AIToday
Large Language ModelsAI Coding AssistantsAI Business & IndustryTomasz Tunguz (Theory Ventures)Published: May 5, 2026, 01:01 JST

Ad-supported AI models can cover compute costs with one search ad every 39 minutes or one content ad every 3 minutes, making ad-only funding viable for open models.

Send on LINE
Ad-supported AI models can cover compute costs with one search ad every 39 minutes or one content ad every 3 minutes, making ad-only funding viable for open models.

3 Key Points

  1. A B200 GPU costs $4.50/hour on spot markets. At Google Search ad rates of $38.40 CPM (cost per thousand impressions), a cluster of 4 B200 Blackwells serving 300 users requires roughly 469 impressions per hour to break even—equivalent to one search ad every ~39 minutes.

  2. Content network ads at $3.12 CPM require ~5,769 impressions per hour, or one ad every ~3 minutes. Even at a reduced effective CPM of $1.50 (accounting for fill rates and revenue share), the frequency doubles to one ad every 90 seconds—still below the ~one ad per minute that hyper-casual mobile games show.

  3. For heavier workloads like agentic coding (which burns 10 to 20x more tokens than passive chat), a hybrid model works: $10/month plus 8 ads per day covers 2 million tokens.

Not sure about something? Ask the AI

Questions and answers are published on this page.

Summaries like this, every morning.

Tomasz Tunguz (Theory Ventures)Read Original Article
Send on LINE

AI news that matters for your work, delivered every morning.

Pick your industry and the AI tools you use, and get news related to your work every day.

Free · 30 seconds with Google · unsubscribe anytimeWhat is AIToday? →

Ask AI

Ask AI anything about this article. The AI reads this article, earlier AIToday articles, and Wikipedia, and cites its sources. Q&As are published on this page for other readers too.

Questions and answers are published on this page.

Related Articles

Next articleAuthor warns against conflating API abuse with legitimate distillation technique as Congress moves toward regulatory action