$Advanced Micro Devices(AMD)$
As a small AMD shareholder, I am not panicking, not only for my small margin.. or after the post-earnings sell-off but I am also not treating every red candle as an automatic buying opportunity.
AMD’s second quarter results were objectively strong. Revenue reached a record $11.54 billion, rising 50% year over year and 13% sequentially. Non-GAAP earnings came in at $1.66 per share, non-GAAP operating margin expanded to 27%, and Data Center revenue surged 107% to approximately $6.7 billion. Data Center alone represented 58% of AMD’s total quarterly revenue.
AMD also guided for third-quarter revenue of approximately $13 billion, plus or minus $300 million, representing roughly 41% annual growth and another 13% sequential increase at the midpoint. Its expected non-GAAP gross margin remains around 56%.
Those are not weak numbers.
Yet the stock initially fell nearly 9% after hours and remained sharply lower during the following session. The problem was not that AMD missed expectations. The problem was that the market’s expectations had already moved far beyond a normal earnings beat.
AMD no longer gets rewarded simply for proving that it belongs in AI
For several years, AMD’s investment story was based on a straightforward question:
Can AMD become a credible alternative to Nvidia?
I believe the answer is now clearly yes. There have been some amazing write ups in this thread on that also.
AMD has competitive CPUs, increasingly capable Instinct accelerators, Pensando networking, ROCm software and the Helios rack-scale architecture. It is no longer pitching customers one isolated component, it is attempting to offer complete AI infrastructure.
AMD also has major commitments behind that strategy. Meta’s expanded partnership covers up to six gigawatts of AMD GPUs, with shipments supporting the first gigawatt expected to begin during the second half of 2026. Anthropic has agreed to deploy up to two gigawatts of MI450-series GPUs through AMD Helios systems, with the first gigawatt expected to begin in the first half of 2027.
These agreements are significant.. but announcements, capacity commitments and customer pipelines are not the same as recognised revenue, free cash flow or sustained market share.
That is where the narrative has changed.
The market is no longer asking whether AMD can sign major AI customers. It is asking:
How quickly can AMD convert these agreements into delivered systems, expanding margins and repeatable earnings?
👀 The three things I am watching before meaningfully adding:
1. The conversion of AI agreements into revenue
AMD has built an impressive pipeline, but the largest MI450 and Helios deployments are only beginning to ramp.
Investors need evidence that these agreements will translate into predictable quarterly revenue not just large headline figures spread across several years.
I will be watching Data Center growth closely over the next two quarters. AMD expects its Data Center business to accelerate during the second half of 2026, so execution now matters more than additional promises.
2. Whether full system AI growth improves profitability
AMD’s 56% non-GAAP gross margin is solid, but Nvidia has conditioned investors to expect extraordinary economics from AI hardware.
AMD is moving from individual processors toward rack-scale systems incorporating GPUs, EPYC CPUs, networking and software. That could increase AMD’s revenue opportunity per customer, but it also introduces greater complexity, implementation costs and execution risk.
The best outcome is not simply selling more hardware. AMD needs to show that its full stack strategy can produce expanding operating leverage and attractive returns.
3. Whether AMD can deliver enough supply on time
Demand appears strong, but demand cannot become revenue without manufacturing capacity, advanced packaging, memory and reliable deployment schedules.
Analysts have highlighted potential constraints involving TSMC’s advanced production processes and CoWoS packaging, with tightness potentially continuing into 2027. That does not destroy the bull case, but it could delay the point at which major customer commitments fully appear in AMD’s financial results.
🐂 The bull case
The bullish interpretation is that this decline reflects an expectations reset rather than a deterioration in AMD’s business.
Revenue is growing rapidly. Data Center has become the company’s largest segment. EPYC continues gaining momentum, Instinct deployments are scaling, and AMD now has genuine support from some of the largest AI customers in the world.
AMD does not need to defeat Nvidia entirely to create substantial shareholder value. The AI infrastructure market may become large enough for AMD to build a very profitable position as the leading alternative.
A period of disappointment could ultimately become healthy if it allows the company’s real operating performance to catch up with the valuation.
The bear case
The bearish interpretation is that AMD’s valuation already assumes near-perfect conversion of its AI pipeline.
Nvidia still benefits from a more established software ecosystem, deeper customer familiarity and stronger full-system adoption. AMD must execute across chips, networking, software, packaging and rack-scale deployment simultaneously.
If major projects experience delays, margins fail to improve or customers continue concentrating their largest workloads with Nvidia, AMD could deliver strong growth while still disappointing shareholders.
That is the danger of owning an excellent company at a price that demands exceptional results.
My position
I currently hold a very small AMD position. This quarter did not make me want to sell, but it also did not automatically convince me to average down aggressively.
My view is cautiously bullish: the underlying business appears stronger than the immediate share-price reaction suggests, but the next stage must be demonstrated through execution.
I would rather see continued Data Center acceleration, evidence of Helios revenue conversion and stable or improving margins than purchase heavily based solely on the size of the decline.
AMD beat the quarter—but the market has moved the examination forward.
The next test is not whether AMD can participate in AI.
It is whether AMD can convert its enormous opportunity into revenue, margins and cash flow quickly enough to justify the expectations already embedded in its share price.
What is your view?
A: The sell-off is a buying opportunity.
B: Hold and wait for MI450/Helios execution.
C: Avoid - the valuation still assumes too much.
Disclosure: I hold a small position in AMD. This post represents my personal opinion and is not financial advice.
Comments