CHAPTER 20

Snowflake, San Mateo, September 2020

The IPO market gained a vote of confidence in the last days of summer when the highly respected venture capital firm Andreessen Horowitz published something of an update to McGurk’s June 2019 direct listing primer. This time, though, McGurk’s former colleagues offered a defense of the conditions that lead to an IPO pop.

Andreessen partners Alex Rampell and Scott Kupor spent 3,800 words and seven charts telling readers why they disagreed with the critics arguing that the IPO market was broken. “There’s a popular narrative that evil investment bankers are intentionally underpricing traditional IPOs to steal from companies, lining banker pockets and those of their fatcat Wall Street clients,” they wrote. “An IPO is far from perfect, but this narrative is almost completely false.”

Technology stocks surged on their first day of trading because investor demand far outstripped the available supply of shares. Lockups and the fact that companies only sold a 10 to 15 percent stake in the IPO crimped the share count, they wrote, at the same time that a variety of investors with different goals clamored for shares. If critics wanted to improve the IPO process, they wrote, they needed to reconsider the traditional lockup structure, aggregate more of the demand into the actual IPO process, and be more creative in “blurring the lines” between public and private companies.

Like Slootman, the authors were also wary of Gurley’s argument. They argued that comparing the IPO price to the trading price was akin to comparing apples to oranges. And they spent several paragraphs explaining the theory behind their belief.

The authors did give some ground, writing lower in the post that “without a doubt, free profits are stupidly being handed away to lucky IPO participants” who are chosen specifically with the expectation that they sell the shares the next day.

The post was quickly shared across Silicon Valley and Wall Street, becoming an often cited rejoinder to critics who argued that the IPO process should be changed. Among others, the VCs thanked Carl Chiou for reading drafts.

Over the following weeks, the IPO candidates ramped up their discussions with institutional investors.

On September 8, Snowflake disclosed that it had signed up two large buyers to act as anchor investors in its IPO. Salesforce Ventures LLC and Warren Buffett’s Berkshire Hathaway Inc. agreed to buy $500 million in stock at the IPO price. Buffett hadn’t invested in an IPO in any industry, CNBC noted, since he bought shares in Ford Motor Co. in 1956. The following day, a Reuters column said that Snowflake’s IPO was “so hot it could melt,” and that the backing of Salesforce and Berkshire could send “Snowflake’s valuation even closer to the sun.”

Riccitiello, Jabal, and Davis had been busy too. That day, Unity disclosed the first price range for its offering, updating its prospectus to say that it planned to sell 25 million shares in an abnormally wide range of $34 to $42 apiece. The range represented indications of interest the company and its bankers had been collecting from investors. At the high end of the range, the company would garner a valuation of $11.1 billion.

The week of September 14, 2020, began with at least a dozen firms scheduled to go public, which Reuters said was the most in a week since 2014. The nearly $7 billion the companies hoped to raise would be the most in a week since Uber went public in May 2019.

Snowflake increased its IPO range substantially early Monday, raising the price to $100 to $110 per share, a 31 percent increase over the midpoint of the previous range. The higher range reflected the orders that institutional investors were calling in to their Wall Street salespeople since Snowflake was using the traditional IPO playbook. Investors were now willing to pay so much for Snowflake’s shares that the company’s valuation had soared above $30 billion.

On Tuesday, September 15, the tech-heavy NASDAQ notched its second consecutive gain, rising 1.2 percent.

That evening, Slootman, members of his board, and Snowflake’s bankers met to settle on a price and allocate the shares. Considering the strong demand from investors, the group felt comfortable increasing the price above the top end of the range. They settled on $120, a full 50 percent higher than the midpoint of the original range. Once the price was set and the shares were allocated to investors, Snowflake had raised $3.4 billion, making it the largest software IPO of all time and giving it a value of $33 billion.

As Slootman huddled with his bankers, a small drama unfolded at JFrog. Based in Israel and advised by Lise Buyer and Carl Chiou, JFrog’s executives had come to understand that bankers tended to market IPOs with a low initial price range to attract as many investors as possible. Chiou believed that the practice anchored investors psychologically to a lower price, making them less willing to raise their bids. Securities rules also contributed to anchoring the price by requiring companies to file an amended prospectus when a new range exceeded the old one by 20 percent. Bankers didn’t like to refile too many times, essentially capping increases.

JFrog had pushed its bankers at Morgan Stanley to increase the initial range, and even boosted it once to a high end of $41, but now that it was time to price the IPO, the bankers advised against raising the price much higher. Company executives texted with Buyer and Chiou that evening, gaining support as they debated the IPO price. JFrog finally sold shares for $44 apiece.

Both companies hoped that their lofty IPO prices would reduce some of the pop in the shares the next day.

Citadel Securities, the same trading firm that had handled the opening trades for Spotify’s and Slack’s direct listings, had been hired by Snowflake to perform the same duty. By then, the firm had become the largest market maker on the floor of the NYSE, with responsibility for opening thousands of stocks every day.

As orders began filling Citadel Securities’ electronic systems that morning, Peter Giacchi soon realized that few of the investors who had received shares the night before were willing to sell them. At 10:23, Giacchi sent his first pre-trade indication. Snowflake shares might open between $155 and $160. The higher price told the Wall Street traders watching for his signal that demand for the shares was outpacing the supply.

Over the next hour, Giacchi increased the price of Snowflake’s shares with a goal of finding enough sellers to match buyers at an agreed-upon price. His indications said as much, climbing to as high as $240 for one share of Snowflake’s stock. When Giacchi finally opened the stock, at 12:38 p.m., he did so at a price of $245. Snowflake had more than doubled the IPO price it had set the previous night.

CNBC broadcast the opening trade to millions of viewers. As the network’s producers flashed a “Breaking News” chyron across their screens, an anchor, Scott Wapner, declared, “Snowflake is now open for business.” Wapner quickly gave a rundown of the IPO’s superlatives—the biggest of the year, the biggest of all time for a software company, the fifth-biggest of all time for a technology company.

The stock shot up to more than $254, an increase of 111 percent from its IPO price. Snowflake’s IPO “is a raging success, to say the least,” Wapner said. His comment showed that despite the protestations of IPO critics, the media was still conditioned to view a large first-day pop as a marker of an IPO’s accomplishment.

All of a sudden, another CNBC colleague on an open mic blurted out, “Halted.” An on-screen graphic showed that the stock’s price had flatlined. In fact, Snowflake’s shares had surged so much that it triggered a pause in trading known as a circuit breaker that was intended to provide relief from wildly swinging prices. Once the stock was allowed to start trading again, it surged as high as $319.

With the shares now trading, Slootman made the customary rounds for interviews with financial journalists. In a video interview with CNBC, he shrugged off a question about Snowflake’s massive price surge. “This is just a hot deal, and we’ll have to live with the consequences of it,” he said.

Perhaps a company could sell shares at a higher price than it did at the IPO price, he told Reuters, but to suggest that the entire IPO amount could be sold at the price where the shares were trading in the market was disingenuous. “The idea that we could have sold all twenty-eight million shares at the highest price we’ve seen today is complete and utter nonsense. Markets don’t work that way,” Slootman said. “That’s why this whole DL [direct listing] narrative and all the noise around it is incredibly misguided.… What an IPO process does, it discovers the price at which you can move your entire offering. And of course that’s a much lower number than the number at which you can move a hundred shares.”

When Slootman appeared on Bloomberg Television that afternoon, it didn’t take anchor Emily Chang long to probe similar ground. “You left a lot of money on the table,” said Chang, speaking in front of a small set that showed bookcases over one shoulder and the Bloomberg logo over the other. “Do you have any concerns that you have given the bankers a little too much here?”

“There are always investors that are willing to buy some number of shares at higher prices,” said Slootman, in lightly accented English that gave away his Dutch roots. “As you know, the IPO is a price discovery process, and we were after a very specific set of institutional investors, people that can hold multibillion-dollar positions, that are willing to hold them for five to ten years, and also people that don’t chase momentum, either up or down. So that’s really how we landed at our price.”

Slootman continued, the corners of his mouth occasionally turning upward into just the hint of a grin that suggested he found the debate humorous. “It’s not unusual for prices to get chased up by momentum players, retail and so on, so there is a lot of noise to it,” he said, beginning to laugh. “The idea that we could have sold the whole thing at the price we closed at I don’t think is reasonable.”

When the trading day ended, Snowflake’s shares closed at $253.93, a 112 percent increase over the IPO price. Snowflake now had a market capitalization of $70.8 billion, more than double its value the previous night.

JFrog’s shares, on the other hand, rose a more modest 47 percent.

Gurley’s influence over the industry dialogue was evident in how the media handled its interviews with Slootman. For much of the day, the venture capitalist refrained from wading into the debate. Then, at 5:40 p.m. EST, Gurley fired off a salvo of Twitter messages.

“Some encouragement to comment on $SNOW IPO,” Gurley began, using the company’s stock ticker. “While it would be easy to do normal post [with regard to] mispricing, it is important to understand what is different here from other IPOs. The most important data is broad (40 years of underpricing, 2020 worst year yet), vs. 1 company.”

In other words, Snowflake was just the latest company to underprice its IPO. Jay Ritter, himself a critic of the first-day pop, found that the average first-day gain for the forty-four companies that went public between July 1 and September 16 was almost 58 percent, the most since 1999 if it was sustained for a full year. Whether it was overexuberance or underpricing, the market was as hot as it had been in two decades.

Gurley sent a second tweet praising Slootman and offering congratulations to a handful of the company’s venture backers, whom he called out by their Twitter handles. And then he sent a third tweet.

“Outside of if the company/shareholders ‘gave up’ anything, the hand allocated investors received $4.3B is one day wealth transfer,” he wrote. “That’s an insane amount of REAL money. That, along with watching the theatre and drama today, it is HARD to say—this is exactly how it should work!”

Gurley wasn’t done. “In many ways, $SNOW is the final proof of just how broken [the] process is,” he tweeted. “Frank Slootman is a HIGHLY experienced IPO CEO. He knows the game, & pushed hard to make sure he wasn’t short-changing the company. But it didn’t matter, because the process is set up to deliver this silliness.”

Months later, Gurley said that he had traded messages with Slootman and come to believe that the CEO’s experience allowed him to drive the process more than most executives could when it was their first IPO. “I think he knows how it works,” Gurley said, “and he is so fucking smart that he figured out a way to game their game.”