The 50/50 Split
WebPulse industry data shows media and entertainment as the most evenly divided sector: approximately 50% legacy frameworks, 50% modern. The BBC runs Next.js. Smaller publishers run WordPress. The divide tracks with organizational size and technical sophistication — but the consequence is the same. Half the media industry produces content optimized for human browsers. Half produces content optimized for both humans and machines.
Content Companies With an AI Visibility Problem
Media companies sell attention. Their product is content. When AI agents crawl, summarize, and cite web content — which they do billions of times per day — the publisher with clean semantic HTML and structured data gets cited. The publisher with WordPress's plugin-injected noise gets skipped or misrepresented.
This is an existential issue for content businesses. AI training data, AI-generated summaries, AI-powered search results — all of these preferentially consume clean, structured content. WordPress outputs 2,000+ lines of HTML per page on average. Astro outputs 200-400 lines of clean semantic markup. The AI agent processes both — but one costs 5-10x more tokens and produces worse extraction results.
The RSS Renaissance
RSS — the protocol media abandoned in favor of social media distribution — is the AI agent's preferred content format. Structured, chronological, machine-readable, minimal noise. Modern frameworks like Astro and Hugo generate RSS feeds by default. WordPress generates them too, but polluted with plugin artifacts. Media companies that rediscover RSS as an AI distribution channel gain a structural advantage in the machine-to-machine web.
