Core Insights - Google is launching its most advanced chip, the Ironwood Tensor Processing Unit (TPU), to attract AI companies by providing custom silicon solutions [2][3] - The Ironwood TPU is designed to enhance performance for large AI models and real-time applications, significantly outperforming its predecessor [3][4] - Google is experiencing strong demand for its AI infrastructure, contributing to substantial growth in cloud revenue [5][6] Product Launch - The Ironwood TPU will be available for public use soon, following initial testing and deployment [2] - This chip can connect up to 9,216 units in a single pod, addressing data bottlenecks for demanding AI models [3] - Major clients, such as AI startup Anthropic, are planning to utilize up to 1 million Ironwood TPUs for their models [4] Market Position - Google is competing with Microsoft, Amazon, and Meta in the AI infrastructure space, focusing on custom silicon advantages over traditional GPUs [3] - The company is enhancing its cloud offerings to be more cost-effective and efficient to compete with AWS and Microsoft Azure [4] Financial Performance - In Q3, Google reported cloud revenue of $15.15 billion, marking a 34% year-over-year increase [5] - The company has secured more billion-dollar cloud contracts in the first nine months of 2025 than in the previous two years combined [5] - Google has raised its capital spending forecast for the year to $93 billion, up from $85 billion, to meet increasing demand [5][6]

Core Insights - OpenAI's $660 billion commitment across multiple cloud providers highlights NVIDIA's pricing power challenges [16][17][19] - Amazon's dual strategy positions it favorably in the AI infrastructure market, leveraging both NVIDIA GPUs and its custom Trainium2 chips [4][32][35] Group 1: Amazon's Strategic Moves - Amazon's $38 billion deal with OpenAI boosted its stock by 5%, while NVIDIA's shares rose by 3%, indicating market confidence in AWS's AI capabilities [1] - Anthropic, OpenAI's competitor, is utilizing 500,000 of Amazon's Trainium2 chips, which are projected to scale to over 1 million by year-end, showcasing Amazon's competitive edge [2] - Trainium2 is reported to deliver 30-40% better price-performance for training workloads compared to GPU-based instances, translating to significant cost savings for Anthropic [5] Group 2: Competitive Landscape - OpenAI's reliance on NVIDIA's GPUs contrasts with Anthropic's use of Amazon's custom silicon, which could disrupt NVIDIA's market dominance and threaten its $5 trillion market cap [3][12] - The shift in AI infrastructure economics is evident as leading cloud providers, including Amazon and Google, ramp up in-house chip development, reducing dependency on NVIDIA [13][14] - NVIDIA's market share is projected to decline from 95% in 2024 to 60% by 2027, while Trainium2 is expected to capture 38% of the market [16] Group 3: Circular Economy Concerns - A significant portion of AI infrastructure demand is circular, with Amazon's investments in companies like Anthropic leading to increased AWS revenue, which in turn justifies further capital expenditures [21][22] - Analysts express concerns about the sustainability of AWS's growth, questioning how much of it is driven by organic demand versus circular revenue from invested companies [24] - The potential for a tech stock correction looms if AI productivity gains are limited or delayed, impacting the broader economy [25] Group 4: Investment Implications - The competition between OpenAI's NVIDIA ecosystem and Anthropic's Trainium2 represents a pivotal moment in AI infrastructure economics [32] - For NVIDIA shareholders, the company is transitioning from an "irreplaceable monopoly" to a leading semiconductor firm with normalizing margins, which may compress its valuation multiple [33] - Amazon's strategy of supporting both NVIDIA and custom silicon positions it as a key player in the evolving AI infrastructure landscape [35]

What's been so interesting about these golden touches from open air for some of the big publicly traded companies is usually nobody's been given a sweetener by in video by AMD. Broadcom, it just seems to be independence that they give to open air. Yeah, and I think the model here is the Google TPU model.I mean, when you think about Broadcom's almost $20 billion run rate for A. I. chips, more than half of that is from Google, TPU.So what openai is saying is you help us get there in terms of the ramp up like ...

Product Innovation & Technology - The iPhone Air is the thinnest iPhone ever at 5.6mm, featuring a flexible front camera and dual recording capabilities [1] - The A19 Pro chip, Apple's most advanced, powers both the Air and 17 Pro, integrating neural accelerators into the GPU cores for on-device AI [2][3] - Apple's new modem, C1X, and wireless chip, N1, signify a move towards controlling the full silicon stack for optimized product solutions [4] AI Strategy & Implementation - Apple is focusing on building the best on-device AI capability, integrating neural processing into GPUs for enhanced performance across applications [8][9][12] - The integration of neural accelerators into each GPU core allows for seamless switching between 3D rendering and neural processing instructions [9] - Apple emphasizes privacy, efficiency, and control as key benefits of on-device AI processing [15] Supply Chain & Manufacturing - Apple is increasing its commitment to manufacture in the US, with TSMC ramping up three-nanometer production in Arizona [35][36] - Geopolitical risks associated with relying entirely on Taiwan for chip manufacturing are a concern for Apple [33] - Apple anticipates using its own modems in all devices within the next two years, including Mac and iPad [30] Competition & Market Dynamics - Apple's move to in-house wireless and Bluetooth chips impacts Broadcom and Qualcomm, though licensing deals will continue [18][21][28] - Qualcomm remains the gold standard for modems, particularly in high-end Android smartphones [28] - Apple's custom modem, C1X, uses 30% less energy compared to the Qualcomm modem in the iPhone 16 Pro [23][24]

Financial Data and Key Metrics Changes - Marvell reported record revenue of $2,006 million, reflecting a 6% sequential increase and a strong 58% year-over-year growth [4][25] - Non-GAAP operating margin expanded by 870 basis points year-over-year to 34.8% [4] - Non-GAAP earnings per share reached $0.67, up 123% year-over-year [4][27] - Operating cash flow was $462 million, significantly up from $333 million in the first quarter [5][27] - GAAP gross margin was 50.4%, while non-GAAP gross margin was 59.4% [25][26] Business Line Data and Key Metrics Changes - Data center revenue was $1,490 million, growing 3% sequentially and 69% year-over-year, driven by custom XPU and XPU attached products [11][25] - Enterprise networking revenue was $194 million, and carrier infrastructure revenue totaled $130 million, with combined revenue growing 2% sequentially and 43% year-over-year [19] - Consumer market revenue was $116 million, up 84% sequentially and 30% year-over-year, primarily driven by gaming demand [21] - Automotive and industrial revenue was $76 million, flat both sequentially and year-over-year [21] Market Data and Key Metrics Changes - The data center end market now accounts for 74% of total revenue, up from 34% in 2024 [32] - The total addressable market (TAM) for data centers is projected to increase to $94 billion by 2028, a 26% increase from previous estimates [8] - The company expects data center revenue to continue delivering strong growth in the mid-thirty percent range year-over-year in the third quarter [12] Company Strategy and Development Direction - The divestiture of the automotive Ethernet business for $2,500 million aligns with the strategy to focus on AI opportunities in the data center market [6][7] - Marvell aims to grow its data center market share from 13% of a $33 billion TAM in 2024 to 20% of a $94 billion TAM by 2028 [8] - The company plans to consolidate non-data center end markets into a new single communications and other end market starting in the third quarter [7][33] Management's Comments on Operating Environment and Future Outlook - Management expressed confidence in the ongoing momentum in the data center market, driven by AI demand [4][23] - The company anticipates overall revenue in the third quarter to be approximately $2,060 million, representing 36% year-over-year growth [22] - Management noted that the custom business is expected to see growth in the second half of the fiscal year compared to the first half [39] Other Important Information - Marvell has repurchased $540 million of stock in the first half of the fiscal year, with approximately $2 billion remaining in its authorization [5] - The company completed a public offering of notes totaling $1 billion, using most proceeds to repay existing debt [28] Q&A Session Summary Question: Guidance for the custom business and headwinds in Q3 - Management acknowledged the lumpiness in the custom business but expects a strong recovery in Q4, driven by optics business growth [37][39] Question: Clarity on design wins and revenue expectations - Management reported significant design activity and new design wins, with many opportunities in the billions of dollars [42][44] Question: Impact of supply constraints and tariffs - Management indicated tight supply chain conditions but has not seen significant impacts from tariffs [45][48] Question: Concentration among lead customers and timing of design wins - Management confirmed that initial programs are ramping and additional design wins are expected to contribute to revenue over the next 18-24 months [52][54] Question: Update on three nanometer XPU program - Management stated that initial programs are ramping and expressed confidence in growth from this program next year [106]

Financial Data and Key Metrics Changes - In Q4, the company achieved record revenue exceeding $1,000,000,000 for the first time, with full-year revenue surpassing $4,000,000,000 and royalty revenue exceeding $2,000,000,000 [7][8] - Total revenue for Q4 was $1,240,000,000, with royalty revenue growing 18% year-on-year to a record $607,000,000 [13][14] - Licensing revenue increased over 50% year-on-year to a record $634,000,000, driven by demand for ARMv9 technology [14][15] Business Line Data and Key Metrics Changes - Smartphone royalties increased by 30% year-on-year, significantly outpacing the 2% growth in shipments, indicating a rising value per device [10][14] - The company launched its first ARMv9 Edge AI platform, which has been adopted by major players in the industry [10][11] - Custom silicon demand is driving both licensing and royalty growth, with significant contributions from data centers, automotive, smartphones, and IoT [8][10] Market Data and Key Metrics Changes - The company expects up to 50% of new server chips at hyperscalers to be ARM-based this year, reflecting strong momentum in the data center market [8][62] - The automotive sector has shown strong double-digit growth, with ARM gaining market share in advanced driver-assistance systems (ADAS) [60] - IoT and embedded markets have seen some slowness, but growth is expected to continue, albeit at a slower pace compared to other sectors [61] Company Strategy and Development Direction - The company is focusing on customization of silicon to differentiate performance and unlock unique features, particularly in automotive and hyperscaler markets [42][44] - There is a strategic shift towards direct relationships with OEMs, which may impact traditional fabless semiconductor companies [42][45] - The company plans to aggressively invest in R&D to support customer needs and capitalize on the growing demand for AI technologies [20][21] Management's Comments on Operating Environment and Future Outlook - Management expressed confidence in healthy growth despite macroeconomic uncertainties, citing strong visibility into customer design pipelines and contracted royalty rates [20][21] - The company anticipates limited direct impact from tariffs on royalty and licensing revenues, with potential indirect effects on end demand [16][26] - Future revenue growth is expected to enable increased investment in next-generation technologies [19][20] Other Important Information - The company signed a significant multi-year agreement with the Malaysian government to develop an ARM-based AI ecosystem, indicating potential for future sovereign-level licensing deals [66][68] - The company is seeing strong demand for its compute subsystems, which are now shipping in volume, boosting both mobile and cloud royalty revenue [11][12] Q&A Session Summary Question: Impact of tariffs on revenue - Management clarified that tariffs do not directly impact costs, and any revenue impact would be indirect, primarily affecting demand elasticity [23][25] Question: ARMv9 adoption rate - The adoption rate for ARMv9 has increased to over 30%, driven by custom silicon solutions [32][33] Question: Strategic direction towards OEMs - Management confirmed a trend towards direct relationships with OEMs for customized silicon, which may affect traditional semiconductor companies [42][45] Question: Licensing performance in Q4 - Licensing revenue growth was strong, with a year-on-year increase of 53%, driven by demand for CSS and AI technologies [76][80] Question: Royalty growth by end market - Management indicated strong growth in smartphones and infrastructure, with automotive also showing double-digit growth, while IoT is recovering slowly [60][61] Question: Future of chiplet technology - Management acknowledged the importance of chiplet technology and its integration with ARM's architecture, emphasizing its role in custom silicon solutions [92][95]

Financial Data and Key Metrics Changes - In Q4, the company achieved record revenue exceeding $1,000,000,000 for the first time, with full-year revenue surpassing $4,000,000,000 and royalty revenue exceeding $2,000,000,000 [6][13] - Non-GAAP operating profit reached a record $655,000,000, with non-GAAP EPS of $0.55, at the high end of guidance [16][19] - Royalty revenue grew 18% year on year to a record $607,000,000, driven by flagship smartphone launches [13][14] Business Line Data and Key Metrics Changes - Licensing revenue hit an all-time high of $634,000,000, increasing over 50% year on year, primarily due to demand for ARMv9 technology [7][14] - Smartphone royalties increased by 30% year on year, significantly outpacing the modest 2% growth in shipments [10][14] - The company launched its first ARMv9 Edge AI platform, which has been adopted by major industry players [10][11] Market Data and Key Metrics Changes - The company expects up to 50% of new server chips at hyperscalers to be ARM-based this year, indicating strong momentum in the data center market [7][8] - The automotive sector has shown strong double-digit growth, with ARM gaining share in advanced driver-assistance systems (ADAS) [55] - IoT and embedded markets have experienced some slowness, but signs of recovery are anticipated [56] Company Strategy and Development Direction - The company is focusing on customization of silicon to differentiate performance and unlock unique features, particularly in automotive and hyperscaler markets [41][42] - There is a strategic shift towards signing deals directly with OEMs, which may increase the total addressable market (TAM) [41][42] - The company plans to continue aggressive investment in R&D to support customer needs and capitalize on AI demand [19][20] Management's Comments on Operating Environment and Future Outlook - Management expressed confidence in healthy growth despite macroeconomic uncertainties, citing visibility into customer design pipelines and rising demand for custom silicon [19][20] - The company does not expect a significant direct impact from tariffs on royalty and licensing revenues, but acknowledges potential indirect effects [16][25] - Future guidance for Q1 indicates revenue growth of 12% year on year, with strong expectations for royalty growth between 25% to 30% [17][19] Other Important Information - The company has expanded its developer community to over 22,000,000, the largest in the world [11] - A significant licensing agreement was signed with the Malaysian government to accelerate the development of an ARM-based AI ecosystem [64] Q&A Session Summary Question: Impact of tariffs on revenue - Management clarified that tariffs do not directly impact costs, and any revenue impact would be indirect, primarily affecting demand elasticity [22][24] Question: ARMv9 adoption rate - The adoption rate for ARMv9 has increased to over 30%, driven by custom silicon solutions [30][32] Question: Strategic direction towards OEMs - Management confirmed a trend towards direct relationships with OEMs for customized silicon, which is expected to continue [41][42] Question: Licensing performance in Q4 - Licensing revenue growth was strong, with a year-on-year increase of 53%, and management remains optimistic about future growth [72][75] Question: Royalty growth by end market - Management indicated strong growth in smartphones and infrastructure, with expectations for continued growth in automotive and recovery in IoT [55][56] Question: Licensing deal with the Malaysian government - The deal is seen as a milestone, with potential for similar agreements with other governments in the future [64] Question: Sequential trends in royalty growth - Management expects Q1 to show strong growth, with a seasonal dip in Q2, followed by growth in the latter half of the year [70][71]