Venture capital giant says, tech bubble? What bubble?

The $50 billion market valuation of the car-hailing service Uber has been held up as an illustration that Silicon Valley is in the middle of another bubble.

The unicorns are multiplying, but if Andreessen Horowitz is right, the highflying venture capital community will have cause for concern when the inevitable thinning of the herd takes place.

The unicorns, of course, are the 60-odd technology startups that have taken investments valuing them at more than $1 billion each. The $50 billion valuation of Uber, the car-hailing service, has been held up as an illustration that Silicon Valley is in the middle of another bubble.

Andreessen Horowitz, the venture capital giant, disagrees.

In a presentation made by three partners at Andreessen Horowitz — Morgan Bender, Benedict Evans and Scott Kupor — to the firm’s limited partners last month, the three argued that the most recent exuberance was different from that of the dot-com era.

The firm released the document — “U.S. Tech Funding — What’s Going On” — to the public, and it is worth a look, not just for the debate over whether a bubble exists but also because it leads indirectly to the fate of the unicorns and the eventual fallout.

The presentation is chock-full of statistics justifying high valuations: E-commerce is only 6 percent of retail revenue in the United States, so there is room to grow. Four billion people are online compared with 40 million in 1995. And technology funding as a percentage of gross domestic product is only 2.6 percent of the economy compared with 10.8 percent in 1999, so the market is not flooded with cash.

So maybe this time is different, but the presentation also talks up the unicorns as the big factor.

The Andreessen Horowitz partners argue that the gains that used to go to the public through initial stock offerings are now being transferred to private investors. For instance, Facebook went public with a valuation of more than $100 billion, and all of its gains went into private hands. For Facebook to match Microsoft’s public market returns, it would have to be worth $45 trillion, according to the presentation.

The unicorns are an extension of this and would have simply gone public earlier in a previous age. But now that the small IPO has essentially vanished, it is the venture capitalists who profit by building a diversified portfolio and investing in these “quasi IPOs,” which grab even more of the gains than Facebook.

It’s easy to dismiss this presentation as Andreessen Horowitz simply talking its book. After all, the firm is at ground zero of the unicorn frenzy and has been accused of bidding up valuations and inflating the bubble. And the presentation may miss the main point in its claims about changing markets. Namely, even if markets are shifting, that doesn’t mean that Instacart is worth $2 billion.

Additionally, these unicorns are not like big public companies, which normally have low bankruptcy rates. They are big bets on the future. If the Andreessen Horowitz partners are right, then the unicorns would have failure rates like big public companies’ — namely only 1 percent to 3 percent. And that is unlikely. Instead, they are more likely to have small company and venture capital failure rates of about a third.

And markets may not be shifting as much as people think. The aphorism about the decline of the small initial public offering has been running for about a decade now. But the truth is that the small IPO went away in 1998. In 1997, there were 180 small IPOs, defined as those with net proceeds of less than $31 million in inflation-adjusted dollars, according to Thomson Reuters. By 2000, there were only 24 as the Internet bubble pushed everyone to invest in bigger companies. Once there, investors continued to invest in bigger companies.

This preference may be because the smaller companies weren’t that great. About a third of the companies that went public after 1996 went bankrupt, and the remainder didn’t grow for the large part. There were no big home runs like Microsoft from previous years that compensated for companies that fizzled. The days when a Microsoft could go public with little or no revenue are gone. Investors now want big initial public offerings with real businesses, not speculation.

The venture capital market now looks like the defunct small IPO market without the big upside. About a third of the companies make money, compensating for the rest of the investments, which either stagnate or go bankrupt.

And this is where the Andreessen Horowitz presentation inadvertently highlights the problem of the unicorns: no exit.

Competition among venture capitalists to get in on technology startups is driving up prices exponentially. But that pricing is justified only if there is an exit. And as Dan Primack noted recently in Fortune, there isn’t one for unicorns. There have been only eight technology IPOs and two unicorn sales this year, including to LinkedIn.

During the dot-com bubble, venture capitalists could exit their investments through the public markets, foisting the eventual losses on them. But with valuations skyrocketing this time around and gains being captured by private investors, there is no IPO or sale potential for many of these companies. Even a sale to Google or Facebook is a flier. And let’s face it, these giants may already have buyer’s remorse. Does Nest still look as if it is worth $3 billion?

Of course, some will be successful and go public. Uber and Airbnb will surely have billions of dollars in value. And Google and Facebook may throw a few billion dollars more at some of these companies. But even the tech giants are wary of these valuations, and many more unicorns are likely to fail as venture capital companies run their normal life cycle. And there will not be much to salvage, especially with companies increasingly agreeing to cover some investors’ losses as part of late-stage financing.

The documents for these financings will make for ugly trench warfare between old investors and new when these unicorns seek to raise more financing as their valuations deflate. Andreessen Horowitz may profit if Uber and a few others make it, but it will most likely have to fight it out in many other investments.

The Andreessen Horowitz presentation may be right that the private markets have crowded out the public ones in Silicon Valley, but it does not justify the valuation spiral or identify who will buy these unicorns. Perhaps most important, it is silent on how to salvage a unicorn investment when it sours.