Originally published on The Searchless Journal
The deals came so fast that the industry barely had time to process one before the next landed.
On July 17, 2026, The New York Times reported that Meta was in talks to lease computing power to Anthropic in a deal valued at roughly $10 billion over two years. Hours later, The Wall Street Journal revealed that SpaceX was positioning itself to provide cloud computing services to the Department of Defense, expanding a client roster that already includes Google and Anthropic. The same day, Netflix confirmed it had acquired Ben Affleck's AI startup for nearly $600 million. And just days earlier, Apple had filed a lawsuit against OpenAI accusing two former Apple employees of helping steal trade secrets, while sending legal warnings to dozens of former OpenAI staff.
Each story on its own is a business headline. Together, they tell a different story: the AI infrastructure layer is consolidating, and the companies that control compute will control what billions of people see when they search for information, products, and services.
The Compute Layer Is the Search Layer
For the past two decades, search was a software problem. Google won because its PageRank algorithm was better. Bing struggled because its results were worse. SEO existed to game ranking systems. The infrastructure layer — servers, data centers, network capacity — was a commodity.
AI search has inverted that logic. Running large language models at scale requires staggering amounts of compute. Anthropic's Claude, OpenAI's GPT models, Google's Gemini, and Meta's Llama all depend on access to GPU clusters that cost billions to build and operate. The companies that control these clusters have structural advantages that no algorithm can overcome.
The Meta-Anthropic deal illustrates this perfectly. Meta built massive GPU capacity to train its own models. Now it is considering leasing that excess capacity to Anthropic for $10 billion. Meta gets revenue. Anthropic gets compute it desperately needs. And both companies tighten their grip on the layer that determines which AI systems can compete.
This is not a market where a better algorithm can bootstrap a competitor. If you do not have access to tens of thousands of GPUs, you cannot run a frontier model. If you cannot run a frontier model, you cannot power an AI search engine. The moat is physical, not digital.
SpaceX and the Militarization of Compute
SpaceX's move into cloud computing is the most strategically significant development, and almost nobody is covering it correctly.
The company already provides compute capacity to Google and Anthropic through its Starlink-connected data centers. Now it is courting the Department of Defense. This means SpaceX is simultaneously:
- Operating the satellite network that provides global internet coverage
- Running the compute infrastructure that powers frontier AI models
- Serving the largest military in the world as a cloud customer
The implications for AI search are profound. If SpaceX becomes a primary compute provider for both AI companies and the defense establishment, it gains influence over which AI systems have access to scalable infrastructure. It also creates a single point of leverage that governments can use to regulate or restrict AI capabilities.
For brands and publishers, this matters because AI search results depend on which models are well-resourced. If three or four companies control the compute layer, they effectively control which AI assistants consumers use. And those assistants determine which brands get cited, which products get recommended, and which businesses get discovered.
The Apple-OpenAI War Is a Distribution War
Apple's lawsuit against OpenAI is not really about trade secrets. It is about distribution.
Apple controls the most valuable computing platform on earth: more than two billion active devices running iOS. When Apple sends legal warnings to 40 former employees who now work at OpenAI, it is signaling that it will not tolerate OpenAI building alternative distribution channels that bypass Apple's ecosystem.
The lawsuit alleges that two former Apple engineers helped OpenAI replicate Apple's Siri integration architecture, which would allow OpenAI to embed its models deeper into competing platforms. If OpenAI can build native integration without Apple's permission, it reduces Apple's leverage over the AI search experience on its devices.
This is the same battle playing out in Europe, where the EU's Digital Markets Act is forcing Google to open Android to rival AI assistants. The question is not which AI model is best. The question is which company controls the distribution surface — the phone, the operating system, the browser — through which consumers interact with AI.
For brands, this means AI search fragmentation is accelerating. Different AI assistants will have different strengths on different platforms. Optimizing for Google's AI Overviews will not be enough if Apple's Siri integrates a competing model, or if Anthropic's Claude becomes the default assistant on Meta-owned apps.
Netflix and the Content-AI Vertical
Netflix's $600 million acquisition of Ben Affleck's AI startup is easy to dismiss as a Hollywood story. It is not.
Netflix has spent a decade building one of the most sophisticated recommendation engines in the world. That engine determines what 300 million subscribers watch, and by extension, what content gets produced. By acquiring AI capabilities, Netflix is verticalizing: it wants to own both the content and the discovery layer.
This is the same pattern we see in search. Google owns the search engine and the content discovery pipeline (AI Overviews, Google Shopping, Google Business Profiles). Meta owns the social discovery layer and is buying compute to strengthen its AI capabilities. Apple owns the device and wants to control the AI assistant layer.
The companies that own both content and discovery will have outsized influence over what consumers see. This is why brands cannot treat AI search optimization as a channel-specific tactic. The same models that power search results also power recommendation engines, social feeds, and voice assistants. A unified AI visibility strategy is the only approach that will work.
What This Means for Brands and Publishers
The consolidation of AI infrastructure has three immediate implications for anyone who depends on being discovered online.
1. Fewer AI systems matter
In 2024, there were dozens of AI chatbots and search assistants vying for consumer attention. By the end of 2026, that number will shrink to perhaps four or five systems that have the compute resources to operate frontier models at scale: OpenAI (backed by Microsoft), Google, Anthropic (backed by Meta, Amazon, and SpaceX), Meta itself, and possibly Apple if it can build or acquire sufficient compute capacity.
Brands should focus their AI visibility efforts on this short list. Optimizing for a long tail of niche AI assistants is wasted effort if those assistants cannot scale.
2. Vertical integration changes the rules
When the same company owns the AI model, the search interface, and the content platform, optimization strategies that worked in the open web era become irrelevant. You cannot buy ads to appear in Claude's responses the way you can buy Google Ads. You cannot rely on backlinks when the AI model generates answers without citing sources.
The new optimization playbook requires:
- Structured data that AI models can parse: Schema markup, clean entity definitions, and consistent brand information across the web
- Authoritative original content: AI models are trained on web content. If your brand is consistently mentioned in high-quality sources, the model will reference you
- Multi-platform presence: Your brand needs to be visible across the major AI systems, not just one, because each system has different training data and ranking signals
3. Infrastructure costs will be passed to advertisers
AI search is expensive. Every query costs significantly more than a traditional Google search because of the compute required to generate responses. The companies building these systems will need to monetize aggressively.
We are already seeing this. ChatGPT launched advertising with Criteo in March 2026 and hit $100 million ARR in six weeks. Google is inserting ads into AI Overviews. Perplexity is testing sponsored follow-up questions.
As compute costs rise — and they will, given the billions being spent on GPU clusters — advertising in AI search will become more aggressive and more expensive. Brands that build organic AI visibility now will have a structural advantage over those that wait and must pay for visibility later.
The Infrastructure Endgame
The deals announced in July 2026 are not isolated events. They are the visible part of a consolidation wave that will determine which companies control AI-powered search for the next decade.
Meta's $10 billion compute lease to Anthropic creates a strategic dependency. SpaceX's defense computing contract gives it government backing and revenue scale. Apple's lawsuit against OpenAI is a shot across the bow in the distribution war. Netflix's AI acquisition signals that every major platform company is verticalizing.
For the search and discovery industry, the message is clear: the infrastructure layer is being locked up. The companies that own compute, distribution, and content will define how billions of people find information. Everyone else — including brands, publishers, and SEO professionals — will need to adapt to rules set by a handful of players.
The time to build AI visibility is now, while the landscape is still taking shape. Once the infrastructure consolidation is complete, the rules will be fixed, and the cost of entry will be measured in billions, not in content quality or schema markup.
The Searchless Journal covers AI search, GEO, and the future of brand visibility. Read more at searchless.ai/journal.
Top comments (0)