Everything Apoorv Agrawal said on any show that made the record, most notable first. Each card names its show and opens the statement there.
Agrawal: Humans Will Not Need to Work to Earn a Living by 2035
“I suspect in five to 10 years, we wouldn't have to work to earn a living anymore.”
Agrawal: Enterprises must plan for multi-model routing over single-lab lock-in
“So if you're a CIO, if you're a CEO, if you're about to make a multi-million, billion-dollar decision of which lab to build your intelligence on, plan for the climate, not for the weather. What does that mean? That's multi-model routing. That's evals. That's t…”
Agrawal: ChatGPT's User Base Is 10x Larger Than All Rivals Combined
“If you took the user count of ChatGPT on one side,
And all the other AI apps on the other side, let's do perplexity, and Claude, and Croc, and Gemini, and you know, long tail of apps that people use. Add them all up, multiplied by 10, it's probably still less …”
Agrawal: OpenAI's Consumer Business Could Reach $200B in Annual Revenue
“We think over time,
The number of users could go up to three to four billion users. We think on the monetization side, over time, it could go from 10 today to 60, 70 dollars per user per year, which is where Meta and Google and other large consumer platforms a…”
Agrawal: Fast shipping velocity is OpenAI's primary defensibility
“Keeping up with OpenAI product updates is a full-time job, and that's the strategy, is ship fast. And I feel like that done at a consistent pace, I mean, just in, just this year alone, you've had Operator and Deep Research and ChatGPT agents and Codex, and the…”
Agrawal: AI Semis Generate $200B, While AI Apps Are Only Earning $40B
“Nvidia earned, Nvidia alone at the chip player earned roughly forty billion dollars in data center revenue in Q one last quarter. So analyze that's about a 160. Let's say the total industry is 170, 180, maybe let's say 200. The infra layer is 20, thirty billio…”
Agrawal: Frontier AI Training Data Is Saturated, Shifting the War to Talent
“The ingredients of call it being on the frontier are compute. CapEx, we spoke about it. Data which is sort of access accessible to everybody. We are saturated on, on, on the training data that exists. Talent. This is where the war is now.”
Agrawal: ChatGPT Daily Time Spent Exceeds Mainstream Social Apps Like X
“On, in terms of time spent, you know, we've published this analysis. They are far and above time spent per user per day compared to any of the other apps AI apps today. Actually not just AI apps. I, you know, one of the analysis we did is they are now larger t…”
Agrawal: ChatGPT exhibits rare retention 'smile curve' like TikTok and Instagram
“And ChatGPT exhibits a smile curve. The other apps that exhibit smile curves are Instagram, and TikTok.”
Agrawal: OpenAI is anti-fragile and has become faster over time
“The single word to define them is anti-fragile. I've seen this for years now. Remember the time that the blip happened when Sam was for a hot second, not at OpenAI, and it felt like, oh, what's going to happen to this org? You know, this is obviously a pretty …”
Agrawal: Switching to OpenAI's New Model Boosted XBOW's Cyber Detection to 81%
“Their benchmark, on their benchmarks, they saw a significant step up with GPT-V, much better than any other coding or design or math startup saw in their ability to find cyber exploits. And, you know, summarizing, basically, the jump was from just under 60% to…”
Agrawal: XBOW's AI agents ranked number one globally on HackerOne
“Expo has been let loose on this platform and couple of weeks in, they became the number one player in the United States. But two weeks ago right at Black Hat, they achieved the number one spot globally”
Agrawal: AI Cybersecurity Startup XBOW Spends More on Tokens Than Human Salaries
“And it is true. We spend more on tokens than we spend on humans. And it is, I guess, the shape of an AI native form.”
Agrawal: If Your AI Startup Has a Polished Roadmap, You're Moving Too Slowly
“Things are moving so fast that if you have a beautiful product roadmap, something's not right. You're probably not moving fast enough.”
Agrawal: Decagon routes roughly 90% of its production through open-source models
“As Jesse from Decagon wrote very eloquently a big portion of their volume goes through open source, about 90% now.”
Agrawal: Global software industry generates roughly $400 billion in annual revenue
“This, and just to put this in context, the entirety of the software industry earns about four hundred billion dollars of revenue.”
Altimeter Invested in OpenAI at $150B Valuation
“We entered last year at the hundred and fifty billion dollar round.”
Agrawal: Just 0.3% of Startups Generate 90% of Venture Gross Profits
“The rough math of our business is there's 5000 companies that raise every year that we track. 500 of those, so 10% of those, will raise from tier one investors, 15 of the 5000 will return over 90% of the gross profits for a given vintage.”
Agrawal: OpenAI is Altimeter's largest investment in firm history
“OpenAI is our largest investment in the history of Altimeter.”
Agrawal: Anthropic Is Beating OpenAI in Enterprise Developer and Coding Markets
“Anthropix crushing it with their API, probably one of the most loved developer products out there, both with the API, but also the Cloud Code product. And so I would say they were winning in their own right in the enterprise segment with developers on the codi…”
Agrawal: The best of Google is ahead of them in AI
“I feel like the best of Google is ahead of them.”
Agrawal: ChatGPT is a personal relationship, not just a product
“ChatGPT is no longer a website. It is no longer a product. It is a relationship that users have with technology similar to the way the relationship that I have with this phone.”
Agrawal: ChatGPT's memory feature creates immense switching costs
“Memory in particular, I think adds so much context for your life that switching out of it would be so hard.”
Agrawal: Public AI benchmarks are saturated; proprietary evals are required
“The benchmarks are the starting line, but they're by no means the finish line. The benchmarks are kind of saturated, right?... What you actually need is more proprietary evals.”