The first VC founded in 2026

@Guo Yunxiao

The first new fund to announce its launch in 2026 is here: Nebulon Ventures.

"Nebulon" means nebula. In astronomy, it's the prelude to a galaxy — hazy, dim, yet pregnant with energy. Capturing one requires the observer to mount a camera and hold the shutter open for an extended exposure — exactly what founder Yongteng Wen had in mind when he chose the name: in an era of rapid AI transformation, the patience to maintain a "long exposure."

At the end of 2025, this post-90s investor left his previous firm. Within four months, Nebulon Ventures had completed its first close, raising tens of millions of dollars.

That speed came from the payoff of a long-term bet: as early as 2021, Wen began systematically exploring AI, making many investments before ChatGPT ever appeared. From then on, he believed AI applications would inevitably go global, so he laid out positions worldwide and gradually became a bridge between entrepreneurial communities on both sides. The result: he backed multiple AI startups with ARR exceeding $100 million, including being one of HeyGen's earliest investors.

As a peer and supporter, he has a natural resonance with young founders and offers them unreserved trust. He noticed that founders of Chinese and American internet companies worth over $100 billion all started before age 30 — but in the AI era, that threshold will certainly be pushed even lower.

Nebulon's fundraising also landed at this inflection point. According to elsewhere, Wen only began formally raising after leaving his previous firm, and closed the first round in four months. One LP's assessment: "You can tell this young man genuinely likes and believes in AI." Perhaps because new faces in dollar-denominated funds have become so rare, many senior veterans, while serving as LPs, also generously shared their hands-on experience.

The high-density cognitive input during fundraising made Wen feel like he was undergoing the industry's "compression algorithm": in a quiet cycle, people were willing to rapidly inject decades of experience into this young player who had just sat down at the table.

Starting a new fund has been the choice of many of the most ambitious investors over the past decade. But if you met Wen, you'd struggle to connect him with such ambition. This reserved person habitually smiles shyly. In our conversation, the dialogue kept unconsciously drifting toward science fiction and idealism.

This resembles the investment life Wen imagines: deep in the starfield, observing long, digging continuously.

Finding the Highest-Compression Young People in the Language and Network Revolution

Wen's definition of AI is quite interesting.

Most people compare AI to the "electricity revolution," using power consumption as an analogy for token burn to find AI projects. But in Wen's eyes, this is more like a "language and network revolution": the internet only changed the speed of language transmission, while AI reconstructs how language operates — it brings machines into humanity's language protocol and makes language directly executable.

Language is the greatest invention in human history, and AI is its epic-level upgrade.

From this angle, Wen believes we should now view future AI industry opportunities through a "network" lens, not a "monolithic application" lens.

The reasoning comes from a biological perspective: nature never evolved an omniscient, omnipotent super-organism, but rather co-evolved through diverse species and populations within species. Future AI will likely follow this law — it probably won't internalize all capabilities into one "model," but will differentiate into countless AI individuals.

Where there are populations, there must be collaboration.

In traditional networks, only humans are nodes that can autonomously initiate connections; but in the future, AI systems will not only initiate tasks proactively, but Agents will collaborate directly through code and standard protocols. When such collaboration costs drop dramatically, future network nodes will no longer be passively click-waiting software, but AI clusters capable of autonomously initiating collaboration.

Therefore, Nebulon is not betting on traditional software interfaces, but on the new ecosystem built around this vast "digital population." "While traditional software will be deconstructed, the wave of applications built around Agent and human networks is just beginning."

If this is what future commerce looks like, then the founders who can build these products naturally cannot be observed through old lenses.

First, "intelligence is compression." This is Ilya Sutskever's famous line. Wen uses it as a yardstick for entrepreneurs: in an era of information explosion, founders must be "high-compression-ratio channels" without redundant baggage, capable of pinpointing the essence of things with extreme precision.

Take HeyGen founder Joshua Xu. When Wen first met him in 2021, Joshua's first sentence went straight to the point: "I want to use AI to replace cameras." This founder from a major tech company could translate signals and trends he observed at work into an extremely streamlined goal, and continuously build around that through-line. In Wen's view, this is extremely high "compression ratio."

Second, "the ability to zero out." In this wave where traditional software and applications are being thoroughly deconstructed, humans themselves become the biggest bottleneck. Those who can truly capture the next generation of product dividends must possess a "zeroing-out ability" — the courage to unconditionally trust the model and resolutely abandon all past successes and failures.

Wen locks his gaze on two types of people: first, young scientists around 28 to 33 years old who are actually writing code on the front lines and have deep algorithmic backgrounds; second, digital natives around 25 to 29 years old who are extremely sensitive to new technologies and global community culture.

The two precise age ranges have different origins. The first is because PhDs typically complete around 26-28, and they need a few years of engineering/commercialization adaptation; the second is because great founders mostly achieve results between 28-32, averaging 7-8 years of "serial entrepreneurship" experience, which AI can compress to 3 years.

In Wen's view, AI is ruthlessly flattening business experience that once took decades to accumulate. At this new table, young people's intuition and decisiveness have instead become the chips with the highest probability of winning.

Investment Is Still a People Business

Wen summarizes the changes he sees in startups into four points: fast industry change, leaner scale, younger teams, and global-facing markets. Many investors see these trends, but few map them back onto their own funds.

As for the Solo GP concept, Wen himself doesn't particularly like this somewhat romantic term. He admits that running a fund alone brings no "joy of freedom" — rather, greater responsibility and occasional loneliness.

His reason for temporarily maintaining a solo structure is actually quite simple: a larger team means more complex operations and greater fundraising pressure, and right now, he wants to get onto the battlefield as quickly as possible.

In his view, in the AI era, information flow speed within an organization determines survival or death. In a track where underlying technology leaps every few months, any communication requiring hierarchical reporting or alignment meetings becomes "thermal noise" that slows flow.

In Nebulon's daily operations, AI has already begun large-scale takeover of back-office functions. Research that once took one to two weeks can now produce a high-quality overview in half an hour; market feedback on a product can be rapidly completed by scraping social media sentiment.

Wen finds he can finally reserve bandwidth for "what AI cannot do." Such as sensing a founder's original intention, judging whether their vision stems from vanity or passion, or providing some human companionship when they're lost.

"Venture capital is fundamentally still a People Business," Wen says. "Receiving an AI-written email, clicking confirm, and the money transfers — at least today, that's not quite realistic."

Or, for example, starting to do some seemingly cost-inefficient "externality" work from day one of the fund's existence.

He plans to support a documentary about young entrepreneurs, because "running a fund is not a great endeavor in itself" — what's truly great are the young people taking risks in this era. Even if no commercial investment ultimately results, he hopes to leave some archival footage for them, to leave a little "positive externality" in the world.

If it were purely about making a profitable business, thinking to this level would be sufficient. But in conversation, Wen keeps unconsciously sliding toward grander things. Such as: when will AGI arrive? Will human civilization ultimately be driven by one or a few tech companies?

Facing these ultimate science-fiction questions, Wen has mostly imagined them, without certain answers. But he feels that if you don't seriously think about these grand problems, when the correct answer truly appears, you won't even be able to recognize it.

Nebulon's emergence is also a slice of this era: extremely fast, extremely lean, extremely idealistic.

Cover image: Provided by interviewee