苏姿丰强调，AMD的Helios是世界上最好的AI机架平台，是与主要客户紧密合作的成果。 AMD CEO苏姿丰5日在CES 2026发表演讲，介绍最新Helios机架级平台，将于今年稍晚推出，搭载的 Venice CPU以台积电2nm制程打造，配置最新的Instinct MI455X加速器。 苏姿丰并预告，下一代Instinct MI500加速器已在研发中，同样以2nm制程打造，预计明年发表。新的 MI500系列将较2023年推出的MI300X提升AI性能达1000倍。 苏姿丰指出，从ChatGPT问世以来，全球AI使用者从100万激增至超过10亿人，远超过当初网际网络普 及速度。她预测，未来五年内将有50亿人使用AI，计算基础设施也在短短几年内成长100倍。但目前计 算能力远远不足以完成人们可能会做的事情，现在模型的功能也愈来愈强大，几年内全球计算能力要再 提升100倍，才能让AI无所不在。 AMD指出，计算基础设施是AI的基石，技术快速演进，将推动全球计算能力从目前的100 zettaflops， 在未来五年内扩展到超过10 yottaflops的空前规模。佑等级（yotta-scale，10的24次方） ...

Core Viewpoint - TD Cowen analyst Joshua Buchalter reaffirms a "buy" rating for AMD with a target price of $290, identifying it as one of the "best picks for 2026" and suggesting that the recent sell-off presents an attractive entry point [1] Group 1 - The upcoming Helios rack-level platform is expected to ignite AMD's AI business, marking a significant turning point for the chip manufacturer [1]

Core Viewpoint - AMD and OpenAI have announced a groundbreaking GPU supply agreement worth up to $90 billion, utilizing an unprecedented "equity-for-purchase" model that could redefine financing for AI infrastructure [3][4]. Group 1: Agreement Structure - The agreement involves OpenAI purchasing up to 6 gigawatts of AMD Instinct GPUs, with potential sales reaching $90 billion. In exchange, AMD will issue warrants for up to 160 million shares at an exercise price of $0.01 per share [3][4]. - If AMD's stock reaches $600, the value of the 160 million shares could total $96 billion, equivalent to the hardware's value in the agreement [3]. - The structure of the warrants is designed to function like performance-based equity incentives, linking AMD's long-term valuation directly to OpenAI's infrastructure growth [6][9]. Group 2: Strategic Benefits - For AMD, this agreement represents an innovative customer acquisition strategy, converting traditional upfront discounts into equity costs tied to future performance, thus securing a minimum revenue of 1 gigawatt while mitigating dilution risks [9][10]. - OpenAI benefits by ensuring a stable source of non-NVIDIA hardware in a tight supply market and creating a potential self-funding pathway through the appreciation of its AMD shares, which can be used to finance future GPU purchases [11]. Group 3: Industry Implications - This transaction signifies a shift in AI computing power from a mere capital expenditure to a financialized and securitized asset class, showcasing a different ecosystem-building paradigm compared to NVIDIA and Intel [12][14]. - AMD's model does not require partners to co-invest directly but maintains the core of commercial transactions, allowing OpenAI to purchase hardware fully while incentivizing them to expand procurement through equity rewards [15][16]. Group 4: Risks and Challenges - The agreement faces execution risks, including the lack of transparency regarding key timelines and technical triggers for equity vesting, making it difficult to predict revenue recognition [18]. - The target of 6 gigawatts is a ceiling rather than a firm commitment, with future purchases dependent on OpenAI's business needs and performance [19]. - AMD must maintain a stable supply chain for the delivery of such a large GPU cluster, with potential disruptions in any part of the supply chain posing significant challenges [20].

Core Viewpoint - The agreement between AMD and OpenAI represents a groundbreaking financial instrument that transforms hardware sales into equity arrangements, linking AMD's valuation directly to OpenAI's infrastructure growth [2][3][4]. Group 1: Agreement Structure - AMD and OpenAI announced a GPU supply agreement potentially worth up to $90 billion, utilizing an unprecedented "equity-for-purchase" model [2]. - OpenAI will purchase up to 6 gigawatts of AMD Instinct GPUs, with AMD issuing warrants to purchase up to 160 million shares at a strike price of $0.01 per share [2][4]. - If AMD's stock reaches $600, the total value of OpenAI's 160 million shares could reach $96 billion, approximately equal to the hardware's value in the agreement [2]. Group 2: Strategic Implications - The agreement is seen as a win-win, providing AMD with a unique customer acquisition method while linking equity dilution to actual business growth [3][6]. - For OpenAI, the deal ensures a stable source of non-NVIDIA hardware and creates a potential self-funding pathway through the appreciation of its AMD shares [6][8]. Group 3: Financial Innovation - The financial structure of the agreement is designed to function like performance-based equity incentives, avoiding traditional equity dilution while maintaining governance control for AMD [4]. - The unlocking of shares is tied to specific performance metrics, including the deployment of the first gigawatt of MI450 GPUs and future GPU purchase volumes [4]. Group 4: Industry Paradigm Shift - This transaction signifies a shift in AI infrastructure financing, moving from mere capital expenditure to a financialized asset class [8]. - AMD's approach contrasts with NVIDIA's model, which relies on direct investments from partners, instead maintaining a core commercial transaction while incentivizing customer investment behavior [8]. Group 5: Risks and Challenges - The agreement faces execution risks, including supply chain stability and the ability to meet the substantial GPU deployment requirements [10][12]. - The lack of transparency regarding key equity vesting timelines and technical triggers complicates revenue recognition and equity allocation predictions [10].