Koah Raises $20.5M Series A to Build 'AdSense for AI' as Conversational Apps Scramble for Revenue

Every new interface layer on the internet eventually gets an advertising business. Web pages got banner ads. Mobile apps got interstitials. Social feeds got sponsored posts. Now conversational AI gets Koah.

The San Francisco-based company announced a $20.5 million Series A on February 24, led by Theory Ventures. The round builds on a $5 million seed from September 2025, bringing total funding past $26 million. Tomasz Tunguz, Theory's founder and general partner, joins the board. Before starting Theory, Tunguz was at Google. Not in some adjacent role. He helped build Google AdSense, the product that turned the open web into a monetization machine for millions of publishers.

That pedigree isn't accidental. Koah is explicitly positioning itself as "AdSense for AI," the infrastructure layer that lets developers monetize generative AI experiences without destroying the user experience that makes those products valuable in the first place.

"Generative AI has changed how people consume information, and monetization must evolve alongside it," said Nic Baird, Koah's co-founder and CEO. "Subscription models alone don't scale given high inference costs, and legacy ad models erode user experience."

Here's the math that explains why this company exists. Running a generative AI application is expensive. Every query a user sends triggers an inference call that costs the developer money. Subscription revenue helps, but most consumer AI apps operate on freemium models where the vast majority of users never pay. The result: a growing base of engaged users generating costs without revenue.

Traditional ad formats don't translate. You can't slap a 300x250 banner into a chat interface without breaking the conversational experience that users came for. Pre-roll video ads make no sense in a text-based environment. Pop-ups would be suicidal for retention.

Koah's SDK addresses this by inserting native ad units that sit within or alongside the conversational flow. The ads are clearly labeled, prominently displaying the word "Ad," and they appear in contextual breaks rather than interrupting the conversation itself. Integration takes minutes, according to the company, requiring just a few lines of code.

The traction suggests the approach works. Koah reports more than 2 million monthly active users across its partner apps, with over 35 million native ad impressions served across 175 million queries. Those aren't massive numbers by Meta or Google standards, but for a company that raised its seed less than 18 months ago, the trajectory is notable.

Koah's biggest risk isn't another SDK competitor. It's the platforms deciding to handle monetization internally. OpenAI is already running ads in ChatGPT through its Criteo partnership, charging $60 CPM with a $200,000 minimum. Google's AI Mode has checkout integration. Perplexity tested ads in late 2024. The largest conversational AI platforms have both the incentive and the infrastructure to build their own ad businesses.

Its counter-argument is that the long tail matters more than the head. OpenAI, Google, and Anthropic will monetize their own products. But thousands of smaller AI apps, vertical AI tools, and developer-built AI experiences don't have ad sales teams, demand-side relationships, or the engineering bandwidth to build custom monetization systems. That's the Google AdSense playbook: you don't need the top 10 websites, you need the next million.

Sources: Criteo press release, Koah press materials, industry estimates

There's an ethical dimension Koah can't avoid. When OpenAI began inserting ads into ChatGPT, Zoë Hitzig, an OpenAI researcher, quit the company publicly. Her concern: AI platforms are "centrally collecting intimate information" about enormous populations and could use that data to target and manipulate users.

Koah's response is architectural rather than philosophical. Its ads operate on contextual signals, not personal data profiles. The company matches ads to what users are actively discussing, not to behavioral histories or inferred demographics. Whether that distinction satisfies privacy advocates is debatable, but it's a meaningfully different model than retargeting-era ad tech.

Still, the line between "contextually relevant" and "manipulative" gets thin fast in a conversational environment. If you ask an AI assistant "should I switch from Verizon to T-Mobile?" and a T-Mobile ad appears mid-conversation, is that helpful context or commercial interference? Koah's design choices, placing ads in breaks rather than inside the conversation, suggest they've thought about this. But at scale, the temptation to move ads closer to the response will be intense.

Koah plans to use the funding for engineering hiring and go-to-market expansion. The company is also introducing analytics tools to track user engagement and ad performance inside AI search and chat, alongside new ad formats built specifically for generative environments.

The bet is fundamentally about timing. AI apps are growing faster than any previous software category, but most haven't hit the monetization wall yet because they're still burning through venture funding. When that changes, and it will, the apps that can generate revenue from their existing user bases without degrading the product will survive. The ones that can't will either raise another round, sell, or shut down.

Tunguz's involvement is the strongest signal that this opportunity is real. He's not an AI tourist. He built the product that proved contextual web monetization could work at internet scale. If he sees the same structural opportunity in conversational AI, the question isn't whether AI apps will need advertising infrastructure. It's whether Koah will be the company that provides it.

For performance marketers, the practical takeaway is simpler. A new advertising channel is forming inside the AI tools your customers are already using three times a day. It's small now. It won't stay that way.

To understand why Koah matters, you need to understand the cost structure of the average consumer AI app. A single conversational query costs between $0.001 and $0.05 in inference, depending on model size and provider. A user who sends 10 queries per day generates $0.01 to $0.50 in daily costs. Multiply by millions of users on free tiers, and the burn rate becomes existential.

Most consumer AI apps charge $10 to $20 per month for premium tiers, converting somewhere between 2% and 8% of their user base. The remaining 92% to 98% are pure cost centers. Even modest ad revenue, say $0.02 to $0.10 per session, can flip the economics for free-tier users from loss-making to break-even.

Source: Industry estimates based on inference pricing from OpenAI, Anthropic, Google, 2026

Those numbers explain why Koah's phone is ringing. The math isn't theoretical. It's survival arithmetic for hundreds of AI startups that raised money on user growth promises and now need to show a path to revenue.

One thing the AI advertising ecosystem hasn't resolved is where contextual advertising ends and surveillance begins. Koah's "we match ads to the conversation, not the user" positioning is elegant but incomplete. Conversations reveal more about a person than their browsing history ever could. When someone tells an AI assistant they're struggling to afford their rent, or researching divorce lawyers, or comparing antidepressants, the context is intensely personal even if it's technically anonymous.

The regulatory framework for conversational AI advertising doesn't exist yet. GDPR and CCPA weren't written for chat interfaces. The FTC hasn't issued guidance. The IAB Tech Lab is still in the "proposing frameworks" stage. Koah is building in a regulatory vacuum, which is simultaneously an advantage (no compliance overhead yet) and a risk (the rules could change overnight).

For marketing ops managers evaluating AI advertising channels, the practical advice is straightforward: test now, but build your compliance infrastructure in parallel. The brands that treat AI ad placements with the same rigor they apply to programmatic display, transparency reports, brand safety lists, frequency caps, will be better positioned when regulation arrives.

The AI advertising gold rush is just starting. Koah is selling the shovels. Whether the gold is actually there, and whether the miners can extract it without poisoning the water supply, remains the open question.