Guides
Evergreen explainers and how-to guides from The Silicon Ledger: background that stays useful beyond the news cycle.
Evergreen explainers and how-to guides from The Silicon Ledger: background that stays useful beyond the news cycle.
- HBM4 vs HBM3E: What Changed in High-Bandwidth Memory HBM4 doubles the interface to 2,048 bits per stack. How it compares with HBM3E on speed, bandwidth, capacity and power, and which chips use it in 2026.
- What Is CoWoS? TSMC's CoWoS-S, CoWoS-R and CoWoS-L Explained CoWoS is TSMC's 2.5D packaging for AI chips with HBM. How CoWoS-S, CoWoS-R and CoWoS-L differ, how large packages get, and TSMC's plans to 2029.
- PCIe Generations Explained: Speeds From PCIe 1.0 to PCIe 8.0 Each PCIe generation roughly doubles lane bandwidth. Transfer rates, real x16 bandwidth, encoding changes and the status of PCIe 7.0 and 8.0 in 2026.
- High-NA EUV vs EUV: What 0.55 NA Lithography Changes High-NA EUV raises numerical aperture from 0.33 to 0.55 for 8 nm resolution. How it differs from standard EUV, its trade-offs and who uses it in 2026.
- Python 3.15 lazy imports explained: how PEP 810 works Python 3.15 lazy imports (PEP 810) explained: the lazy keyword, -X lazy_imports, __lazy_modules__, startup gains, pitfalls and how to test your code.
- TSMC process nodes explained: N3 vs N2, N2P, A16 and A14 TSMC's N3, N2, N2P, A16 and A14 nodes compared: transistor type, production dates and the speed, power and density gains TSMC claims for each.
- Why Is RAM So Expensive? The 2026 DRAM Shortage Explained RAM costs more in 2026 because AI servers and HBM absorb scarce DRAM capacity. The causes, who makes DRAM, new fab dates and the signals to watch.
- Windows 11 version history: release dates and end-of-support dates Every Windows 11 version from 21H2 to 26H2 with release dates, build numbers and end-of-support dates for Home, Pro and Enterprise, as of October 2026.
- 12V-2x6 vs 12VHPWR: what changed and how to avoid melting connectors 12V-2x6 vs 12VHPWR explained: shorter sense pins, longer contacts, the ATX 3.1 sense-pin power table, why 16-pin plugs still melt and how to seat them.
- Big Tech AI capex tracker 2026: Microsoft, Alphabet, Amazon, Meta, Oracle Quarterly capex from filings for Amazon, Alphabet, Microsoft, Meta and Oracle: Q2 2026 spend, 2026 guidance, finance leases and unsigned lease commitments.
- California DROP: How to Delete Your Data From Data Brokers Use California DROP to delete your data from 600+ data brokers: eligibility, sign-up steps, status meanings, exemptions and what to do if data returns.
- Nvidia GPU Architectures List: Tesla to Blackwell, Rubin and Feynman Every Nvidia GPU architecture from Tesla (2006) to Blackwell, plus Rubin (2026), Rubin Ultra (2027) and Feynman (2028), with nodes and key products.
- BIS Affiliates Rule (50% Rule) Explained: Status and Suspension What the BIS Affiliates Rule (50% rule) does, how ownership is counted, and where its suspension stands as of October 2026, with official sources.
- GAAFET vs FinFET: Gate-All-Around Transistors Explained GAAFET vs FinFET: how gate-all-around nanosheets differ from fins, and how Samsung MBCFET, TSMC N2 and Intel RibbonFET use them. Updated October 2026.
- Source-available licenses explained: BSL, SSPL, FSL and ELv2 Source-available licenses explained: BSL, SSPL, FSL and ELv2 compared, what each restricts, who uses them and when code becomes open source.
- Why SSD and hard drive prices are rising in 2026 SSD and HDD prices are rising as AI data centers absorb QLC flash and nearline drives. What TrendForce and Seagate say, and when relief may come.
- Android developer verification explained: dates, fees and sideloading Android developer verification: enforced since 30 Sept 2026 in four countries, USD 25 fee, free 20-device accounts and how sideloading still works.
- Backside Power Delivery Explained: Intel PowerVia vs TSMC Super Power Rail What backside power delivery (BSPDN) is, how Intel PowerVia on 18A and TSMC Super Power Rail on A16 differ, and the heat and cost trade-offs.
- Entity List explained: how US export blacklisting works Entity List explained: what it restricts, who is on it and how to search the BIS list in plain English, with Federal Register citations.
- CHIPS Act Explained: Grants, the 48D Tax Credit and the 2026 Deadline What the CHIPS Act is, how its $39B in grants work, and how the 35% Section 48D tax credit and its December 31, 2026 construction deadline apply.
- Global Privacy Control (GPC) explained: browsers and state laws What the Global Privacy Control signal is, which browsers send it, and which US states require sites to honor it, with the Sephora case and AB 566.
- Google Summer of Code 2027, Outreachy and LFX: how to apply GSoC 2027 dates are not announced yet. Use the 2026 timeline, plus Outreachy and LFX Mentorship eligibility, stipends and how to apply.
- microSD Express explained: speeds, Switch 2 and compatibility What microSD Express is, how fast it runs (up to 985 MB/s), how it differs from UHS-I and UHS-II, and why the Nintendo Switch 2 needs it.
- GB200 vs GB300 vs Vera Rubin NVL72: Nvidia Racks Explained GB200, GB300 and Vera Rubin NVL72 compared: GPUs, HBM, NVLink and why NVL144 became NVL72, using Nvidia's own published specs as of October 2026.
- NVLink vs UALink vs PCIe: AI Interconnects Explained NVLink, UALink and PCIe compared: what each link does, bandwidth per GPU, who backs it, and where InfiniBand and NVLink Fusion fit (as of Oct 2026).
- QLC vs TLC SSD: Endurance, Speed and Which to Buy QLC stores 4 bits per cell, TLC stores 3. See how that changes endurance (TBW), write speed and price, and when each SSD type is the right pick.
- Thunderbolt 5 vs USB4 v2: Speeds, Features and Differences Thunderbolt 5 is Intel's certified take on USB4 v2. Compare 80 and 120 Gbps speeds, PCIe, DisplayPort, charging and cables before you buy.
- TPU vs GPU vs NPU: How AI Chips Differ and When Each Is Used TPU vs GPU vs NPU explained: architecture, memory, software stacks and why TOPS misleads, with official Ironwood and Blackwell specs as of October 2026.