accretive to GM.

Key risks / swing factors


Bottom line for investors

The quarter itself is largely pre-announced ($11.2B guide). The share price reaction will hinge on three things: (1) whether gross margin holds in the mid-50s as the dilutive MI450 ramp begins; (2) confirmation that the server CPU surge (+70% YoY) and the escalating CPU TAM are real and sustainable; and (3) crisp, credible detail on the 2H Helios/MI450 ramp and 2027 "tens of billions" AI trajectory. With expectations elevated but the stock already down mid-teens from its highs, AMD needs to beat and raise the second-half framing to re-rate toward the $580+ bull case; a merely in-line print risks testing support in the mid-$400s.

Note: Figures reflect AMD's Q1'26 results/guidance and third-party consensus and news gathered ahead of the report; actual consensus varies slightly by data provider (revenue ~$11.3B, EPS ~$1.55–$1.64).