Google changed what it thinks is important: Here’s what you need to know (November)
Google changed what it thinks is important: Here’s what you need to know (November)
The global rollout of Gemini 4 Hydrogen in Search on 12 November means you can focus on your niche and show up in broader questions.
In a blog post published last week, Google’s Chief AI Architect Koray Kavukcuoglu confirmed the global rollout of Gemini 4 Hydrogen across many of its properties, from Google AI Studio, the Gemini API, Gemini for business, the Gemini app, Google Search, and paid Google AI Pro and Ultra subscribers in Google Antigravity.
With Hydrogen’s advanced reasoning, we’re also upgrading our query fan-out technique in Search, connecting you to a broader range of helpful links than before, helping you find sites you previously might not have discovered.
For marketing and communications teams, you can focus on your niche while contributing to a much bigger conversation.
What has Google actually done?
Google introduced query fan-out with AI Mode’s experimental Search Labs launch in March 2025, before beginning its public US rollout that May. Hydrogen improves this existing capability, which we will focus on here.
Fan-out effectively allows AI Mode to deal with more complex or layered searches and queries. It allows the AI to break down the query into its component parts, providing a more comprehensive answer taking into consideration each of those components.
For example, should a user ask (this part is illustrative):
“Where can we go for a warm holiday in April with two young children, without flying, where there’s plenty to do and we won’t need a car.”
That looks like one question, breaking down the query, we can see different layers:
- Where can the user reach by train or ferry?
- What is the public transport like at the destination?
- Which countries traditionally have low-rainfall in April?
- What is the predicted temperature for locations currently in April (this might differ from historical data)?
- Are there child-friendly activities located nearby?
- What available child-friendly accommodation is there?
- What are rainy day alternatives?
- Are there local shops nearby?
- How many child-friendly restaurants are there?
Previously, the AI may have been informed by articles written about family holidays in August, but there is no guarantee these articles are still accurate, or know the specifics about the reader (e.g., where the starting location is).
Why is this different?
The focus here is seemingly to improve how thoroughly Google can investigate the parts of a question and find relevant sources for each
If we refer back to the family holiday query above, a rail specialist could explain the journey, a local guide could identify children’s activities, and a transport operator could explain how to get around without a car. Each source contributes something useful without having to dilute their authority to answer the entire holiday question.
Breaking the question down does not guarantee that every source is accurate or up to date. But it allows Google to assemble an answer from sources relevant to the individual requirements, rather than relying on content that tries to cover the whole subject.
What does this mean?
The example query is one many would relate to, but the same theory applies to much more niche or technical queries as well. Instead of a family holiday, the query could ask about the complexities of building an AI data centre.
Using its improved query fan-out, Google would be able to separate the query into areas such as permitting delays for data centre interconnect across multiple jurisdictions, or the risks of a dry summer in a region which overindexes on wind generated power.
Businesses with expertise in one part of a bigger question may find themselves more visible in those answers. This suggests Google will search for the best sources for each of the components, before assembling a response to the query itself. For content marketing and communications, the opportunity is to show how specialist knowledge contributes to something the audience wants to understand.
Whereas marketers may have been tempted to be more generic to make their content more accessible, they can narrow the focus on what they are genuinely experts on.
It fits nicely into the advice we always offer. Start with the audience’s problem, and then work back to your solution. You can connect your expertise to bigger economic or societal themes while staying within the areas where you can demonstrate authority.
Bring us the raw thinking
And we'll turn it into a content marketing plan which engages across different platforms and audience profiles
Tell us the idea
And we'll show you how you can layer the argument, build in redundancy, and support other channels




