This is an excerpt from Alex Heath’s Sources, a newsletter about AI and the tech industry that is syndicated to The Verge subscribers only once a week.
I spent yesterday in San Francisco at Eric Newcomer’s Cerebral Valley conference, now in its third year. I have attended this event for three years in a row because Eric does a great job of organizing the speakers and audience, and the talks are more meaningful than a typical industry event.
This year was no exception; However, I found the most interesting part of the day to be when the results of an anonymous audience survey were shared on stage. The survey included more than 300 participants, primarily AI company founders, followed by investors, other industry professionals (including product leaders and engineers), and members of the media.
Here are the survey results as they were shared on the platform:
1. What will OpenAI’s annual revenue be at the end of 2026?
Average answer: $30 billion.
2. What will Nvidia be worth at the end of 2026?
Average answer: $6 trillion.
3. In which year will an independent committee of experts declare that we have reached AGI, as per the Microsoft-OpenAI agreement?
4. Which venture capital firm’s AI portfolio are you most jealous of?
The top three, from first to last, received the most votes: Andreessen Horowitz, Khosla Ventures and Sequoia.
5. If you could invest money in any private technology company today, which would it be?
The top five companies, from first to last: Anthropic, OpenAI, Cursor, Anduril, SpaceX, and OpenEvidence.
6. Which global company model will top the LMArena web development leaderboard at the end of 2026?
In order from first to last: OpenAI, Anthropic, Gemini, Grok, Quen.
7. If you could short a startup with a valuation over $1 billion, which one would it be?
First place was Perplexity. OpenAI got the second place. Other names shown on stage: Cursor, Figure, Harvey, Mercor, Mistral and Thinking Machine.
What I learned from these results (newbie has published slides for his paying customers):
- Softening on OpenAI: Given that Sam Altman has said that OpenAI plans to end this year with annual revenues of $20 billion, this group of AI insiders do not expect next year to be as rapid for the business as the leap from 2024 to 2025. The prediction that AGI will not be declared until 2030 suggests a lack of confidence in meaningful improvement in model progress in the near term, although this answer may also be blurred by the complexity of how OpenAI and Microsoft should compromise. It has been decided. (I’m still waiting for either company to share information about who its “independent committee of experts” will be and how they’ll make decisions.) It was also notable that more attendees wanted to buy Anthropic stock than OpenAI, despite the consensus that OpenAI would lead LMArena next year.
- Meta was not in the conversation. It was not named in the list of models likely to lead the LMArena next year. The presence of a Chinese model (Alibaba’s Quen) in the top five signals a shift that is already underway, as many companies prefer the open-source Chinese model rather than Llama. Meta has a lot to prove if it wants to re-enter the model race.
- Confusion is controversial. But everyone working in AI already knows this.
Other findings from Cerebral Valley:
What’s driving? reverse acquisition, I attended a breakout session about AI acquisitions, like Meta’s deal with ScaleAI to hire Alexander Wang and Google’s deal with Character and Windsurf. I’ve covered many of these deals closely over the years, but it was interesting to hear the group’s perspective on what motivates them. Antitrust investigations of Big Tech certainly play a factor, but some involved in these types of transactions have also said that larger companies are competing with each other to attract talent and grow faster than their competitors. As one member of the group put it, they appear to have “unlimited money”, and view it as a game of betting on a very limited pool of talent. One AI founder in the group that presented several such proposals recalled a member of a corporate development team at a Big Tech company asking Him He wanted to know how much importance his startup should be given for a deal.
No one cares about AGI anymore. At the first Cerebral Valley conference, the topic of AGI was a major topic. One startup founder on the forum said that “we will be dead” by the time OpenAI releases GPT-10. This year, many conversations on the platform saw how AGI barely registered as a topic of discussion. Instead, most of the interviews focused on business applications of AI. Many of the companies previously represented on stage at the Cerebral Valley event did not exist and are now worth billions of dollars. There was fear of an AI bubble all day long, but most people were focusing on how they can win market share and provide products people want to pay for.
Standout quotes from interviews on stage:
- Replit CEO Amjad Massad: “If you’re competing on price, you probably don’t have a business.”
- Elad Gil: “Most companies should sell at some point. There is often a market-max moment where you can get the best deal. A very small number of companies should never sell.”
- San Francisco Mayor Daniel Lurie: “People are starting to complain about traffic. Thank God. I want those complaints. We still have a lot of empty office space.”
- Anthropic CPO Mike Krieger: “I can tell you that time spent is not on any dashboard that I look at. It’s just not a core consideration.”
- Jimmy Bae, co-founder of XAI: “Knowledge is simply crystallized calculations from the past.”