The deep integration with OpenAI has bolstered AMD’s stock price, doubling it on the back of AI-driven optimism. However, its earnings guidance failed to impress, leading to a sentiment-driven pullback amid the AI frenzy.
According to Zhitong Finance, NVIDIA's long-time major competitor in the data center and PC chip sectors, AMD (AMD.US), released its latest earnings report early Wednesday morning Beijing time. AMD is one of the few companies globally capable of competing at the GPU-level computational power against NVIDIA’s near-monopolistic 'NVIDIA AI GPU series' in the AI infrastructure domain. As such, some top Wall Street investment firms are optimistic about AMD gradually eroding NVIDIA's dominant market share—estimated at 90% of a hundred-billion-dollar market—and achieving explosive revenue growth far outpacing NVIDIA's. Following AMD’s agreement with OpenAI for a 6GW AI GPU procurement deal and an innovative 'equity-AI computational supply wager contract,' market sentiment toward AMD's fundamental outlook and future valuation has grown increasingly bullish.
However, after the most recent wave of investment fervor in the AI computational power supply chain significantly drove up share prices of NVIDIA, AMD, Broadcom, Micron, and other leaders in computational power, as well as major AI procurement deals that propelled AMD’s stock higher—raising Wall Street’s expectations for AMD’s performance—the company’s latest official earnings guidance failed to impress institutional investors. AMD’s latest results did not fully reinforce the 'AI faith' as some investors had hoped.
In a recent statement, the chip company indicated a midpoint revenue outlook of approximately $9.6 billion for the fourth quarter—with a guidance range of $9.3 billion to $9.9 billion. The expected gross margin for Q4 under NON-GAAP standards is approximately 54.5%, aligning with the average analyst forecast.
Although this outlook exceeds the average analyst expectation of around $9.2 billion, some analysts had projected figures as high as $9.9–$10 billion. Clearly, given the more than 3% drop in AMD’s shares during after-hours trading following the earnings release, the market was less than satisfied, even slightly disappointed. This reaction came despite Wall Street continuously revising upward its earnings estimates and target price for AMD after the company struck significant procurement agreements with OpenAI and Oracle. Positive news flow and bullish market sentiment had driven AMD’s stock up over 55% since October, raising its market capitalization from $270 billion before these deals to approximately $410 billion currently.
Following major cooperation agreements reached with OpenAI and later in October with cloud computing leader Oracle (ORCL.US), investors placed aggressive bets on simultaneous increases in AMD’s stock price and financial performance. These companies plan to make large-scale purchases of AMD’s AI GPUs for their AI infrastructure buildouts, rather than relying solely on NVIDIA’s Blackwell architecture-based AI GPUs. The market is eager for AMD to eventually challenge NVIDIA’s absolute dominance in the AI computational infrastructure space.
However, AMD’s latest forward-looking guidance suggests that the realization of profits and returns from its AI GPUs may be slower than some investors had anticipated.
Following the announcement, the stock fell nearly 4% in after-hours trading. As of Tuesday’s U.S. market close, prior to the earnings report, AMD’s stock had doubled year-to-date, closing at $250.05.
Data center and PC businesses drive strong growth for AMD in Q3
In terms of Q3 performance, overall results exceeded market expectations. However, the stimulative effect of Q3 results on stock prices and market sentiment was far less impactful than the outlook for future performance. AMD’s total revenue in Q3 increased by 36% year-over-year to $9.25 billion, surpassing the Wall Street consensus estimate of $8.7 billion. The majority of this contribution came from the data center business unit — the very division that provides the strongest GPU computing power competitor to NVIDIA's Blackwell AI GPU as well as high-performance CPUs for data centers.
AMD’s operating profit in Q3 surged 75% year-over-year to $1.27 billion. Excluding certain items, AMD’s adjusted earnings per share (EPS) for Q3 were $1.20, higher than the average analyst estimate of $1.17 per share compiled by institutions, compared to an adjusted EPS of just $0.92 in the same period last year.
During the same period, AMD’s data center revenue increased by 22% year-over-year to $4.3 billion, surpassing the analysts’ average estimate of approximately $4.14 billion. Revenue from client and gaming businesses, which are closely related to personal computers, surged 73% year-over-year to $4 billion, significantly exceeding the analysts’ average forecast of about $2.6 billion.
In the earnings statement, AMD CEO Lisa Su said that these figures mark a “clear upward momentum in our growth trajectory,” as AMD anticipates a substantial increase in revenue and profitability driven by the AI boom.
During her earnings call with global investors and analysts, she stated that by 2027, businesses closely related to AI training/inference computing clusters are expected to generate “tens of billions of dollars” in annual revenue. Meanwhile, she noted that the company’s PC chip business is projected to grow significantly faster than the overall market.
AMD’s recent major long-term cooperation agreements with OpenAI, Oracle, and the U.S. Department of Energy reflect a substantial increase in market demand for its MI series of AI GPU product lines. These products compete directly with NVIDIA’s Blackwell AI GPU and are used for training large-scale AI models and handling massive AI inference workloads.
Unlike NVIDIA, AMD is also one of the largest suppliers of GPUs and high-performance central processing units (CPUs) for personal computers and data center servers. Last month, one of its major competitors, Intel (INTC.US), pointed out that under the current AI boom, there is strong market demand for new AI laptops featuring innovative AI capabilities as well as enterprise-grade server CPUs for data centers. In these key business areas outside of AI GPUs for data centers, AMD has been capturing market share from its long-time rival, Intel.
AMD is the world’s second-largest supplier of GPU series chips, which form the infrastructure for running trillion-parameter AI models in data centers. At the same time, its central processing unit (CPU) series competes directly with Intel’s CPU products in both the personal computer and server markets.
Similar to NVIDIA, AMD has also been drawn into the escalating Sino-U.S. trade standoff. Washington has imposed strict restrictions on the export of its most powerful AI accelerators to certain Asian countries, citing security concerns. U.S. chip giants have been lobbying to ease these rules, arguing that conducting more business in the Chinese market would strengthen U.S. AI capabilities, but the Trump administration has remained cautious about such lobbying efforts.
AMD had previously warned that these export restrictions to China would result in a revenue loss of approximately $1.5 billion this year. The company stated on Tuesday that its Q3 results did not include revenue generated from shipments of its MI308 series AI chips to China.
The super heavyweight collaboration with OpenAI creates significant room for imagination regarding AMD's stock price and future performance prospects.
Recently, prices of high-performance memory products in the global DRAM and NAND series have continued to soar. Additionally, OpenAI, the world’s most highly-valued AI startup, has secured AI infrastructure deals exceeding $1 trillion. Furthermore, Taiwan Semiconductor, Samsung, and SK Hynix have reported exceptionally strong earnings surpassing expectations and raised their revenue growth forecasts for 2025 and 2026. These developments have collectively reinforced the long-term bull market narrative for AI computing infrastructure sectors such as AI GPUs, ASICs, HBM, data center SSD storage systems, liquid cooling systems, and core power equipment.
However, at present, AMD’s performance outlook has evidently failed to reinforce the “AI bull market narrative logic.” The chip giant’s latest performance prospects clearly fall short of the immense imaginative potential brought by its super heavyweight collaboration with OpenAI.
In its latest bullish research report, HSBC reiterated its 'Buy' rating on NVIDIA’s long-time competitor AMD (AMD.US) and significantly raised its target price from $185 to $310, the highest target price level on Wall Street. Analyst Lee wrote that the company’s recent deal with OpenAI “presents a more favorable bullish outlook” and enhances clarity in earnings forecasts as well as bullish market sentiment. He further added: “There may be additional upside potential in the future from premium pricing and expanded shipments of AI GPUs.” As of Tuesday’s U.S. market close, AMD shares closed at $250.05.
Year-to-date, AMD’s share price has surged over 110%, with the majority of the gains concentrated in the past month. AMD’s market capitalization ranking among S&P 500 components has also jumped from 44th place at the beginning of the year to within the top 20, surpassing $400 billion and continuously setting new all-time highs. The primary catalyst behind AMD’s recent strong rally is a multi-billion-dollar key AI computing infrastructure cooperation agreement between the chip company and OpenAI, a leader in artificial intelligence applications. This move not only validates the robust capabilities of AMD’s AI computing infrastructure technology but has also made Wall Street investment institutions more optimistic about its future financial prospects.
The agreement signed between AMD and OpenAI includes a multi-year, multi-generation supply contract for 6GW worth of AI GPUs, with the first deployment of 1GW scheduled to begin in the second half of 2026, featuring AMD Instinct MI450 series AI GPUs. Even more significant is that AMD will issue up to approximately 160 million stock warrants to OpenAI at an exercise price of about $0.01 per share. If fully exercised, this would equate to granting OpenAI the option to acquire up to approximately 10% of AMD’s equity. The vesting of these warrants follows specific triggers: reaching the first 1GW deployment milestone, expanding to 6GW of GPU deployments, and AMD’s stock price hitting the target level of $600.
The collaboration between AMD and OpenAI represents what the media refers to as the “chips-for-equity model”: OpenAI commits to substantial computing power procurement and helps AMD establish itself as a leading AI computing infrastructure provider, in return receiving a package of extremely low-cost long-term equity options, aligning the long-term interests of both parties.
The MI450 purchased by AMD demonstrates a leap in AI performance compared to AMD’s previous generation MI300X/MI350X, earning it the title of a “memory monster.” In terms of process technology, it adopts Taiwan Semiconductor’s most advanced 2nm (N2) process based on the CDNA 5 architecture. Memory bandwidth surges dramatically from 5.3 TB/s in the MI300X to 19.6 TB/s, and memory capacity increases from 192GB to 432GB. For a major player like OpenAI, which is working on ultra-large-scale multimodal models with extended context lengths for large language models (LLMs), the value of ultra-large memory and ultra-high bandwidth often outweighs incremental TOPS of compute power per card in terms of cost-effectiveness. Additionally, for OpenAI, this establishes a second AI GPU computing front centered around “ultra-large memory/bandwidth + more advanced process technology,” separate from NVIDIA’s Blackwell AI GPU ecosystem.
