<?xml version="1.0" encoding="UTF-8"?>
<feed xmlns="http://www.w3.org/2005/Atom" xml:lang="en">
  <title>The Silicon Valley</title>
  <subtitle>Independent coverage of the Silicon Valley tech industry — startups, AI, semiconductors, policy, climate, and the people building the future.</subtitle>
  <link rel="self" href="https://thesiliconvalley.news/atom.xml"/>
  <link href="https://thesiliconvalley.news/"/>
  <id>https://thesiliconvalley.news/</id>
  <updated>2026-05-28T12:47:57.857Z</updated>
  <entry>
    <title>The Bay Area's awkward second act: rebuilding office culture in 2026</title>
    <link href="https://thesiliconvalley.news/bay-area-rto-culture-rebuild-2026"/>
    <id>https://thesiliconvalley.news/bay-area-rto-culture-rebuild-2026</id>
    <updated>2026-05-27T18:00:00.000Z</updated>
    <published>2026-05-27T18:00:00.000Z</published>
    <author><name>Maya Chen</name></author>
    <category term="Culture"/>
    <summary>Three years into a sustained return-to-office push, Bay Area tech is discovering that physical proximity does not, on its own, rebuild the trust that distributed work eroded. The result is a quieter, more uncertain office than the one anyone remembers.</summary>
  </entry>
  <entry>
    <title>Commonwealth Fusion Systems targets SPARC start-up by year-end as Google PPA looms</title>
    <link href="https://thesiliconvalley.news/commonwealth-fusion-sparc-late-2026-google-ppa"/>
    <id>https://thesiliconvalley.news/commonwealth-fusion-sparc-late-2026-google-ppa</id>
    <updated>2026-05-26T14:00:00.000Z</updated>
    <published>2026-05-26T14:00:00.000Z</published>
    <author><name>David Park</name></author>
    <category term="Climate"/>
    <summary>CFS has reiterated that its SPARC tokamak will be operational in late 2026 or early 2027, with first plasma the proximate milestone. Behind it sits a Google power purchase agreement for half the output of Arc, the commercial follow-on.</summary>
  </entry>
  <entry>
    <title>Brussels widens AI antitrust scrutiny to the full stack — chips, cloud, models</title>
    <link href="https://thesiliconvalley.news/eu-antitrust-ai-stack-full-investigation-2026"/>
    <id>https://thesiliconvalley.news/eu-antitrust-ai-stack-full-investigation-2026</id>
    <updated>2026-05-25T10:00:00.000Z</updated>
    <published>2026-05-25T10:00:00.000Z</published>
    <author><name>Julia Romero</name></author>
    <category term="Policy"/>
    <summary>EU competition chief Teresa Ribera has signalled the bloc will examine how Big Tech 'entrenches corporate power' across the entire AI value chain, not only at the application layer. The U.S. FTC, separately, has opened an Arm Holdings probe.</summary>
  </entry>
  <entry>
    <title>NVIDIA and AMD resume AI chip sales to China after May policy reversal</title>
    <link href="https://thesiliconvalley.news/nvidia-amd-china-h200-mi308-resume-may-2026"/>
    <id>https://thesiliconvalley.news/nvidia-amd-china-h200-mi308-resume-may-2026</id>
    <updated>2026-05-24T17:30:00.000Z</updated>
    <published>2026-05-24T17:30:00.000Z</published>
    <author><name>David Park</name></author>
    <category term="Chips"/>
    <summary>Both vendors confirmed in May that export licenses cleared for the H200 and MI308 — the de-facto China-tier successors to Hopper and MI300. The policy whiplash matters less than the volume implication: a market that lost $15B+ in pent-up demand is reopening.</summary>
  </entry>
  <entry>
    <title>Nexthop AI raises $500M Series B as networking becomes its own AI category</title>
    <link href="https://thesiliconvalley.news/nexthop-ai-500m-series-b-networking"/>
    <id>https://thesiliconvalley.news/nexthop-ai-500m-series-b-networking</id>
    <updated>2026-05-23T13:00:00.000Z</updated>
    <published>2026-05-23T13:00:00.000Z</published>
    <author><name>Julia Romero</name></author>
    <category term="Startups"/>
    <summary>The company is building open-source-based switching infrastructure specifically for AI data centers. The size of the round signals that the network layer is now a standalone bet — not a feature of GPUs, not a feature of clouds.</summary>
  </entry>
  <entry>
    <title>Google bets on speed: Gemini 3.5 Flash overtakes the flagship at I/O 2026</title>
    <link href="https://thesiliconvalley.news/google-io-2026-gemini-3-5-flash-beats-flagship"/>
    <id>https://thesiliconvalley.news/google-io-2026-gemini-3-5-flash-beats-flagship</id>
    <updated>2026-05-22T16:00:00.000Z</updated>
    <published>2026-05-22T16:00:00.000Z</published>
    <author><name>Maya Chen</name></author>
    <category term="AI"/>
    <summary>At I/O 2026 Google unveiled a model that is cheaper, faster, and — on most benchmarks — better than its own top-end Gemini 3.1 Pro. Paired with Antigravity 2.0 and Managed Agents, the message is unmistakable: the company is done selling chatbots and is now selling agents.</summary>
  </entry>
</feed>