Hello Flavius Thank you for your interest in our models. We like to schedule a call with our potential new clients to better understand your needs as well as to give a high-level explanation of each of our offerings. Please find a time on our calendars to schedule a time to chat. https://meetings.hubspot.com/dwolfson Please let me know if you have any other specific questions about the models or their contents before the call. This helps me show up as prepared as possible for our conversation. I have provided a brief description of each offering below. *Product Links and Descriptions* - AI Accelerator Model : Quarterly forecasts (2023–2027) for AI accelerator shipments and revenue by vendors (NVIDIA, AMD, ASICs) and customers (Microsoft, OpenAI, Google, etc.). Includes supply-demand analysis for wafer capacity, packaging yields, and HBM. - AI Datacenter Model : Forecasts for 5,000+ global datacenters, tracking power usage, carbon output, and infrastructure growth through 2030. Built from permitting data, satellite imagery, and other primary sources. - AI Total Cost of Ownership Model (TCO Model) : Cost-of-ownership analysis for GPU-based AI systems. Covers unit shipments, performance (training / inference), TDP, and future cost curves. - Wafer Fab Equipment (WFE) Model : Forecasts for advanced logic WFE (EUV) by model, units, and revenue. Includes lithography demand by layer and vendor (ASML, Applied, KLA). - AI Optical and Networking Model : TAM and architectural breakdowns for Infiniband and Ethernet in AI datacenters. Includes vendor analysis (Marvell, Arista, Broadcom, etc.) and cost projections for optics, switches, and transceivers. - Tokenomics Model : Macro view of AI token economics use that quantifies global token demand, supply, and spend across training and inference by provider. - Foundry Model : Node-level supply, pricing, utilization, and cycle times for TSMC, Samsung, Intel, SMIC, and others with links to GPU/CPU wafer starts. - Memory Model : Provides a comprehensive, bottoms-up analysis of the global DRAM and NAND flash markets. It tracks the industry from the fab floor to the spot market, enabling clients to gauge historical trends and future supply-demand dynamics. - Industrials Model : The Industrials Model maps the datacenter construction supply chain across 74 equipment categories and 483 suppliers, projecting equipment demand and bottlenecks through 2028. - Energy Model : Translates AI datacenter buildout into power infrastructure demand, providing a granular view of the supply-demand gaps across major electricity markets in the United States. - Chipbook : A monthly AI and semiconductor industry tracker which provides both timely and targeted insights into the overall state of the industry, as well as detailed information about specific sectors and trends. - Core Research - Financial Markets Coverage : Weekly updates on vendor developments, regulation, and AI model launches, combining industry and financial analysis. Contains snapshots of the data, but not the data itself. I'm looking forward to connecting! On Sun, Apr 26, 2026 at 9:31 PM SemiAnalysis wrote: > > Flavius Pop > flavius@abnito.ai > Date: April 27, 2026 at 1:31 am > Source: Memory Model > Device: Chrome (Mobile) > IP address: [image: 🇺🇸] 76.244.38.94 > Name > Flavius Pop > Email > flavius@abnito.ai > Website > www.abnito.ai > Message > — > Mark as spam > View > in dashboard > > [image: Jetpack] Powered by Jetpack Forms >

Interactions 0