{"schema_version":"2026-09-15.story-v1","story":{"slug":"beyond-the-gpu","title":"Beyond the GPU: memory and packaging","description":"Read the processor, memory, and packaging records together, with each source and assessment attached.","type":"reading-guide","editorial_question":"Which parts of the accelerator supply chain does a chip specification leave unexplained?","limits":"These records do not establish a single bottleneck for every accelerator or manufacturer. Product specifications, supplier commentary, and capacity forecasts have different scopes.","canonical_url":"https://www.manufacturing.ai/stories/beyond-the-gpu"},"policy":{"editorial_prompts_are_not_evidence":true,"publication_is_not_human_verification":true,"preserve_assertion_assessments":true,"evidence_entries_may_share_a_source":true,"linked_documents_do_not_add_verification":true,"do_not_treat_a_connection_as_causation":true,"full_topic_records":"https://www.manufacturing.ai/api/v1/topics/{slug}"},"sections":[{"title":"Begin with the processor","editorial_prompt":"Establish what the component does before comparing product specifications or claims about performance.","topic_slug":"gpu","concept_slug":"gpu","topic_url":"https://www.manufacturing.ai/topics/gpu","evidence_url":"https://www.manufacturing.ai/topics/gpu#gpu","assertions":[{"id":"2fb18ebf-dcdc-4b8b-bc7e-c959787ce838","statement":"NVIDIA's own documentation states the architectural bargain: a GPU devotes more transistors to data processing rather than to data caching and flow control, trading single-thread speed for parallel throughput and hiding memory latency with queued independent work.","assessment":"reported","editor_reviewed":false,"reference_only_count":0,"evidence":[{"id":"79362ffa-cfbe-4d5a-a12d-7c67e9933199","stance":"supports","is_primary_for_assertion":true,"origin_independence":"unknown","note":null,"quote":null,"source":{"title":"CUDA C++ Programming Guide","publisher":"NVIDIA","url":"https://docs.nvidia.com/cuda/cuda-c-programming-guide/","sourceType":"company_website","retrievalStatus":"fetched","inspected":true,"publishedAt":null,"content_inspected":true,"documents":[]}}]},{"id":"de81b1b1-0569-48e3-ade4-fbb2c8f53141","statement":"NVIDIA defined the GPU at the GeForce 256's launch on 11 October 1999 as a single-chip processor with integrated transform, lighting, triangle setup/clipping and rendering engines processing at least 10 million polygons per second — 17 million transistors on a TSMC 220 nm process, and the first consumer part to move geometry work off the CPU.","assessment":"reported","editor_reviewed":false,"reference_only_count":0,"evidence":[{"id":"b4de929b-344c-4894-ad07-094384d0271e","stance":"supports","is_primary_for_assertion":false,"origin_independence":"unknown","note":null,"quote":null,"source":{"title":"GeForce 256","publisher":"Wikipedia","url":"https://en.wikipedia.org/wiki/GeForce_256","sourceType":"other","retrievalStatus":"fetched","inspected":true,"publishedAt":null,"content_inspected":true,"documents":[]}}]}]},{"title":"Follow the memory connection","editorial_prompt":"Read the memory interface and packaging description alongside the processor record, rather than treating their specifications as interchangeable.","topic_slug":"hbm","concept_slug":"high-bandwidth-memory","topic_url":"https://www.manufacturing.ai/topics/hbm","evidence_url":"https://www.manufacturing.ai/topics/hbm#high-bandwidth-memory","assertions":[{"id":"a457d882-6787-4f75-9c52-c78bcf3c741e","statement":"HBM stacks DRAM dies vertically over a logic base die using through-silicon vias, then packages the stack beside a processor for higher bandwidth and lower power than conventional off-package memory.","assessment":"reported","editor_reviewed":false,"reference_only_count":0,"evidence":[{"id":"b9064445-8ee5-4445-bbdf-47432b858453","stance":"supports","is_primary_for_assertion":false,"origin_independence":"unknown","note":null,"quote":null,"source":{"title":"What is HBM (High Bandwidth Memory)? Deep Dive into Architecture, Packaging, and Applications","publisher":"Wevolver","url":"https://www.wevolver.com/article/what-is-hbm-high-bandwidth-memory-deep-dive-into-architecture-packaging-and-applications","sourceType":"other","retrievalStatus":"fetched","inspected":true,"publishedAt":null,"content_inspected":true,"documents":[]}}]}]},{"title":"Inspect the packaging evidence","editorial_prompt":"Separate a statement about today's supply from a forecast or an announced expansion. Follow each number to its original source.","topic_slug":"advanced-packaging","concept_slug":"packaging-capacity-constraint","topic_url":"https://www.manufacturing.ai/topics/advanced-packaging","evidence_url":"https://www.manufacturing.ai/topics/advanced-packaging#packaging-capacity-constraint","assertions":[{"id":"02e6caac-72ea-4fe4-9a8f-381769183b7f","statement":"TrendForce reported in June 2026 a CoWoS supply-demand gap of around 20% narrowing to about 10% by year end, with TSMC at 120,000-140,000 wafers a month and OSAT partners adding 50,000-60,000 — figures the report attributes to institutional investors via Economic Daily News, not to TSMC.","assessment":"reported","editor_reviewed":false,"reference_only_count":0,"evidence":[{"id":"8fff22c7-f816-462d-884a-22dc6563edef","stance":"supports","is_primary_for_assertion":false,"origin_independence":"unknown","note":null,"quote":null,"source":{"title":"TSMC CoWoS Supply-Demand Gap Reportedly Seen Narrowing from 20% to 10% by End-2026","publisher":"TrendForce","url":"https://www.trendforce.com/news/2026/06/15/news-tsmc-cowos-supply-demand-gap-reportedly-seen-narrowing-from-20-to-10-by-end-2026-as-capacity-expands/","sourceType":"journalism","retrievalStatus":"fetched","inspected":true,"publishedAt":"2026-06-15","content_inspected":true,"documents":[]}}]},{"id":"8209c308-593e-4d5a-a2c9-637c338f585b","statement":"TSMC describes packaging as having been 'once considered simple backend processes, almost an inconvenience' and now 'critical to a product's performance, function, and cost', with its 3DFabric family spanning both wafer-level frontend and backend technologies.","assessment":"reported","editor_reviewed":false,"reference_only_count":0,"evidence":[{"id":"b9d56fe0-8841-423a-81c3-a4e122dbec58","stance":"supports","is_primary_for_assertion":true,"origin_independence":"unknown","note":null,"quote":null,"source":{"title":"TSMC 3DFabric","publisher":"TSMC","url":"https://3dfabric.tsmc.com/english/dedicatedFoundry/technology/3DFabric.htm","sourceType":"company_website","retrievalStatus":"fetched","inspected":true,"publishedAt":null,"content_inspected":true,"documents":[]}}]}]}]}