Distributive Power

Does today's market leader possess a moat that ensures its defensibility in the face of evolving challenges and uncertainties in the AI landscape tomorrow?

Witnessing ChatGPT postulate on the definition of life in the form of a Shakespearean poem for the first time was an unforgettable experience. It echoed the sentiment of those memorable moments people share when they first engaged with their PC or first browsed on the internet. Such magical encounters over the past year propelled OpenAI's ChatGPT to 100M Users in just 2 months, the fastest ever consumer app to reach that milestone.

OpenAI has surpassed $1.3B as of October 2023, putting them up about 550% from an estimated $200M at the end of 2022 (Sacra)

OpenAI is the undisputed leader in the current AI race. Its core competitive advantage is its technology, namely the GPT Language Model (LM) series. Its GPT-Series (>= v3.5) is the gold standard against which every company, new model, and research advance compares itself. However, while OpenAI currently maintains a technological advantage over its closest competitors, the gap is rapidly shrinking.

“Great products are certainly a positive, but the cold and unromantic fact is that a good product with great distribution will almost always beat a great product with poor distribution."
Reid Hoffman, in his book "Blitzscaling"

Distribution Dilemma: The Key to Sustaining Leadership

Unless OpenAI achieves Super Intelligence or AGI, I believe its current technological advantage may reach a plateau, exposing the company to significant market disruptions. In such a scenario, the criticality of OpenAI’s distribution and go-to-market strategy becomes as, if not more, pivotal than its technology.

Madrona investors Vivek Ramaswami and Sabrina Wu (

Aspiring for Intelligence

), in their Predictions for AI in 2024 , predicted that “[in 2024] open-source language models are poised to eclipse closed-source counterparts [such as OpenAI].” They cite benefits like enhanced data and privacy, cost savings, faster iteration cycles, and customisation as reasons for the predicted surge in open-source models, additionally calling out that major tech companies, like Tesla and Uber, are expected to contribute more significantly. The potential emergence of open-source technology as a frontrunner, coupled with large companies making a more meaningful entry into the space, poses a threat to both revenue scaling and defensibility. This raises questions about the sustainability of OpenAI's current position.

Performance of Mistral’s latest Mixtral8x7B Model to the Llama 2 family and the GPT3.5 base model. Mixtral matches or outperforms Llama 2 70B, as well as GPT3.5, on most benchmarks.

OpenAI's Three-Pronged Approach to Distribution

With OpenAI’s technological advantage shrinking, potentially even trailing its Open-Sourced peers in the coming years, the question around distribution advantage becomes more apparent.

OpenAI’s current distribution strategy is effectively 3-part:

1. Consumer Subscription (ChatGPT)

   

   1. Risk: Depending solely on a consumer subscription model for revenue leaves OpenAI vulnerable to shifts in market preferences, economic downturns, or changes in user behavior. If the subscription model faces challenges, such as a plateau in user growth, increased competition, or difficulties in retaining existing subscribers, OpenAI may struggle to maintain the desired level of revenue.

2. Enterprise Licensing (OpenAI Enterprise API)

   

   1. Risk: The AI market is dynamic and competitive. Should open-source alternatives surpass the performance of the GPT series, enterprises stand to gain increased value at a reduced cost. This shift could impact OpenAI's Enterprise Licensing, posing challenges in scaling and maintaining revenue.

3. Enterprise Partnerships

   

   1. Risk: OpenAI's reliance on partnerships, such as the one with Microsoft, could be a potential vulnerability. If OpenAI's distribution strategy heavily depends on a limited number of key partnerships, any changes or challenges in those partnerships could negatively impact the company's reach and revenue.

Fragility of Market Leadership

And I think that OpenAI is acutely aware of the fragility of its current market leadership and the challenge it faces ahead, despite the large-scale success it is currently experiencing.

For one, OpenAIs GPTs symbolise a meaningful push towards User Engagement and Ecosystem Ownership. By fostering a developer community, encouraging 3rd Party Applications, and creating a platform akin to an app store, OpenAI has focused on creating a more extensive ecosystem around its models. This ensures that users not only utilise its technology but also engage with a broader suite of applications and services, powered by its technology.

Additionally, OpenAI’s recent partnership with Humane suggests a move towards a more vertically integrated approach. Evaluating the collaboration, Freda Duan (Principal at Altimeter) discussed that such a partnership has deeper implications, than just a simple grab for more screen time. She observes that “OpenAI can't be an app store without being an OS itself.” That is, “LLMs have to create their hardware. Due to conflicts of interest, giants like Apple will never integrate external AI systems like ChatGPT (esp. GPT Stores) into their ecosystems.” This underscores the importance of adopting a comprehensive approach, particularly for technology-first companies such as OpenAI, who need to take ownership of the entire consumer experience stack. This involves not only providing underlying plug-and-play technology but also owning both the hardware and software components.

An Uncertain Future

In a recent chat with Elad Gil, Brad Gerstner (Chairman and CEO at Altimeter) discussed how this current situation is analagous to the canonical example of the internet: “what is going to have durable value capture? Am I getting head faked? Is this Alta Vista? Is this Lycos? Is this AskJeeves? Or is this Google? Right? And why did Google ultimately capture all the value? Because everybody went to Google, right? They owned the top of the funnel. They became the place. Not only did they own search, they were so crafty. They said, you know, this address bar that people used to actually type in URLs, we'll just turn that into a search box so that anything you type in there becomes a search too.” Analogously, we may ask how OpenAI will be able to capture the top of the funnel better than its competitors, through its own craft, to ensure its viability and vitality as a business in the long-term?

Whilst a Microsoft acquisition remains a continuous possibility, the rapid development in the field and the touted entry of significant players like Apple leave the future of OpenAI intriguing and uncertain. The question lingers: Where does OpenAI go from here? The answer may lie in a holistic strategy that encompasses both technological innovation and strategic positioning in this ever-evolving landscape.

Thanks for Reading,

Ary