Why does it matter if AI companies own both the models and the delivery channels?

History shows that owning both content and distribution under one roof typically creates tension and underperformance. Verizon and AT&T tried to move into media but eventually retreated; Comcast just announced it is cutting loose NBC and Sky News to refocus on broadband. In banking, regulators pushed firms to separate distribution from content, and oil drillers sold off their gas station networks. The article suggests AI may face the same pressure to choose between being the content or the pipes, not both.

What is an example of successful pure-distribution in tech today?

ChatGPT has a billion monthly users, suggesting that a single large platform can work. Alternatively, the article notes that a big software company like Salesforce or Microsoft could become an open highway for any and all AI models—a neutral distributor rather than a competitor to them.

Back to articles

AI's content kings want to own the pipes too—but the industry may not need them

Semafor Tech14h ago5 min read

Key takeaway

The world's leading AI companies are following a historically risky playbook: trying to own both the AI models they create and the consumer apps and APIs that deliver them. Past industries—from media to banking—have learned that vertical integration of content and distribution typically fails, often forcing companies to spin off one side. The question for AI is whether it will repeat history or break the pattern.

Summaries like this, in your inbox every morning.

3 Key Points

What happened
The article examines how AI companies like OpenAI, Google, and Anthropic are trying to build both the AI models (the content) and the consumer-facing apps and developer APIs that deliver them to users. This mirrors historical patterns in media, banking, and energy, where companies have repeatedly tried to own both content and distribution.
Why it matters
History suggests this vertical integration often fails. Verizon and AT&T retreated from media after failed Hollywood investments; Comcast just announced it is spinning off NBC and Sky News to focus on broadband. The pattern indicates that distribution networks and creative content typically work better as separate businesses, raising questions about whether AI companies' dual strategy will succeed or whether the industry will eventually split into content-only and distribution-only players.
What to watch
Three scenarios could unfold. ChatGPT's billion monthly users could anchor one large integrated player; an existing software company like Salesforce or Microsoft could become a neutral distribution highway for any AI model; or new AI-native hardware from Sam Altman and Jony Ive could create a distribution channel that makes a separate 'phone company' unnecessary.

FAQ

Why does it matter if AI companies own both the models and the delivery channels?: History shows that owning both content and distribution under one roof typically creates tension and underperformance. Verizon and AT&T tried to move into media but eventually retreated; Comcast just announced it is cutting loose NBC and Sky News to refocus on broadband. In banking, regulators pushed firms to separate distribution from content, and oil drillers sold off their gas station networks. The article suggests AI may face the same pressure to choose between being the content or the pipes, not both.
What is an example of successful pure-distribution in tech today?: ChatGPT has a billion monthly users, suggesting that a single large platform can work. Alternatively, the article notes that a big software company like Salesforce or Microsoft could become an open highway for any and all AI models—a neutral distributor rather than a competitor to them.

Discussion

No discussion yet for this article

Stay ahead with AI news

Get curated AI news from 200+ sources delivered daily to your inbox. Free to use.

Get Started Free

Free · takes 30 seconds · unsubscribe anytime

1 minute a day. The AI essentials.

200+ sources · Email / LINE / Slack

Get it free →