Release Date: August 04, 2026
For the complete transcript of the earnings call, please refer to the full earnings call transcript.
Positive Points
- Record revenue and profitability in Q2 2026, with revenue up 50% year-over-year to $11.5 billion, driven by strong adoption of EPYC, Instinct, Ryzen, and embedded processors.
- Data center segment revenue more than doubled year-over-year to a record $6.7 billion, now representing 58% of total revenue, fueled by strong demand for EPYC CPUs and Instinct accelerators.
- Announced a new strategic partnership with Anthropic to deploy up to 2 gigawatts of MI450 series GPUs in Helios, with the first gigawatt deployment beginning in the first half of 2027, and expanded partnership with Microsoft for Helios deployment on Azure.
- Raised long-term market expectations, now projecting the data center AI accelerator market to grow over 45% annually to approximately $1.4 trillion by 2030 and the server CPU market to grow over 50% annually to approximately $220 billion by 2030.
- Expects data center segment revenue to more than double year-over-year in 2027, with server CPU revenue projected to grow over 70% and data center AI business to grow well over 100%.
- Gross margin expanded to 56% in Q2 2026, up over 200 basis points year-over-year, reflecting a favorable product mix and growing contribution from the data center business.
- Embedded segment returned to strong growth, with revenue up 19% year-over-year to $977 million, driven by broad-based demand and new design wins, tracking towards a record year with over $18 billion in new design wins.
Negative Points
- Gaming segment revenue declined 31% year-over-year to $779 million, primarily due to lower semi-custom sales at this stage of the console cycle and higher component costs impacting graphics card demand.
- Client business faces a softer PC market in the second half of 2026 due to higher memory and component costs, which could weigh on demand despite expected outperformance relative to the market.
- Data center AI business gross margins are expected to be slightly below the corporate average, potentially creating a headwind as the business ramps significantly in 2027.
- Server CPU supply chain remains tight in the first half of 2026 due to unforecasted demand, which could limit near-term growth opportunities despite plans to increase capacity.
- The company faces potential challenges in ramping complex new products like Helios, with yields expected to improve over the first few quarters, potentially impacting initial gross margins.
- Operating expenses increased 40% year-over-year to $3.4 billion as the company continues to invest heavily in R&D to support long-term growth, which could pressure near-term profitability.
- Higher industry-wide component costs, particularly memory, are impacting the broader market and could affect AMD's ability to maintain margins and meet customer demand for its products.
Q & A Highlights
Q: Can you provide more color on the 2027 data center growth targets, specifically the "more than" language used for server CPU growth of over 70% and overall data center segment growth of over 100%?
A: Lisa Su (Chair and CEO): We are trying to provide a framework for 2027. The server CPU growth of over 70% is a very strong statement given the current business size. The overall data center business is expected to be "well over 100%" growth, driven by the data center AI business growing "well over 100%" due to the strength of our strategic customers and the ramp of Helios. The "more" in our statements indicates we see potential for results to exceed these initial targets.
Q: How much gigawatt of compute does AMD have line of sight into for 2027, and what is the monetization per gigawatt?
A: Lisa Su (Chair and CEO): We have a very significant ramp into 2027. Our strategic anchor customers (OpenAI, Meta, Anthropic) are ramping compute demand, and we have supply to more than meet our guidance. We expect the Helios ramp to be significant over the next few quarters, with Q3 being the beginning, Q4 a step-up, and continued growth into 2027. Regarding revenue per gigawatt, we have previously stated "double-digit billions," and we remain in that range.
Q: Can you unpack the comment that Helios is ahead of your original forecast? Is this related to volumes or yields, and how should we think about gross margins in the early ramp?
A: Lisa Su (Chair and CEO): The comment about Helios being ahead of forecast refers to the overall demand volumes for 2027. Customers have high confidence in Helios as a great addition to the AI portfolio, particularly for inference. Regarding yields, we expect them to improve as we progress through the ramp over the next few quarters, which is typical for a highly complex product like this.
Q: Can you help us conceptualize the sizing of the server CPU market buckets (general purpose, agentic AI sandboxes, and AI front-end nodes) within the new $220 billion TAM by 2030?
A: Lisa Su (Chair and CEO): The agentic AI or "sandbox" segment is expected to be the largest and fastest-growing piece of the TAM by 2030, although it is the smallest today. Our Venice family is designed to be leadership across all these categories, and we see workloads expanding with each new EPYC generation. This gives us confidence to grow substantially ahead of the market.
Q: The press release states data center accelerates in the second half. However, Q2 grew 107% year-over-year, and Q3 guidance implies lower growth. How should we interpret this?
A: Jean Hu (CFO): When we talk about acceleration, we are referring to the second half versus the first half of the year. We do expect the year-over-year growth rate for the data center business in the second half to be higher than the first half, which is the reflection of that statement.
Q: In the server CPU business, are you supply constrained now, and will that tightness persist? How is the supply chain able to support the level of growth you are talking about for next year?
A: Lisa Su (Chair and CEO): The server CPU supply chain is tight now and has been for the first half of the year due to unforecasted demand. However, as we get into 2027, demand is better forecasted, and we expect the supply situation to improve. We feel good about satisfying the over 70% year-over-year growth target, and there may be opportunities for that growth to go higher. We are working across the entire supply chain, including wafers, back-end capacity, and substrates.
Q: How should we think about the revenue cadence for the Instinct business in 2027, and what are the underlying gross margins for that business as it ramps?
A: Lisa Su (Chair and CEO) & Jean Hu (CFO): Your data center AI number for 2027 is probably too low; it should be "well over 100%" growth. The ramp will be a progression over the next few quarters. For gross margins, the 2027 margin will be determined by the pace of the server CPU ramp and the data center AI ramp. While the mix may change quarter-over-quarter, we are optimistic about managing the MI450 ramp, and the strong server business growth will provide a good offset. Other levers like the embedded business will also contribute to overall margin improvement.
Q: Can you provide color on the customer diversity expected in the early stages of the MI400 series ramp, and are there any data center readiness constraints?
A: Lisa Su (Chair and CEO): We have large frontier model companies like OpenAI, Anthropic, and Meta consuming through various CSPs. There are also many other customers interested in Helios at a more regular scale, so we expect good customer diversity, especially in the Q4 and Q1 timeframe. We do not see anything that would give us pause about hitting our targets. The range of outcomes will depend on the ability to bring on more capacity in a timely fashion, and we are working closely with data center operators to accelerate build-outs.
Q: How should we think about OpEx growth relative to the overall 40% market growth you discussed for the next few years?
A: Jean Hu (CFO): We will continue to invest given the large opportunities, but you should expect us to manage OpEx increase to be less than the top-line revenue growth. This is designed to drive more operating leverage and deliver earnings per share that is significantly higher than the $20 target we outlined at our Financial Analyst Day.
Q: Can you talk about the fast inference market and the partnership with Cerebras?
A: Lisa Su (Chair and CEO): The inference market is growing very substantially. Fast inference is becoming more relevant, and our partnership with Cerebras combines their technology with Helios to provide a strong solution for customers. This solution is expected to become available in Q4 in the Cerebras cloud and extend into 2027. We view this as an important part of the market and continue to look at ways to optimize our technologies for various workloads.
For the complete transcript of the earnings call, please refer to the full earnings call transcript.
This stock alert was generated using automated technology and GuruFocus financial data to provide readers with timely and accurate market reporting. This content was reviewed by GuruFocus editorial team prior to publication. Please send any questions or comments about this story to [email protected].
