I Think OpenAI Is Undervalued

The market's sentiment toward OpenAI feels a little strange lately.

On one hand, people still recognize it as one of the defining icons of the AI era. On the other hand, the discussion around it has started to carry a clear sense of pessimism: Anthropic has overtaken OpenAI in B2B, Google has full-stack advantages, and OpenAI seems to be falling behind in B2B while being chased in consumer.

I think the market has overreacted.

OpenAI is still one of the few companies that simultaneously has model capability, compute, talent, distribution, and product imagination. That combination is extremely rare.

The market tends to adjust narratives quickly based on short-term benchmarks, product release cadence, and competitor momentum. But maybe the better questions are:

Does this organization still have the ability to keep training frontier models?

Does it still have enough talent density?

Can it quickly turn research into product?

Does it have distribution that keeps users inside the product?

Across these dimensions, OpenAI remains very strong. GPT-5.5 shows that OpenAI is still at the frontier in reasoning, agentic coding, and long-horizon workflows; ChatGPT Images 2.0 also shows that it remains strong in multimodality.

The Market May Be Overestimating B2B First-Mover Advantage

Another common narrative is that Anthropic has already captured a large share of the B2B market, and OpenAI's lag is enormous.

I think this judgment is wrong.

The B2B AI market is not a fixed market where the cake has already been divided. It is an absolute growth market that is only just beginning. Many companies are still in the early adoption phase: buying seats, running pilots, connecting APIs, testing internal agents, exploring coding workflows, customer service, financial analysis, data cleaning, and so on.

This is still far away from true enterprise transformation.

The B2B market is just getting started, and the landscape is only beginning to take shape.

B2B is also not a winner-take-all market. Different companies, departments, and workflows will have different model preferences, procurement logic, and integration paths.

As long as OpenAI's core model and compute capabilities do not have fundamental problems, it still has enough HP and counterattack potential.

Moreover, OpenAI's PE-related enterprise deployment venture is a very interesting distribution move: it has the chance to enter a large number of mid-sized companies through PE portfolio companies. Execution will still be difficult, but this step does feel like timing, positioning, and resources coming together.

The Integration of ChatGPT and Codex

I am especially bullish on OpenAI's recent integration of ChatGPT and Codex.

This is a unification of business logic and infrastructure logic.

The future product form of AI is a unified intelligent workspace: a place where you can write code, research, edit documents, debug systems, call tools, execute tasks, and manage context.

ChatGPT has massive consumer distribution. Coding and agentic backend capabilities are essential infrastructure for supporting the next generation of interaction.

If these two pieces are integrated well, OpenAI has the chance to further elevate ChatGPT's capabilities and push it into a deeper, higher-ARPU productivity center.

From the outside, OpenAI is quickly converging its product experience around ChatGPT, Codex, and agentic workflows. That speed of integration itself is worth paying attention to.

If this were happening inside a much larger company, the politics and internal division of resources alone could take a long time to resolve. Large companies often suffer from fragmented ownership, complex product lines, and organizational structures that make integration painful.

As a pure-play AI company, OpenAI has a chance to respond to the market with much greater speed.

High Volatility

Of course, any investment thesis has to face the bear case. OpenAI's uncertainty is also very real.

OpenAI as a company contains two different temperaments at once:

A-side: very sharp, very ambitious, extremely good at capturing historical opportunities.

B-side: organizational turbulence, ambiguous governance, and high volatility in commercialization path and public messaging.

That volatility has discounted the price.

But I think this uncertainty has already been overly priced into the market narrative. OpenAI's chaos is real, but so is its alpha. Many great tech giants in their early stages were not stable, perfectly orderly machines. They were complex systems with high energy, high conflict, and high speed.

So I think the market has been too quick to re-rate OpenAI from AAA to BBB. The cards in OpenAI's hand are still very solid. In one month, we will also have a chance to test the flexibility of its organizational structure. Let's see.