When OpenAI announced plans to test advertising in ChatGPT, the headlines wrote themselves. The AI darling was finally adopting Google’s playbook. Another startup forced to monetize through ads because nothing else scales.
Except that’s exactly backward. Both companies are racing away from advertising dependence as fast as they can, and the revenue strategies they’re pursuing reveal a far more interesting future than banner ads in chatbots. We’re watching two fundamentally different visions of how AI companies make money collide in real time, and the winner will define not just these companies but the entire AI economy for the next decade.
Here are five revenue strategies beyond advertising that are actually moving the needle in 2026, and what they reveal about where this industry is headed.
1. Tiered Subscriptions Are Becoming the New Cable Bundles
OpenAI just introduced a fascinating middle ground that exposes how subscription economics are evolving in the AI era.
ChatGPT Go launched at $8 per month, slotted strategically between the free version and the premium Plus tier at $20 monthly and Pro at $200 monthly. This isn’t random pricing experimentation. It’s sophisticated market segmentation designed to capture everyone from casual users to power professionals. The strategy mirrors how streaming services evolved, but with a crucial difference: the product is fundamentally the same across tiers, just throttled or enhanced by usage limits and model access. Google is taking the opposite approach, integrating Gemini features within existing Google One and Workspace plans rather than creating standalone AI subscriptions. The philosophical divide is revealing. OpenAI believes people will pay specifically for AI intelligence. Google believes AI should enhance products people already pay for. One creates a new market. The other expands an existing one.
2. The Enterprise Play Is About Embedding, Not Selling
Consumer subscriptions grab headlines, but the real money is in making AI infrastructure invisible to business users.
OpenAI signed a three-year deal with ServiceNow to integrate its models directly into ServiceNow’s business software. This is brilliant strategic positioning. Instead of asking enterprises to adopt a new tool, OpenAI becomes the intelligence layer inside software businesses already depend on. The company is developing bespoke AI products specifically tailored for governments and large-scale enterprises, targeting customers who will pay premium prices for customization and security. Google is seeing accelerated growth in Google Cloud Platform driven by a strong backlog of enterprise AI projects, leveraging existing customer relationships. The winner here won’t be who builds the best standalone AI product. It’ll be who gets embedded so deeply into enterprise workflows that removing them becomes unthinkable.
3. Custom Chips Are the Ultimate Competitive Moat
Here’s a development that sounds technical but has massive strategic implications: both companies are investing heavily in proprietary chip development.
OpenAI is developing its own custom inference chips currently being taped out to optimize internal costs and potentially create new hardware-related value streams. Meanwhile, Google is exploring the sale of its proprietary TPUs to a market facing limited supply of high-end AI chips. Think about what this means. Whoever controls the silicon controls the cost structure, the performance ceiling, and potentially an entirely new revenue stream selling infrastructure to competitors. Google already secured a partnership with Anthropic to use Google Cloud TPU chips to train Claude models, utilizing over 1GW of capacity. That’s not just revenue. It’s competitors becoming dependent on your infrastructure. The long game here is obvious: in a world where AI models become commoditized, the companies making money will be those selling the picks and shovels.
4. Commerce Fees Are the Quiet Revolution
Everyone obsesses over chatbot subscriptions while missing the more lucrative opportunity: taking a cut of transactions that AI facilitates.
OpenAI is exploring affiliate commerce fees through features like Instant Checkout, charging merchants a fee for purchases made via its AI. This is potentially transformative. If AI assistants genuinely change how people shop, whoever controls that interface controls access to consumer intent at the moment of highest value. Google has a massive head start here, with deep integration of its Shopping Graph containing 50 billion products within Search. But the dynamics are different. Google sends traffic to merchant sites. OpenAI’s model suggests completing transactions without users ever leaving the chat interface. The question is whether merchants will accept giving OpenAI a percentage of sales in exchange for convenience, or whether they’ll resist creating a new middleman. Early signs suggest merchants might not have a choice if consumer behavior shifts enough.
5. Infrastructure Megaprojects Are Becoming Business Strategies
The Stargate project represents something unprecedented: a $500 billion infrastructure initiative that’s simultaneously a technical necessity and a potential revenue source.
This massive investment in sovereign data centers and compute resourcing isn’t just about serving OpenAI’s own needs. It’s about creating capacity that can be monetized through APIs, enterprise partnerships, and potentially direct infrastructure sales. OpenAI provides intelligence through APIs, a model designed to scale revenue in direct proportion to the value of intelligence delivered. The more valuable AI becomes, the more they can charge per API call. Google is playing a similar game, securing multi-year partnerships like the deal to have Gemini power Siri and Apple Intelligence features. These aren’t just distribution deals. They’re recurring revenue streams measured in billions, locked in for years, independent of advertising fluctuations.
The Uncomfortable Economics
The advertising announcements distract from a more fundamental truth: both companies are building business models that look nothing like the internet giants that preceded them.
OpenAI is targeting hundreds of billions in revenue by 2030. Not millions. Not billions. Hundreds of billions. That doesn’t happen through subscriptions alone, even at $200 per month. It requires enterprise deals measured in eight figures, infrastructure sales at massive scale, and transaction fees on commerce that runs through AI interfaces. Google is leveraging established advantages in cloud computing, hardware manufacturing, and platform distribution to create revenue streams that compound rather than compete.
The real divergence isn’t Google’s advertising dominance versus OpenAI’s subscription focus. It’s two completely different theories about how AI companies capture value. One believes intelligence itself is the product, priced by consumption. The other believes intelligence enhances existing products and platforms, creating value through integration rather than isolation.
Which model wins? Probably both, serving different markets and customer needs. But the companies that think AI monetization is about choosing between ads and subscriptions are missing the actual game being played. The real money is in infrastructure control, enterprise embedding, transaction facilitation, and platform dominance.
The question for 2026 isn’t whether Google or OpenAI will make more from advertising. It’s which company will control more of the value chain when AI becomes the operating system for how we work, shop, and access information. And on that question, the scoreboard looks very different than the headlines suggest.
– Manpreet Jassal
Leave a Reply