--- title: "State of CSS 2026 results land, Chrome 151 ships native SPA metrics" url: https://daily.dev/posts/state-of-css-2026-results-land-chrome-151-ships-native-spa-metrics-4wnincbeb source_url: https://daily.dev/posts/state-of-css-2026-results-land-chrome-151-ships-native-spa-metrics-4wnincbeb type: freeform source: "Web Digest" published: 2026-08-06T04:19:28.166Z updated: 2026-08-06T04:19:49.526Z tags: ["webdev", "css", "google-chrome", "tanstack"] reading_time: 5 upvotes: 1 comments: 0 language: en --- > ## Documentation Index > Fetch the complete documentation index at: https://daily.dev/llms.txt > Use this file to discover all available pages before exploring further. # State of CSS 2026 results land, Chrome 151 ships native SPA metrics **[Web Digest](https://daily.dev/sources/webdev_digest)** · 5 min read · 1 upvotes · 0 comments ## Summary The State of CSS 2026 survey is out with 4,902 respondents, and anchor positioning topped both the 'favorite new feature' and 'wish it had better browser support' lists. Chrome 151 lands stable soft-navigation and interaction-contentful-paint entries on the Performance Timeline, and the web-vitals library jumps to v6 — but it's not a drop-in upgrade. TanStack is having a busy week: Charts hits pre-alpha, plus two new lightweight libraries for Markdown and syntax highlighting that cut a docs page from 1.1 MiB to 27 KiB. Figma published numbers on Code Connect's impact on AI agents: 29.5% fewer tokens, 19.6% faster task completion, and a measurable quality improvement. ## Content **TLDR:** The State of CSS 2026 survey is out with 4,902 respondents, and anchor positioning topped both the 'favorite new feature' and 'wish it had better browser support' lists. Chrome 151 lands stable soft-navigation and interaction-contentful-paint entries on the Performance Timeline, and the web-vitals library jumps to v6 — but it's not a drop-in upgrade. TanStack is having a busy week: Charts hits pre-alpha, plus two new lightweight libraries for Markdown and syntax highlighting that cut a docs page from 1.1 MiB to 27 KiB. Figma published numbers on Code Connect's impact on AI agents: 29.5% fewer tokens, 19.6% faster task completion, and a measurable quality improvement. --- ## Chrome 151 ships native SPA metrics, web-vitals v6 requires careful migration Chrome 151 lands stable support for soft-navigation and interaction-contentful-paint entries on the Performance Timeline — the first time SPAs can be measured natively without workarounds. The web-vitals v6 library ships alongside it with a one-second requestIdleCallback cap and improved INP reporting after bfcache restores. The catch: one document can now produce multiple metric lifecycles, which means existing dashboards risk double-counting, population shifts, and false regressions. The recommended path is a parallel run — comparing v5 and v6 side by side and reconciling percentiles before cutting over. Chrome 152 Beta also introduces a coarse CPU Performance Tier API for device segmentation. [Read more](https://daily.dev/posts/Ia9s4dkOC) ## State of CSS 2026: anchor positioning is developers' favorite feature they can't actually use The 2026 State of CSS survey results are in, with 4,902 respondents. Anchor positioning topped both the 'favorite new feature' list and the 'wish it had better browser support' list simultaneously — which tells you everything about where CSS enthusiasm and browser reality are currently misaligned. The most-used newer features are :has(), aspect-ratio, and CSS nesting. CSS gap decorations (covered in yesterday's digest) shipped in Chrome and Edge 149. [Read more](https://daily.dev/feed-by-ids?id=SBmWLqr5Y&id=95LEJbgQd) ## TanStack ships Charts pre-alpha, plus Markdown and Highlight libraries that cut 1.1 MiB to 27 KiB TanStack Charts 0.3.1 is a typed, tree-shakable charting library built on a grammar-of-graphics approach — the framework-neutral scene weighs 8.12 KiB gzip. Separately, TanStack Markdown and TanStack Highlight are in alpha, built to replace a bloated pipeline on tanstack.com that was transferring over 1.1 MiB of script per docs page (358 KiB of that was Shiki alone). Together they bring the content rendering footprint to roughly 27 KiB transferred, which also let the team remove the React Server Components pipeline they'd been using to hide the dependency cost. TanStack also unveiled a new visual identity, hiring Andy Beutler as its first dedicated lead designer. [Read more](https://daily.dev/feed-by-ids?id=h2X7lJrT7&id=sKYnr4t3U&id=h3KuBuJMf) ## Figma measures Code Connect impact: 29.5% fewer tokens, 19.6% faster AI tasks Figma ran 27 test cases using Claude Sonnet 4.5 and Opus 4.7 to measure what happens when AI coding agents have Code Connect context versus generic React representations. With Code Connect, median token usage dropped 29.5%, task duration fell 19.6%, and code quality improved by a full point on a 1-4 Likert scale. The key driver was coverage — the more components are connected, the better the results. A Coinbase case study confirmed agents stopped fabricating components and started using correct imports directly. [Read more](https://daily.dev/posts/a99y4btH1) --- ## Also notable - **CSS keylogger attack presented at Black Hat USA 2026:** PortSwigger researcher Gareth Heyes demonstrated at Black Hat USA 2026 that CSS alone — no JavaScript, no attachments — can be used to build keyloggers and exfiltrate data from webmail platforms, with major vendors quietly patching vulnerabilities after initially dismissing his disclosures. [Read more](https://daily.dev/posts/h8kWZyKKi) - **Node.js 26.7.0 released with Perfetto tracing and --test-coverage-include-all:** Node.js 26.7.0 adds Perfetto tracing support for diagnostics, crypto support for loading private keys through STORE loaders with updated NSS 3.125 root certificates, Symbol.dispose in ModuleHooks, and a --test-coverage-include-all flag for the test runner. [Read more](https://daily.dev/posts/qWDptySuw) - **Qwen 3.8 Max benchmarks at rank 12, open-weight 27B version anticipated:** Alibaba's Qwen 3.8 Max is a 2.4 trillion parameter model priced at $2/1M input tokens and $6/1M output tokens with a 1M context window, benchmarking at rank 12 behind DeepSeek and Grok 4.5 — underwhelming for its size, though it delivers comparable results to Claude Opus 5 at roughly one-eighth the cost for frontend and visual reasoning tasks. [Read more](https://daily.dev/posts/5TwBWIzlk) - **vlt claims 73% faster package installs with built-in malware blocking:** vlt is a JavaScript package registry and client positioning itself as an npm alternative, claiming 73% faster installs and 38% faster registry performance, with automatic malware blocking on every public package request and restricted lifecycle scripts by default. [Read more](https://daily.dev/posts/ZgJ5x2gIt) - **TIME serves AI crawlers a separate markdown site with embedded sponsored content:** TIME is serving ClaudeBot, PerplexityBot, and OAI-SearchBot a stripped-down markdown version of its site roughly 23x smaller than the human-facing HTML, with sponsored content blocks from advertisers like Ally Bank billed by token count rather than pageview — TIME reportedly sees more bot traffic than human traffic on most days. [Read more](https://daily.dev/posts/KRTrmg2fa) ## Similar posts on daily.dev - [Issue \#753: The new CSS feature devs love most but can't rely on yet — Frontend Focus](https://daily.dev/posts/issue-753-the-new-css-feature-devs-love-most-but-can-t-rely-on-yet-frontend-focus-95lejbgqd) · Frontend Focus · 1 upvotes · 0 comments --- Tags: [#webdev](https://daily.dev/tags/webdev), [#css](https://daily.dev/tags/css), [#google-chrome](https://daily.dev/tags/google-chrome), [#tanstack](https://daily.dev/tags/tanstack) [View this post on daily.dev](https://daily.dev/posts/state-of-css-2026-results-land-chrome-151-ships-native-spa-metrics-4wnincbeb)