Insiders Reveal the Truth About the AI Race

Insights from a recent gathering of AI industry leaders reveal a fascinating shift in sentiment and strategic focus. At a prominent San Francisco conference, an anonymous survey of more than 300 attendees, primarily founders, investors, and engineers, provided a snapshot of expert predictions and prevailing attitudes. The results highlight both ambitious financial forecasts and a notable cooling of enthusiasm for certain high-profile players.

Survey participants projected that OpenAI will reach $30 billion in annualized revenue by the end of 2026, a figure that suggests growth will continue but may not maintain its previous explosive pace. In a striking prediction, the median response estimated Nvidia’s valuation will hit $6 trillion within the same timeframe, underscoring its perceived dominance in supplying the computational backbone for AI development.

When asked about the arrival of artificial general intelligence (AGI), the most common answer was the year 2030. This timeline indicates that many insiders believe meaningful progress toward human-level machine intelligence remains several years away. The selection process for the independent expert committee tasked with making this declaration, as outlined in the Microsoft-OpenAI agreement, remains unclear and was noted as a point of uncertainty.

In terms of investment appeal, venture firm Andreessen Horowitz was identified as having the most enviable AI portfolio, followed by Khosla Ventures and Sequoia. If given the opportunity to invest in private tech companies today, attendees favored Anthropic first, then OpenAI, Cursor, Anduril, SpaceX, and OpenEvidence. For the model expected to lead the LMArena leaderboard in web development by the end of 2026, the order of preference was OpenAI, Anthropic, Gemini, Grok, and Qwen.

The survey also highlighted some surprising targets for skepticism. Perplexity was the top choice for a startup that attendees would short if given the chance, with OpenAI surprisingly appearing in second place. Other names on that list included Cursor, Figure, and Mistral.

A clear theme emerged from these results: a noticeable softening toward OpenAI. While the company is still expected to perform well, the fact that more respondents expressed a desire to own Anthropic stock points to a shifting competitive landscape. Equally telling was the absence of Meta from the LMArena leaderboard predictions. The inclusion of Alibaba’s Qwen model signals a growing global influence, with many companies now opting to fine-tune open-source Chinese models instead of Meta’s Llama.

Beyond the survey, conference discussions revealed other critical industry dynamics. A session on “reverse acquihires”, where large tech companies acquire startups primarily for their talent, highlighted how antitrust scrutiny and the fierce competition for a limited talent pool are driving these deals. One participant described big tech as having “infinite money” to place bets on the best minds, with corporate development teams sometimes asking founders to name their price for an acquisition.

Perhaps the most significant cultural shift noted was the declining emphasis on AGI. Whereas the first iteration of this conference was saturated with existential debates about artificial general intelligence, this year’s conversations were far more pragmatic. The focus has moved squarely toward business applications, market share, and building products that customers are willing to pay for. A undercurrent of concern about an AI bubble was present, but the dominant mood was one of commercial determination.

Several onstage interviews yielded memorable quotes that capture the current mindset. Replit CEO Amjad Masad observed, “If you are competing on price, then maybe you don’t have a business.” Investor Elad Gill advised that “Most companies should sell at some point. There’s often a market-maximizing moment where you’re going to get the best deal you can.”

San Francisco Mayor Daniel Lurie welcomed a particular urban problem, stating, “People are starting to complain about traffic. Thank goodness. I want those complaints. We still have a lot of empty office space.”

Anthropic’s Chief Product Officer Mike Krieger offered a product development insight, noting, “Time spent is, I can tell you, not on any of the dashboards that I look at. It’s just not a main consideration.” Meanwhile, xAI co-founder Jimmy Ba provided a philosophical perspective, remarking, “Knowledge is just crystalized computation from the past.”

Collectively, these insights paint a picture of an industry maturing rapidly. The initial wave of speculative excitement is giving way to a more grounded phase focused on commercialization, talent wars, and measurable business outcomes.

(Source: The Verge)