5x Returns on an Early-Stage AI Fund: The Trust and Returns Game in Venture Capital

A Single Tweet That Sparked VC Reflections

Recently, a venture capitalist posted a brief but impactful update on Twitter: "My first fund just hit 5x returns!"

Behind this tweet lies an intriguing backstory — last year, a potential GP (General Partner) declined to invest, citing uncertainty about whether this fund manager could deliver 5x returns. Now, reality has provided the most direct answer.

In the structure of a venture capital fund, GPs and LPs (Limited Partners) form a classic principal-agent relationship. The GP handles day-to-day fund management and investment decisions, typically charging an annual management fee of around 2% of fund size plus approximately 20% of excess returns (known as Carried Interest, or Carry). LPs are the primary capital providers, including institutional investors such as pension funds, university endowments, and family offices. It's worth noting that the "potential GP who declined to invest" mentioned in the tweet actually refers to a GP who could have participated as an LP — in the venture capital industry, GPs of established funds frequently invest as LPs in other funds, and this "GP-invests-in-GP" model is a crucial part of the industry's trust network.

This investor also revealed that their second fund is performing quite well, and a third fund is currently being raised. From being questioned to letting performance speak for itself, this story reflects some thought-provoking patterns in early-stage venture capital.

What 5x Returns Mean in the VC Industry

Performance Benchmarks for Early-Stage Funds

In venture capital, the fund return multiple (TVPI, Total Value to Paid-In) is one of the core metrics for measuring fund performance. TVPI is calculated as (Distributed Value + Residual Net Asset Value) ÷ Paid-In Capital. Related metrics include DPI (Distributions to Paid-In) — the ratio of actual cash distributed to LPs versus paid-in capital, which is the hard metric for measuring "real money" returns — and RVPI (Residual Value to Paid-In), the ratio of unrealized portfolio valuations to paid-in capital. The relationship is TVPI = DPI + RVPI. Sophisticated LPs evaluate all these metrics simultaneously and apply different weighting depending on where a fund is in its lifecycle.

For early-stage venture capital funds:

• 1-2x returns: Barely passing — after deducting management fees and carry, LPs are left with very little
• 3x returns: Above-average industry performance, enough to reach the Top Quartile
• 5x and above: Top-tier performance, typically achieved by only the top 10% of funds

Driven by the current AI wave, early-stage funds that bet on the right sector can indeed achieve high return multiples in a relatively short timeframe. Since ChatGPT's release in late 2022, venture capital in the AI space has experienced an unprecedented round of valuation inflation. Taking large model companies as an example, OpenAI's valuation surged from approximately $29 billion in early 2023 to over $80 billion in 2024, while Anthropic completed multiple large funding rounds within just two years. The underlying logic behind this valuation growth is that generative AI is viewed as the third platform-level technological shift after the internet and mobile internet, with a potential market size covering virtually all knowledge work domains. For early-stage funds, if they entered these companies at seed or Series A valuations in 2021-2022, the valuation increases from subsequent funding rounds alone would be sufficient to achieve multiples — or even tens of multiples — in paper returns.

The Gap Between Paper Returns and Actual Exits

It's important to note that "5x returns" when a fund hasn't fully exited likely includes a significant amount of unrealized gains. In a fund's early stages, TVPI is often primarily composed of RVPI — meaning most returns exist only on paper. Only when portfolio companies achieve exits through IPOs, acquisitions, or secondary market transactions does RVPI convert to DPI — the cash returns that are truly in the bank. Given the significant valuation volatility in the tech industry, actual final returns may differ from current figures. The high valuations in AI also come with considerable uncertainty — the sustainability of business models, the evolution of compute costs, and competitive pressure from open-source models against closed-source models could all impact ultimate exit returns.

That said, reaching 5x TVPI in the early-to-mid stages of a fund's lifecycle is still an exceptionally strong signal.

The Trust Paradox in GP Selection

The Fundraising Dilemma for Emerging Managers

The most thought-provoking part of this tweet is the potential GP who "declined to invest because they weren't sure about 5x returns." This reveals a classic paradox in venture capital:

Emerging Managers who most need capital to prove themselves are precisely the group that finds it hardest to secure funding.

Emerging managers typically refer to GPs managing their first or second fund who haven't yet established a complete fund-level track record. According to research from institutions like Cambridge Associates and the Kauffman Foundation, emerging manager funds statistically tend to outperform established large funds. Reasons include: smaller fund sizes make it easier to concentrate investments in high-return opportunities, the manager's personal reputation is tightly linked to fund performance creating stronger incentive alignment, and emerging managers often possess unique deal flow channels. However, institutional LP allocation processes typically require GPs to have at least one complete fund's track record, creating the classic chicken-and-egg dilemma. In recent years, some Fund of Funds specializing in emerging managers have emerged to fill this market gap, but overall, fundraising difficulty for emerging managers remains far greater than for established GPs.

Institutional investors and Fund of Funds rely heavily on historical track records as a core decision-making criterion when screening GPs. But for first-time fund managers, there's no complete fund-level performance to reference — they can only rely on their personal investment judgment, industry networks, and deal pipeline to convince LPs.

The Risk of Adverse Selection

This screening mechanism inherently carries adverse selection risk:

• Over-reliance on historical data may cause investors to miss emerging managers with forward-looking judgment
• In rapidly evolving fields like AI, past performance may not predict future results
• Investors who can truly capture paradigm-shifting opportunities are often those who hold "non-consensus" views

The GP who declined to invest used a traditional evaluation framework to assess an emerging manager and ultimately missed out on 5x returns. This case reminds us that at critical junctures of technological change, overly conservative decision-making frameworks can carry enormous opportunity costs. As an old saying in venture capital goes: "The best returns often come from investments that looked the most uncertain beforehand."

Implications for the AI Investment Ecosystem

Investment enthusiasm in the AI space continues to run high, with capital flowing in at unprecedented speed — from foundation models to the application layer, from chips to data infrastructure. In this environment, several trends are worth watching:

1. The window for early-stage funds is narrowing: As more capital floods in, valuations for quality deals rise accordingly, giving early entrants a significant first-mover advantage. Take seed rounds as an example: in 2021, the median seed-round valuation for AI companies was approximately $10-15 million, but by 2024, seed valuations for comparable projects have generally climbed to $30-50 million or even higher. This means later entrants need much higher exit valuations to achieve the same return multiples.
2. The value of specialized funds is becoming more apparent: In a technology-intensive field like AI, specialized funds with deep technical understanding are better positioned to identify truly valuable projects. Distinguishing genuine technical breakthroughs from beautifully packaged pitch deck fundraises requires GPs to have solid technical backgrounds or deep industry networks.
3. Performance validation accelerates trust-building: Once a first fund proves its capability, subsequent fundraising becomes significantly smoother — just as this investor's second and third funds demonstrate. In venture capital, this is known as the "flywheel effect": strong performance attracts higher-quality LPs, ample capital helps GPs access better deals, and better deals further improve performance.

From being questioned to achieving 5x returns, the core takeaway from this story is perhaps quite simple: In the early stages of a technological revolution, those who dare to place bets and stick to their convictions often earn outsized returns. Meanwhile, those who wait for "certainty" may never find the best time to enter.