Navigating the AI Frontier: VCs Debate the Bubble, Defensibility, and the Rise of Los Angeles

In a candid, high-level discussion hosted by TechCrunch’s StrictlyVC in El Segundo, California, two of the venture capital industry’s most prominent voices on artificial intelligence took the stage to dissect the current state of the market. Carter Reum, co-founder of M13, and Chang Xu, a partner at Basis Set Ventures, provided a masterclass in modern investing. Between them, they represent nearly $3.5 billion in assets under management and a front-row seat to the most explosive technological cycle in modern history.

The conversation, held in a sun-drenched venue just miles from the aerospace corridors that define much of L.A.’s tech identity, moved beyond the usual industry platitudes. Instead, it tackled the existential questions currently haunting boardrooms: Is the AI infrastructure bubble about to burst, how can startups survive the "hyperscaler" onslaught, and what will the impending liquidity of a SpaceX IPO mean for the future of the Southern California tech ecosystem?

The Paradox of Growth: Is AI in a Bubble?

The central tension of the evening revolved around valuation. When companies like ChatGPT achieve $40 billion in revenue in record time, the traditional metrics of venture capital seem to disintegrate.

Defining the Bubble

"There’s both a bubble and not a bubble," Chang Xu argued. From her perspective, the "not a bubble" case is rooted in unprecedented growth curves. She pointed to a portfolio company, OpenArt, which scaled from $1 million to $70 million in annual recurring revenue (ARR) in just two years while remaining cash-flow positive. "The bar for what is good growth has totally changed," she noted. When compounding growth is this rapid, the high price tags attached to early-stage deals are simply a reflection of the projected terminal value.

However, Xu offered a stern caveat: if investors apply these hyper-growth assumptions to every single deal, the math will inevitably fail for the broader portfolio.

Historical Parallels

Carter Reum took a broader, historical view, suggesting that the current hysteria is not as unique as founders might like to believe. "We’ve seen this before—with the cloud, the iPhone, and even the automobile in the 1920s," Reum observed. "People feared for their jobs then, too, yet life went on."

What makes this cycle distinct, according to Reum, is the intensity of the competition. In previous tech eras, startups competed against other startups. Today, they are locked in a cage match against the largest, most well-funded incumbents in history—companies like Google, Microsoft, and Amazon that possess a rare trifecta: unlimited capital, proprietary data, and elite talent. "For the first time in history, the incumbents actually have the advantage," Reum said, warning that rapid rise could just as easily lead to a rapid fall.

Strategy and Defensibility: Investing Above and Below the AI

With hyperscalers like OpenAI and Anthropic constantly expanding their feature sets, investors are struggling to identify which startups have true "moats."

The "Friction as a Moat" Thesis

Reum emphasized that the most dangerous place to build is in the obvious, low-friction markets like marketing or generic software. Instead, M13 focuses on "friction as a moat." He pointed to a successful exit in a company disrupting 911 call centers—a sector where regulation, complexity, and life-or-death reliability create natural barriers to entry. "The hyperscalers might go there eventually," Reum noted, "but as a few-billion-dollar outcome, they aren’t going there anytime soon."

The Infrastructure Layer

Xu’s firm, Basis Set Ventures, approaches the problem by looking at the technical stack. She divides the world into "below the AI" and "above the AI."

• Below the AI: The infrastructure layer (databases, version control, deployment) was built for human engineers. Now that AI agents are performing the work, that infrastructure must be rethought. "Last year, I would never have thought you’d need a new GitHub," she said. "This year, I can count on two hands how many teams are going after being the GitHub for agents."
• Above the AI: Here, the focus must remain on long-term differentiation. In "velocity markets," where fast followers can easily copy a product, speed is everything. In "depth markets"—like one of her portfolio companies using transgenic chickens to manufacture drugs—the complexity of the task provides the defense.

The Evolution of the Startup Ecosystem

The discussion then shifted to the nature of innovation, moving from the "obvious" first wave to the "hidden" second and third waves of the AI boom.

The Rise of "Bad Ideas"

"The story of VC is that it’s constantly a story of bad ideas becoming good again," Xu reflected. She pointed to Cursor, a coding assistant that many dismissed as a mere "AI wrapper." Today, it is a $60 billion success story. She noted that researchers, once relegated to the poverty-level wages of academia, are now the most influential figures in the industry, dictating the direction of the market via social media.

The "Ripples" Theory

Reum likened the current tech cycle to a rock thrown into a pond. "The first wave is always the most obvious and crowded," he explained. "I get excited about two, three, four years from now. There are going to be business models we can’t even imagine today." These second and third-wave bets are difficult to execute, but they offer the most attractive ROI because they exist outside the glare of the mainstream spotlight.

Los Angeles: The Future of "Taste" in Tech

Perhaps the most significant takeaway from the evening was the bullish outlook for the Los Angeles technology sector, particularly in light of the impending SpaceX IPO.

A New Class of Wealth

Reum noted that when SpaceX eventually goes public, it will trigger a massive transfer of wealth into the L.A. basin. Unlike the liquidity events of the past, this will be widely distributed among employees, creating a surge of capital and entrepreneurial energy. "Every major liquidity event generates a second wave," Reum said. "The previous L.A. cycle produced Riot Games, Tinder, and Snap. This is a different order of magnitude."

The "Taste" Advantage

While San Francisco remains the global hub for technical, engineering-heavy innovation, both panelists agreed that L.A. is uniquely positioned for the next phase of the AI revolution.

"The next frontier in AI isn’t more compute—it’s taste," Xu asserted. As models become increasingly adept at writing code and optimizing data, the value shifts toward the ability to connect with human emotion and culture. "San Francisco has extraordinary technical talent, and that is exactly what the models are getting good at automating. L.A. has taste in spades."

Implications for Founders and Investors

The session concluded with advice for founders navigating this volatile landscape: "You need a microscope in one eye and a telescope in the other," Reum advised. The microscope is for the brutal, day-to-day execution required to survive the current quarter. The telescope is for the long-term, structural shifts that will determine if a company survives the next three years.

For investors, the message was clear: The "easy" money has been deployed. The next phase of the AI boom will reward those who can distinguish between fleeting fads and fundamental shifts in how machines interact with the real world. Whether it is through the lens of regulated healthcare, agent-based infrastructure, or the creative, culture-first output of Los Angeles, the industry is moving into a more mature, more complex, and potentially more lucrative chapter.

For further insights, you can watch the full conversation from StrictlyVC LA 2026 via the embedded video above.