First, we need to define two important metrics.
DPI (Distributions on Paid-In Capital). This is the cash distributed to LPs after management and performance fees. This is by far the most important metric for LPs because it can’t be bullshit. It’s the truest measure of fund performance.
TVPI (Total Paid-In Capital). This is the mark-to-market value of all of the fund’s assets, also known as “paper earnings.” VCs have a lot of discretion over how to mark their assets (more on this below), so TVPI is often inflated. It’s easy for a good LP to dig deep in due diligence and see through a fake TVPI.
1. Most funds with higher TVPI have lower DPI.
This is the famous "VC Ponzi scheme" described by Chamath, showing your paper returns to LPs, thereby raising a large number of giant funds and earning generous management fees.
Especially in the cryptocurrency space, VCs have a lot of wiggle room when deciding how to bookkeeping. For example, a VC invests in a project that launches a token with a lockup period of several years and low liquidity on exchanges. Many will choose to mark their investment at the current spot price of the token without any discount, even though they can't possibly sell the token to realize such a gain. So, this is why new shiny L1 VC chains keep getting funded.
Therefore, it is prudent to achieve a good DPI over time to show that your returns are real. It is nice to have a high TVPI, but you should let it fluctuate with the market and write down the investment when appropriate.
2. The DPI of most 2017-18 vintage funds is lower than that of a16z crypto Fund I.
Obviously, I can’t share exact numbers, but a16z getting the fund X times bigger is pretty impressive.
a16z is a great brand, and for large institutional LP investment committees that have to reach consensus on decisions, investing in a16z is the equivalent of "no one got fired for choosing IBM."
Why would an LP choose a16z over other lesser-known funds if their returns are likely to be worse and they also have to take on more reputational risk? This is especially true for individual GPs who are currently raising a second fund and whose first fund performed poorly.
3. In the 2021 bull market, one-year funds were among the worst performing funds.
One-year funds are when a fund quickly raises and deploys all of its capital in one year, then raises the next fund in the following year. By asking institutional LPs, who have been investing in venture for decades, I could not find a historical example of a one-year venture fund that performed well.
Any institutional LP will tell you that the most important factor in fund returns is vintage, not the ability to get good deal flow or pick good companies. But as a VC, you can’t control the macro environment. Your job is to discover the best founders and themes within your area of expertise and deploy slowly over multiple vintages to spread out the macro risk.
2021 has been a particularly bad year for crypto VCs. Seed rounds were done at crazy $50M+ valuations before product, and the risk/reward made no sense. So many startups were raising money that VCs felt pressure to deploy capital. Additionally, startups were raising follow-on rounds in a short period of time, forcing early investors to exercise pro rata equity to maintain their ownership percentage before seeing meaningful product traction. This led to VCs deploying capital faster than expected and going back to LPs to raise the next round sooner than expected.
4. It’s either a small seed fund or a large index fund, there’s no-man’s land in between.
This is a great example of dialectic (opposite truths in extreme situations). Seed funds are "snipers", being the first money in a startup at an attractive valuation, where the risk/reward means only one seed investment project needs to succeed for the fund to make a return. Large funds are "aircraft carriers", indexing the market and having meaningful ownership in most companies post-market.
Fund size is often a legitimate signal for VCs to compete openly. New managers feel peer pressure to raise billions of dollars in AUM, but the fund size game often favors existing brands. Seed funds that were free to raise larger funds during the bull market are now in the middle, too big to get upside in seed investing, too small to become household names. They can’t raise funds of that size again anytime soon, and shrinking their fund size would reduce management fees.
Just because you can raise a bigger fund doesn’t mean you should.
5. Unfortunately, in bear markets, it is becoming increasingly common for VCs to withdraw from term agreements and demand refunds from investee companies.
Pulling a term sheet means that the VC had agreed to invest at specific terms, but while waiting for the legal documents to be completed, the cryptocurrency market crashed and the VC backed out of the deal. This does not mean the VC has to actually sign the document; see YC's handshake agreement for what is considered industry standard. Regardless, it can set founders back months because they have to waste time raising money again in a worse environment.
Asking for a refund from a portfolio company means that the VC funded the company during a bull market, but now regrets it and wants the company to return their money. This isn’t as bad as withdrawing the term sheet, since obviously the founders are under no obligation to agree to return the funds, but it’s still a stain on the VC’s reputation for being “founder-friendly.”
So far, I've heard from founders and other investors about five prominent crypto VC funds that have done this. The worst offenders have done this to at least five different companies. I've also noticed that generally the more a VC invests in their public image, the more they feel they can get away with bad behavior like this behind the scenes.