TradingKey - That Advanced Micro Devices (AMD) has become a market darling once again is evident. On Monday, shares in the chipmaker climbed 4 per cent, just 24 hours before its Tuesday afternoon earnings release. The rally followed a week in which the stock notched its seventh consecutive session of gains, marking its most sustained winning streak in nearly a year.
Why is the market so optimistic?
For a long time, AMD’s struggle to wrest meaningful share from Nvidia (NVDA) in the server graphics processing unit (GPU) market has acted as a persistent overhang on sentiment. However, that concern is increasingly being offset by a rapidly accelerating opportunity in server central processing units (CPUs).
In the server CPU arena, where AMD locks horns with Intel (INTC), demand remains remarkably robust. Investors appear heartened by the resilience of the traditional server business; according to FactSet consensus, analysts are bracing for a quarterly revenue of $9.4bn and adjusted earnings of $1.23 per share.
The fundamental drivers are becoming clearer. A recent assessment by KeyBanc suggests that AMD’s server CPU orders for the year are now "near sold out." As hyperscale data centres scramble to lock in capacity, the company is likely to enjoy tighter pricing power, with average selling prices expected to rise by 10 to 15 per cent. This supply crunch is the result of a dense architectural upgrade cycle: global cloud providers are accelerating the retirement of legacy systems in favour of AMD’s fifth-generation EPYC “Turin” and Intel’s Xeon “Granite Rapids” platforms, driving an estimated 30 per cent jump in sales for these products this year.
Compared with Intel, its primary rival in the x86-based chip market, AMD holds a distinct advantage in performance and total cost of ownership (TCO). Because its chips require fewer server racks and promise lower long-term electricity and facility costs, large-scale customers are increasingly tilting their budgets toward AMD. Furthermore, Intel’s own supply constraints are expected to hand more market share to AMD, as cloud giants look to diversify their order books during capacity expansions.
The path is not without its hurdles. Competition from Arm Holdings (ARM) is set to intensify as the designer of the eponymous architecture gains ground. Simultaneously, rising prices for memory chips could throttle growth in the second half of the year. This chain of "suppressed demand, lower-tier configurations, and delayed launches" threatens to act as a drag on PC CPU shipment growth.
While CPU momentum is the current talking point, investors remain fixated on the future of the GPU business—specifically, AMD’s gambit with OpenAI.
In October 2025, OpenAI announced a multi-year deal to deploy 6 gigawatts of AMD’s Instinct AI accelerators over the next five years. The pact includes warrants for up to 160m AMD shares—roughly 10 per cent of its float. Coupled with a 500-megawatt installation deal with Oracle (ORCL), Bank of America (BAC) estimates these contracts could yield over $20bn in net revenue. Analysts suggest this marks "undeniable progress" in AMD’s AI ambitions. Yet, OpenAI remains the only anchor customer for the Helios system; the company’s AI narrative hinges on the success of this partnership and its ability to broaden its customer base for rack-scale products.
At CES 2026, AMD showcased its MI450 AI accelerator and the Helios rack-level system, with deployment slated for the latter half of 2026. The company also unveiled its Venice EPYC processors, built on TSMC’s 2-nanometre process—a move analysts believe will bolster market share and gross margins. In the upcoming earnings call, management is expected to reiterate the production timelines for the Instinct MI455 and Helios, with analysts at Bernstein noting a "high degree of confidence" in the gigawatt-scale opportunity.
Nonetheless, the competitiveness of AMD’s GPUs against Nvidia’s remains an open question, leaving Wall Street "mixed" on the stock’s prospects. From a valuation standpoint, AMD trades at 54 times 2026 expected earnings, far exceeding its historical average. Bulls argue that the AI opportunity justifies this premium, especially if CEO Lisa Su delivers on her target of $100bn in data centre revenue by 2030. Bears, however, caution that any execution missteps—be it production delays for the MI450, high customer concentration, or stricter export controls—could trigger a sharp correction.
The onus is now on the company to provide the granular detail needed to show it can maintain its server hegemony while finally staging a breakout in the AI accelerator race.