Investing.com — AMD (NASDAQ:) is set to host its “Advancing AI” event this week, where analysts expect to hear about new product announcements that could lift sales and drive the chipmaker’s stock price.
Meanwhile, analysts at Lynx Equity Strategies believe this will be “a fade-the-news event.”
“AMD stock has had a modest run-up over the past week on investor expectations for clarity on MI325X and its follow-ons. We think investors are likely to be disappointed,” they emphasized.
Lynx’s team is skeptical about AMD’s ability to close the gap with Nvidia (NASDAQ:), noting that the latter continues to dominate the market. Despite AMD’s ambitious claims at the MI300X launch nearly a year ago, it has made minimal progress in challenging NVIDIA’s stronghold.
“We do not expect MI325X to change the narrative meaningfully,” the analysts state, suggesting that AMD is likely to “remain a distant second” to its rival.
While AMD is expected to highlight developments in data center CPUs and AI PCs, Lynx believes investor attention will remain firmly on the MI-series. The analysts warn that unless AMD can demonstrate a clear driver for market share gains, the stock could potentially retest its yearly lows.
After AMD’s Computex presentation a few months ago, there was some optimism around the company’s next-gen data center GPUs, especially given the higher HBM density on AMD’s MI325X—288GB compared to NVIDIA’s 141GB on the H200.
However, Lynx notes that customer traction has been weak in the months since. Although AMD management may boast superior benchmark results compared to NVIDIA’s Hopper series, the firm’s analysts remain unconvinced, noting that “while running real workloads in the field, we think MI300X continues to trail NVDA’s H100.”
“We do not expect MI325X to change the narrative meaningfully. Superior HBM density is merely one of the factors at work; the software stack, backplane networking and power consumption matter just as much,” analysts continued.
They believe AMD has high hopes for the MI325X and is expanding its chip-on-wafer-on-substrate (CoWoS) capacity in anticipation of future demand. Notably, Broadcom Inc (NASDAQ:)’s strong guidance for its ethernet connectivity business at GPU-based AI data centers next year could indicate growing demand from AMD AI servers. Microsoft (NASDAQ:) also appears to be a key customer, though Lynx remains uncertain about “who the needle-movers are.”
Adding to the uncertainty, Meta Platforms (NASDAQ:), which endorsed AMD’s launch of MI300X last year, may have seen its commitment to AMD diminish recently.
“We believe META may have firmed up its plans for internally developed silicon as a future alternative to NVDA. And Amazon/AWS too, we doubt if AMD is seen as a viable alternative to NVDA,” Lynx explains.
The firm also mentions Oracle (NYSE:) as a potential customer, along with interest from enterprises, but it cautions that the MI-series, along with its software stack, may not be the plug-and-play solution AMD has marketed it as.
“Outside of MSFT we do not believe end customers are likely to spend resources to help bridge the gap as it is unclear what would be the payoff,” said the analysts.
“We expect the stock to trade down after the event today and likely to revisit the lows of the year.”
As for potential surprises at the event, Lynx suggests that while Amazon (NASDAQ:) or Google (NASDAQ:) might be announced as new customers, this would likely be limited to externally facing cloud instances.
