OpenAI’s first sales leader, Aliisa Rosenthal, has found a new career: venture capital. He was involved Capital Acres as a general partner, working with founding partner Lauren Kolodny and other corporate partners, Rosenthal and Kolodny told TechCrunch.
Rosenthal left OpenAI about eight months ago after a three-year sprint at the AI lab that saw the launch of DALL·E, ChatGPT, ChatGPT Enterprise, Sora, and other products. “I don’t want to join a VC fund,” he told TechCrunch. “I’m out there meeting with a lot of AI startups.”
But after growing the company’s OpenAI sales team from two people to hundreds, they saw the appeal when Kolodny sold them on venture capital. Instead of helping a single startup with a go-to-market strategy, they can help their portfolio.
During her time at OpenAI, “I learned a lot about behavior, both on the part of buyers, how people think about these purchases, and the gap between what most organizations think is possible and what can actually be deployed today,” she said.
For example, he has first-hand insight into the kind of moat that AI startups can build that won’t be vulnerable when model makers like OpenAI launch competing products.
Is OpenAI “just building everything and putting all the companies out of business? You know, they’ve done so much: they’re in the consumer, they’re in the enterprise, they’re building devices. I don’t think they’re going to go after every potential enterprise application,” he said.
So one moat for AI startups offers specialization.
Techcrunch event
San Francisco
|
13-15 October 2026
Context is moat
Additionally, he thinks the key to a good startup moat is “context” — or information that AI stores in context window memory as it works on requests.
“Context is dynamic. It can adapt. It can be scaled. And I think what we’re seeing is going to extend the basic RAG to this idea of a context graph, which is continuous,” he said referring to Retrieval-Augmented Generation (RAG) de facto method from 2025 to minimize hallucinations by training LLMs on reliable and specific sources (and having LLMs mention them).
There is still a lot of technology to be developed for this area, from memory to reasoning beyond pattern recognition.
“I expect real innovation here. I think this year we will see a new approach – the idea of context and memory,” said Rosenthal.
But beyond startups working directly on context engineering, Rosenthal thinks the company’s applications that create it will reap the benefits.
“Ultimately, when we talk about the trenches, I think having and managing this layer of context is going to be a big advantage for AI products,” he said.
Another opportunity he sees: startups don’t build on top of the most advanced models in the lab, with high prices.
“I think there is room in the market for cheaper models that are lighter in weight and innovation in terms of inference costs,” he said. This is a model that is not, perhaps, at the top of the leaderboards of various benchmarks but “still very useful” and more affordable.
“Where I am very excited to invest in the application layer. I am really interested in what will be durable applications built in all different models, not only in the basic model,” she said. They’re looking for startups with “interesting use cases” or those that use AI to help their company’s employees work more efficiently.
Where he will find these startups, he will work to network among OpenAI alumni for startups. Now that the AI outfit is 10 years old, the alum network has grown. Many have founded startups that have raised huge amounts of money at high valuations, ranging from OpenAI’s biggest competitor, Anthropic, to busy early-stage companies like Safe Superintelligence.
There is also ample precedent for high-level ex-OpenAI people to become seed-stage investors. About a year ago, Peter Deng, former head of OpenAI consumer products joined Felicis. He’s been crushing it ever since, and is clearly having fun, getting big deals for hot startups like that LMA and Periodic lab.
“I actually had a phone call with Peter a few months ago, and he helped me make the decision,” Rosenthal said of his choice to become an investor.
But Rosenthal may have a secret weapon for winning bids. They also have deep contacts among the users of AI companies – the kind of buyers and beta testers that these AI startups need.
Companies still don’t know how much AI can do. “There’s a huge gap that I’m optimistic can be filled. It leaves a huge green field for applications and companies.”