Is Advanced Micro Devices (AMD)'s AI Catch-Up Strategy Enough? Analysts Remain Cautious

Advanced Micro Devices, Inc. (NASDAQ:AMD) is one of the AI Stocks to Watch. AMD may own a dominant position in the consumer and server CPU markets, but analysts believe it is still a catch-up AI accelerator player. On February 12, DA Davidson analyst Gil Luria initiated coverage on AMD (NASDAQ: AMD) with a Neutral rating and a price target of $220.

According to the firm, AMD’s Instinct GPUs compare great on paper, particularly the MI300X, MI350X, and MI400X offering specifications that appear broadly competitive on a single-node basis.

However, the reality is that customers are not buying spec sheets and that success in frontier AI workloads depend less on standalone chip metrics and more on interconnect and systems-integration. These, it noted, are areas that Nvidia has a hold on and AMD doesn’t.

The result is that real-world model FLOPs utilization on AMD’s hardware at scale runs materially below what the spec sheet implies, which means the effective cost per useful FLOP is worse than advertised.

Customer adoption patterns are a useful indicator, with OpenAI a telling sign. Companies such as the AI startup prefer competing technologies over AMD’s accelerators.

We’d argue the most telling sign is OpenAI, who is arguably the single most compute-hungry lab without a captive relationship to any hyperscaler’s proprietary silicon. And yet, we’d ascertain that OpenAI likely ranks AMD behind not only NVIDIA, but its own Broadcom ASIC program, and potentially even Trainium and TPU if they had access.

Advanced Micro Devices, Inc. (NASDAQ:AMD) develops and sells semiconductors, processors, and GPUs for data centers, gaming, AI, and embedded applications.