Timing note: AMD is scheduled to report today, Tuesday, August 4, 2026, after the market closes—not tomorrow. The earnings call begins at 5:00 p.m. ET / 2:00 p.m. PT.
AMD’s second-quarter numbers matter, but the stock’s larger question is whether management can convert an increasingly impressive list of AI customers and commitments into a timely, profitable, supply-supported revenue ramp.
The central debate has shifted. Investors are no longer asking whether AMD can participate in AI infrastructure. Partnerships with Meta, OpenAI, Anthropic, Microsoft and others have largely answered that question. The focus is now on:
A routine Q2 beat may not be enough. The most important information will be the third-quarter outlook, fourth-quarter commentary and 2027 visibility.
AMD issued the following guidance in May:
| Metric | Q2 2026 guidance | Q1 2026 actual | Implied change |
|---|---|---|---|
| Revenue | $11.2B ± $300M | $10.25B | Approximately +9% QoQ and +46% YoY at midpoint |
| Non-GAAP gross margin | Approximately 56% | 55% | Approximately +1 point QoQ |
| Non-GAAP operating expenses | Approximately $3.3B | $3.15B | Continued aggressive AI investment |
| Non-GAAP other income | Approximately $60M gain | — | Positive |
| Non-GAAP tax rate | 13% | 13% framework | Stable |
| Diluted shares | Approximately 1.66B | 1.65B | Slight dilution |
Applying management’s assumptions mechanically implies roughly:
That EPS figure is an estimate derived from AMD’s guidance, not an explicit company forecast or analyst consensus.
Q2 2025 included approximately $800 million of inventory and related charges associated with U.S. restrictions on MI308 exports to China. Reported non-GAAP gross margin was only 43%, but AMD said it would have been approximately 54% excluding those charges.
Consequently, year-over-year EPS and reported-margin comparisons will look unusually strong. Investors should focus on the underlying progression rather than the headline percentage increase.
AMD said server CPU revenue should grow more than 70% year over year in Q2, following growth of more than 50% in Q1. Management attributed that acceleration primarily to unit growth rather than pricing, with fifth-generation Turin processors ramping alongside continued Genoa demand.
This may be the most important near-term source of upside because EPYC:
At its July Advancing AI event, AMD said it had reached 46% server revenue share last quarter and raised its estimate for the 2030 server CPU market to more than $200 billion, citing agentic-AI workloads. That forecast is far above the $120 billion estimate AMD discussed on its May earnings call, illustrating how rapidly management’s expectations have increased.
What investors need: Evidence that the Q2 acceleration is based on durable deployments and share gains—not temporary supply tightness, pricing or customer inventory building.
AMD expected Data Center AI revenue to increase by a double-digit percentage sequentially in Q2. However, the most important product ramp—MI450-based Helios—was not expected to become meaningful until later in the year.
At the July event, AMD said:
This makes Q2 AI accelerator revenue less important than management’s commentary about:
The cleanest positive signal would be a firmer statement that the Q4 ramp is intact—or accelerating—without sacrificing too much margin.
Q1 client revenue increased 26% year over year to $2.9 billion, supported by Ryzen demand, a richer mix and commercial PC share gains. AMD expected solid Q2 demand and modest sequential growth for the combined Client and Gaming segment.
The risk lies in the second half. Management previously warned that higher memory and component costs could weaken PC demand, particularly in more consumer-sensitive categories.
Key items to watch:
Client is unlikely to determine the stock’s immediate direction unless management materially changes its second-half forecast.
AMD said second-half Gaming revenue could decline more than 20% versus the first half, reflecting:
Radeon growth can soften the decline, but investors should not expect Gaming to contribute meaningfully to AMD’s near-term growth narrative. A worse-than-expected console outlook would place more pressure on Data Center to carry the company.
Embedded revenue increased 6% year over year in Q1 to $873 million. AMD expected double-digit year-over-year and sequential growth in Q2.
This business remains relevant because its relatively high segment margins can help the corporate mix. Investors should look for:
Q3 represents the bridge between AMD’s current business and the expected Helios-led inflection. A constructive outlook should contain several elements:
AMD guided Q2 non-GAAP gross margin to approximately 56%, up from 55% in Q1. The mix currently benefits from:
The complication is MI450. AMD has said the initial MI450 ramp will be below the corporate average gross margin. Management expects EPYC, Client mix and Embedded to offset part of that dilution.
The best outcome would be:
If AI revenue rises rapidly but corporate margins decline substantially, investors may question how much of the economic value AMD is retaining.
AMD’s agreement with Anthropic is strategically significant: up to 2 gigawatts of next-generation infrastructure, with deliveries expected to begin in the first half of 2027. AMD also agreed to invest up to $5 billion in Anthropic as milestones are met and has reportedly discussed support for data-center lease commitments.
Investors should ask:
The deal is a powerful endorsement of AMD’s platform, but it introduces capital-allocation, counterparty and “circular financing” concerns.
AMD’s arrangement with Core Scientific secures more than 500 megawatts of data-center capacity beginning in 2027, potentially expandable to 2.5 gigawatts. The reported $14 billion figure refers to potential contracted revenue for Core Scientific—not direct AMD revenue.
For AMD, the agreement demonstrates that delivering AI systems increasingly requires coordination across silicon, racks, software, power and real estate. It also highlights how capital-intensive the competition with Nvidia is becoming.
AMD used its July event to introduce rocm.ai, an AI-assisted development and optimization platform intended to make AMD GPUs easier to deploy and tune. Management presented substantial performance improvements from AI-generated kernel and workload optimization.
The software discussion should be evaluated through customer evidence rather than benchmark claims alone:
Hardware availability creates the opportunity. Software determines whether AMD can broaden beyond a small number of deeply supported hyperscale customers.
AMD closed at $484.68 on August 3, approximately:
The pullback has reduced some of the immediate risk, but the stock still reflects substantial confidence in AMD’s 2027 AI opportunity. This is therefore not a low-expectations setup.
Investors are likely to reward:
They may look through a modest Q2 beat if Q3 guidance or Helios timing disappoints.
Read-through: AMD is becoming a credible second full-scale AI infrastructure platform, with CPU strength funding its accelerator expansion.
Read-through: The thesis remains intact, but the stock’s next move depends on actual late-year shipments.
Read-through: AMD has strong customer interest, but the path from announcements to attractive earnings may be longer and more expensive than investors expect.
AMD enters Q2 earnings with excellent operating momentum and stronger strategic positioning than it had a quarter ago. EPYC appears to be accelerating, major AI customers are validating Helios, and the company now has several multi-gigawatt opportunities.
But much of that upside is still prospective.
The decisive issue is not whether AMD reaches its $11.2 billion Q2 revenue target. It is whether management can demonstrate that the expected late-2026 and 2027 AI ramp is:
A beat-and-raise with firm Helios timing would support the bull case. A respectable quarter accompanied by vague Q4 commentary could still disappoint given the expectations already embedded in the shares.