⚡🤔 Sam Altman's Venture Math Problem

In partnership with

OpenAI released some usage data last week, and if you squint at it the right way, it tells a different story than the one everyone's been telling about AI.

The headline numbers: work usage of ChatGPT dropped from 40% to 28% over the past year, while personal use jumped from 60% to 72%. Writing tasks fell from 36% to 24% of conversations. Information seeking doubled from 14% to 24%.

You could read this as "enterprise AI adoption is struggling" (which is partly true), but I think the more interesting story is what's happening on the consumer side. ChatGPT's monthly retention hit 90%, higher than YouTube's 85%, which has been best-in-class for years.

Daily usage has also been aggressively climbing, averaging 29 minutes per person. For context, that's approaching Instagram territory (48 minutes) and well above Snapchat or Pinterest.

Those are remarkable numbers for any consumer product, let alone one that's been around for less than two years. The retention data suggests something shifted from "useful tool I sometimes remember to use" to "part of my daily routine." When you get to 90% monthly retention, you're not really in the software business anymore, you're competing with social media and entertainment for people's habitual attention.

The Enterprise Distraction

This creates an odd situation where OpenAI might be succeeding at something different than what they're trying to succeed at. The conventional AI narrative goes: AI transforms work → enterprises pay billions → winner takes all in B2B software. But the usage data suggests the real traction is happening elsewhere.

It's not that enterprise AI is failing entirely, there are pockets where it's working well. Coding assistance has found real product-market fit (though mostly through specialized tools like Cursor and Replit rather than ChatGPT directly). Customer support automation is delivering measurable ROI. But the broad "AI transforms all knowledge work" story seems to be hitting more friction than expected.

Meanwhile, people are spending half an hour a day asking AI about random topics, getting help with personal projects, using it as a research assistant for everything from dinner planning to understanding quantum physics. The shift toward information seeking (doubled to 24% of conversations) suggests people are treating ChatGPT less like Microsoft Word and more like Google with personality.

This puts OpenAI in an interesting strategic position. They've built what might be the stickiest consumer product since social media took off, but they're structured and funded like an enterprise software company trying to compete with Microsoft and Google in B2B markets.

The Google Precedent

There's actually a pretty clear historical precedent for this situation. Google spent their first decade building consumer products (Search, Gmail, Maps, Photos, YouTube, Android) before making a serious enterprise push. Each consumer product expanded their data collection and advertising inventory. Only after establishing dominance in consumer attention did they use those profits to subsidize expansion into enterprise with Cloud, Workspace, and other B2B offerings.

OpenAI seems positioned to execute a similar playbook, just compressed into internet time. They already have the consumer engagement (90% retention, 29 minutes daily). The recent hire of Instacart's former CEO as head of applications suggests they're thinking about expanding their consumer surface area - building ChatGPT equivalents of Maps, Photos, etc.

The monetization path is straightforward. Information seeking doubled to 24% of conversations, which maps well to intent-based advertising. When someone asks ChatGPT "what's the best laptop for video editing," you can capture that commercial intent. Keep users in-chat for recommendations, only send them out for purchases or when they need real-time information.

Interestingly, despite ChatGPT's growth, Google searches actually increased 21.64% year-over-year. Queries with 8+ words grew 7x, technical terminology usage surged 48%. Rather than displacing search, AI seems to be teaching people to ask more sophisticated questions. This suggests market expansion rather than zero-sum competition.

The Funding Math Problem

Which brings us to the tension at the heart of OpenAI's strategy. The consumer data suggests they've built something genuinely valuable and differentiated. But venture economics create pressure to be much more than a consumer AI company.

OpenAI hit $13 billion ARR faster than almost any company in history. But they still need to raise another $115 billion through 2029, on top of the $65 billion already raised. That's more than Uber ($25B), Tesla ($19B), WeWork ($14B), Amazon ($8B), Nvidia ($4B), Netflix ($3B), and Facebook ($2.3B) raised combined before going public.

To justify that level of investment, you likely need a story bigger than "really good consumer AI with great retention." Investors want to hear about transforming GDP, not about people chatting with AI for 29 minutes a day (even though that level of engagement should probably command a premium valuation).

So OpenAI is pursuing everything simultaneously: consumer apps, enterprise models, data centers, hardware devices, consulting services, chip development, becoming a cloud hyperscaler. They're trying to be Amazon, Google, Microsoft, and Facebook all at once.

This isn't necessarily bad strategy, Sam Altman is too thoughtful for that. But it does create interesting trade-offs. The optimal strategy (sequential consumer dominance like Google) conflicts with the necessary funding strategy (everything-company to justify unprecedented venture rounds).

Cross-Subsidization Warfare

There's a deeper strategic logic here that becomes clearer when you think about competitive dynamics. Consumer dominance could provide the cash flow to subsidize expansion into competitive enterprise markets.

Anthropic, for instance, can't really compete with ChatGPT on the consumer side - OpenAI has too much brand recognition and engagement momentum. So Anthropic focuses on B2B API business, particularly serving developers and AI coding companies. But if OpenAI uses consumer profits to underprice enterprise models, they could squeeze competitors out of those niches too.

This is basically Amazon's playbook: use retail profits to subsidize AWS until you dominate cloud infrastructure. OpenAI could use consumer subscription and advertising revenue to subsidize enterprise offerings until competitors can't match on price. Eventually you own both markets through cross-subsidization.

The timeline pressure makes this tricky, though. Amazon had a decade to build retail profits before making their big enterprise push. OpenAI is attempting to execute both strategies simultaneously while raising unprecedented amounts of capital.

What Could Break

The obvious risk is that Google and Microsoft decide to play the same game with even deeper pockets. If it becomes a pure subsidization war, infinite corporate cash flows beat venture funding every time.

But OpenAI has something their competitors don't: genuine consumer engagement with meaningful switching costs. When users maintain 90% retention and spend 29 minutes daily with your product, switching becomes psychologically difficult rather than just economically inconvenient. People develop habits, build up months of conversation history, get attached to the AI's personality quirks, and rely on its memory of their preferences and context.

This creates a very different competitive dynamic than enterprise AI markets. In coding, for instance, you have Anthropic's Claude, Google's models, and dozens of specialized tools like Cursor and Windsurf all competing intensely. Switching between coding assistants is relatively frictionless, it's mostly about which model performs better on your specific tasks today.

Consumer AI has much less direct competition and much higher switching costs. There's no real consumer equivalent to ChatGPT's combination of engagement, personality, and memory. Google's Gemini feels corporate, Anthropic's Claude lacks the consumer focus, and most other AI products are built for specific use cases rather than general conversation. Once someone builds up a relationship with ChatGPT, with all their conversation history, preferences, and behavioral patterns, recreating that somewhere else becomes genuinely difficult.

The question is whether they can expand their consumer ecosystem fast enough to build substantial advertising revenue before enterprise competitors entrench. Building secondary applications, an ad engine while simultaneously executing enterprise expansion and infrastructure plays is... ambitious.

How 433 Investors Unlocked 400X Return Potential

Institutional investors back startups to unlock outsized returns. Regular investors have to wait. But not anymore. Thanks to regulatory updates, some companies are doing things differently.

Take Revolut. In 2016, 433 regular people invested an average of $2,730. Today? They got a 400X buyout offer from the company, as Revolut’s valuation increased 89,900% in the same timeframe.

No wonder 10K+ everyday people are taking the chance on Pacaso.

Founded by a former Zillow exec, Pacaso’s co-ownership tech reshapes the $1.3T vacation home market. They’ve earned $110M+ in gross profit to date, including 41% YoY growth in 2024 alone. They even reserved the Nasdaq ticker PCSO.

The same institutional investors behind Uber, Venmo, and eBay backed Pacaso. And you can join them. But not for long. Pacaso’s investment opportunity ends September 18.

Invest While You Still Can

_{Paid advertisement for Pacaso’s Regulation A offering. Read the offering circular at invest.pacaso.com. Reserving a ticker symbol is not a guarantee that the company will go public. Listing on the NASDAQ is subject to approvals.}