AI Data Pipeline Forum 2026
2026-10-13
San Jose, United States
| title | author | length |
|---|---|---|
| Bernie Wu, MemVerge | theCUBE + NYSE Wired: AI Data Pipeline Forum 2026 | Bernie Wu | 12m |
Argues that memory architecture is the primary AI bottleneck and that CXL becomes a new midtier between DRAM and SSDs, expanding the hierarchy; describes… | ||
| Pradyumna Narendra, Dell Technologies | theCUBE + NYSE Wired: AI Data Pipeline Forum 2026 | Pradyumna Narendra | 17m |
Dell unveils an 18th-generation PowerEdge portfolio that ranges from Core+ compute-optimized single-socket servers to rack-scale GPU clusters that target… | ||
| Hao Zhong, ScaleFlux | theCUBE + NYSE Wired: AI Data Pipeline Forum 2026 | Hao Zhong | 13m |
Claims a memory bottleneck as GPUs/TPUs outpace data paths and announces PCIe Gen6 SSD controller and CXL DRAM controller chips that sit beyond HBM to feed… | ||
| Rakesh Chopra, Cisco | theCUBE + NYSE Wired: AI Data Pipeline Forum 2026 | Rakesh Chopra | 16m |
Cisco announces the USFP optical form factor with about 4x density and 1.3x power efficiency versus OSFP 1.6T and supports 400 gig per lane, arguing optics… | ||
| Andy Hock, Cerebras | theCUBE + NYSE Wired: AI Data Pipeline Forum 2026 | Andy Hock | 20m |
Claims wafer-scale silicon more than 50 times larger than any conventional chip and a system that delivers up to 15x faster inference by keeping compute and… | ||
| Ofer Shapiro, Resolight | theCUBE + NYSE Wired: AI Data Pipeline Forum 2026 | Ofer Shapiro | 24m |
Claims that nanosecond optical circuit switching replaces millisecond MEMS OCS to behave like a data switch, cutting latency from hundreds of microseconds to… | ||
| Pichai Balaji, ScaleFlux | theCUBE + NYSE Wired: AI Data Pipeline Forum 2026 | Pichai Balaji | 16m |
Claims that NAND must handle 95% of GPU-side memory are backed by concrete numbers: GPUs hit 20–22 TB/s, DRAM sits roughly 37–40× slower and traditional… | ||
| Gopi Sirineni, Axiado | theCUBE + NYSE Wired: AI Data Pipeline Forum 2026 | Gopi Sirineni | 24m |
Axiado reveals a rack-scale Trusted Compute Unit called GOSI 2 and a PIC called the Platform Intelligence Controller that runs entirely out‑of‑band to give… | ||
| Vaughn Stewart, VAST Data | AI Data Pipeline Forum 2026 | Vaughn Stewart | 25m |
Argues that fragile, best-of-breed AI pipelines lose to a single unified platform that combines high‑performance storage and data processing; claims DASE… | ||
| CT Sun, AIC | theCUBE + NYSE Wired: AI Data Pipeline Forum 2026 | CT Sun | 13m |
Argues that memory matters more than compute for modern AI, claiming tiered architectures from HBM to cooler storage keep hot context close while offloading… | ||
| Samir Rajadnya, Microsoft | theCUBE + NYSE Wired: AI Data Pipeline Forum 2026 | Samir Rajadnya | 22m |
Microsoft's pathfinding team scouts new memory and storage tech and claims composable memory is the lever that turns scarce DRAM into usable capacity by… | ||
| Siamak Tavallaei, Samsung | theCUBE + NYSE Wired: AI Data Pipeline Forum 2026 | Siamak Tavallaei | 11m |
Argues that OCP stitches standards like JEDEC PCIe Ethernet and CXL into deployable systems so accelerators shift from bolt‑on cards to core components, with a… | ||
| Ben Lee, Supermicro | theCUBE + NYSE Wired: AI Data Pipeline Forum 2026 | Ben Lee | 14m |
Argues that modern AI factories must be designed as a single, data-first computer, not GPU islands: rising SSD and memory costs force co‑design of storage and… | ||
| nothing matches | ||