The AI Frontier: Navigating the Paradox of Unprecedented Growth and Market Volatility

In the sun-drenched corridors of El Segundo, California, the latest installment of TechCrunch’s StrictlyVC event served as a high-stakes town hall for the venture capital elite. As the artificial intelligence gold rush matures from its experimental infancy into a trillion-dollar infrastructure play, the narrative surrounding the sector has shifted from "Will it work?" to "How do we survive the fallout?"

Two of the industry’s most perceptive voices—Carter Reum, co-founder of M13, and Chang Xu, a partner at Basis Set Ventures—took the stage to dissect the current state of the AI market. With M13 managing $2.5 billion and Basis Set Ventures operating as a specialized AI-native fund with nearly $1 billion in assets, both speakers offered a sobering look at a landscape defined by paradox: record-breaking revenue growth coupled with extreme existential risk.

The State of the AI Bubble: A Paradoxical Reality

The central tension of the current investment cycle is whether we are witnessing a genuine paradigm shift or a classic speculative bubble.

"There’s both a bubble and not a bubble," asserted Chang Xu. The argument against the "bubble" narrative is rooted in the velocity of adoption. Xu pointed to the meteoric rise of OpenAI’s ChatGPT, which scaled from zero to $40 billion in annualized revenue in less than six months. She cited her own portfolio company, OpenArt, which saw its annual recurring revenue (ARR) jump from $1 million to $10 million in its first year, and from $10 million to $70 million in its second, all while maintaining cash-flow positivity with a lean team of only 20 employees.

However, Carter Reum offered a more cautionary historical perspective. "We pretend like this is the first time in venture capital land, but we’ve seen this before—with cloud, with the iPhone, with the car in the 1920s," he noted. Reum emphasized that while the current cycle is undoubtedly faster and steeper, the fundamental mechanics of market competition remain unchanged.

The differentiating factor, according to Reum, is the intensity of the incumbents. Unlike previous cycles where startups primarily competed against other startups, today’s innovators are battling the largest, most well-capitalized tech companies in history. "For the first time in history, the incumbents actually have the advantage—the tech, the capital, the data, and the talent," Reum said.

Chronology of the Current Cycle: From Infrastructure to Application

To understand where the market is headed, investors are looking at the evolution of the AI stack. The timeline of this boom has moved with dizzying speed:

1. The Foundation Phase (2022–2023): The explosion of generative models (DALL-E, Stable Diffusion) brought AI into the public consciousness. During this period, the focus was on the raw capability of LLMs and basic image generation.
2. The Tooling Phase (2023–2024): As companies began to integrate AI, the need for "below the AI" infrastructure became apparent. Databases, version control systems, and deployment tools—originally built for human developers—required a complete overhaul to accommodate autonomous AI agents.
3. The Application/Agentic Phase (2024–Present): We are currently in the era of specialized agents. As Xu noted, the industry has shifted from wondering if a tool is a business to realizing that even simple "wrappers" can command massive valuations if they solve a specific, high-value problem with speed and precision.

Supporting Data: Why "Depth Markets" Outperform "Velocity Markets"

Investment strategies have split into two distinct schools of thought regarding defensibility. Xu categorizes these as "Velocity Markets" versus "Depth Markets."

In velocity markets, the key to success is pure speed of execution. Because the frontier of AI capability shifts monthly, companies must iterate faster than their competitors to maintain an edge. However, this is high-risk; a competitor can often replicate a feature set in weeks.

Conversely, "Depth Markets" involve industries where the barrier to entry is not just technical, but structural or regulatory. These are sectors where "hard things remain hard." Xu highlighted a portfolio company using transgenic chickens to manufacture complex proteins—a process that is cheaper and more efficient than traditional laboratory methods. Despite the AI hype, the biological cycle of a chicken remains fixed, providing a natural, tangible "moat" that software-only competitors cannot easily disrupt.

Reum echoed this, pointing to his firm’s success in the 911 call center industry. By targeting highly regulated, high-friction industries, startups can create value that hyperscalers like Google or Microsoft may find too cumbersome to prioritize. "Friction is a moat," Reum remarked.

Official Perspectives: The "Microscope and Telescope" Mandate

Both investors agreed that the primary challenge for founders today is the inability to forecast the future with traditional metrics.

Reum advocates for a dual-vision approach: "I tell every founder: you need a microscope in one eye and a telescope in the other." The microscope is necessary for the day-to-day execution required to survive in a high-burn environment, while the telescope is essential for tracking the shifting tectonic plates of the broader AI landscape.

When evaluating deals, the "cocktail napkin math" remains the first line of defense. Investors are increasingly skeptical of startups that promise to revolutionize broad categories like "brand marketing" without a clear path to sustainable unit economics. If the math doesn’t check out on the back of a napkin, the venture firm is unlikely to proceed, regardless of the hype surrounding the AI model being utilized.

The Future: The Los Angeles Ecosystem and the "Taste" Economy

A significant portion of the discussion turned toward the local impact of the upcoming SpaceX IPO. Reum predicted that this event would act as a massive catalyst for the Los Angeles tech ecosystem, noting that the sheer volume of liquidity will likely trigger a second and third wave of entrepreneurial activity.

"The previous L.A. cycle produced things like Riot Games, Tinder, and Snap," Reum observed. "This is a different order of magnitude."

The consensus among the two investors is that the next phase of the AI revolution will move away from raw compute and toward "taste." While San Francisco remains the global hub for technical and engineering talent, Los Angeles possesses a unique advantage in creative industries, content, and emotional resonance.

"The next frontier in AI isn’t more compute—it’s taste," Xu noted. "Making films, making videos, making things that connect with specific cultures—that is where the human element becomes the primary differentiator."

Implications: The Second and Third Ripples

As the market continues to evolve, the most lucrative opportunities may lie in the "second and third ripples" of the technology. The first wave of any tech cycle is typically the most crowded and obvious, leading to inflated valuations and intense competition.

For investors like Reum and Xu, the real value will be found in the companies that emerge two to four years from now—businesses built on models and applications that are currently unimaginable. These future companies will likely benefit from more reasonable valuations and a clearer understanding of how to integrate AI into existing, complex, and regulated human workflows.

The verdict is clear: The AI revolution is not ending; it is simply entering its more difficult, more creative, and more selective phase. For those who can balance technical rigor with the human element of "taste," the opportunities ahead are unprecedented.