#310: Private Company Valuations Appear To Be Falling, & More

Last week, Bloomberg reported that Instacart, the private grocery delivery startup, had dropped its valuation by almost 40%, from the $39 billion announced roughly a year ago in March to $24 billion. Then, Axios noted – curiously – that the drop in valuation pertained only to Instacart’s 409a common stock which, for the most part, employees own, not to the preferred stock that investors own. However, at the beginning of 2021, the difference between Instacart’s 409a valuation and its preferred stock valuation was only 5%, and as early as January 2022, Fidelity lowered the fair value of Instacart’s Series I preferred stock by 25%[1] – both pieces of data suggesting that Instacart’s preferred stock could experience a similar valuation cut as Instacart’s common stock.

After more than a year of a bear market in innovation stocks that are not part of broad-based public equity benchmarks, we believe Instacart’s move marks the first major publicly announced valuation cut by a private startup. One of its public competitors, DoorDash, DASH on the NYSE, also has dropped roughly 40% since March 2021.[2] Instacart still could be overvalued relative to DASH, however, as its revenue growth last year was less than a third of DoorDash’s, 21% versus 69%, while its price-to-sales ratio is more than 50% higher at 13x compared to DASH’s 7.5x.[3]

In last week’s newsletter, we noted that inflated valuations often hit employees disproportionately relative to private investors with liquidation preferences, suggesting that talent might be hesitant to join startups with high valuations. Instacart might have published its 409a valuation decline to address this concern. According to Bloomberg, Instacart hopes to “boost recruiting and retention efforts by aligning new equity awards with the updated valuation that could have more potential for gains over the long term if the company regains favor in the markets.”

Meanwhile, data from private secondary markets also suggest that startup valuations are responding to the drawdown in public markets. Secondary marketplace Forge reported that prices for companies that traded on its platform both in the fourth quarter and in February declined roughly 10%, and that the percent of indications of interest (IOI’s) on the sell side was 60%, the highest level since the depths of the coronavirus crisis during the first quarter of 2020 and a stark reversal from the 60% of buyer IOI’s in 2021.

[1] Based on the Fidelity’s Growth Company Fund (FDGRX) 2021 Annual Report, which lists the acquisition price of Instacart Series I shares at $125 per share in February 2021, and its latest Monthly Holdings Report, which reported the value of the shares at $94 per share as of 1/31/22. Both documents available here.
[2] 40% calculated based on average Doordash share price of $162 in February and March 2021 and $98 in February and March (year to date) 2022. Share price data sourced from ycharts.com.
[3] Bloomberg (https://www.bloomberg.com/news/articles/2022-03-25/instacart-slashes-its-valuation-by-almost-40-to-24-billion?sref=1f7Aj053). Instacart’s 2021 revenue grew by 20% to $1.8 billion. $24B/$1.8B = ~13x
Yahoo Finance (https://finance.yahoo.com/quote/DASH?p=DASH). Doordash’s 2021 revenue grew by 69% to $4.9B and currently has a market capitalization of ~$36.5B. $36.5B/$4.9B = ~7.5x.

OpenAI recently announced two new GPT-3 capabilities, edit and insert, which enable the model to revise existing text in a wide-variety of ways, including rewriting a paragraph based on a change in context and converting story texts from first- to third-person. Until now, GPT-3 could generate new text but was unable to edit existing text.

The new capabilities also work on software code. In Codex, software engineers can use edit and insert functionality to refactor code, autocomplete functions with contextual knowledge of other code sections, translate between languages, and more, all of which should increase the productivity of software engineers significantly. In ARK’s Big Ideas 2022, we suggested that AI tools could double the productivity of software engineers by 2030. Codex’s progress has increased our confidence in this forecast, as have conversations with software engineers and executives who are extolling compelling examples of real-world use.

This week, Bloomberg reported that Exxon Mobil Corp. is running a pilot program using natural gas that typically would be flared and wasted to mine bitcoin in North Dakota. The oil giant also is considering similar operations in Alaska, Nigeria, Argentina, Guyana, and Germany.

According to anonymous sources, in January 2021 Exxon struck an agreement with Crusoe Energy Systems “to take gas from an oil well pad in the Bakken shale basin to power mobile generators used to run Bitcoin mining servers on site.” This pilot program uses up to 18 million cubic feet of gas per month that otherwise would be flared or burned off into the atmosphere.

Although unconfirmed by both Exxon and Crusoe, we believe the program could become a crucial and economically beneficial strategy as major oil and gas producers attempt to meet the ESG standards that many investors are demanding. Natural gas is composed largely of methane, a substance up to 80 times more potent than carbon dioxide in causing global warming during its first 20 years in the atmosphere. Exxon’s project follows ConocoPhillips’ push to supply Bitcoin miners with excess gas, also from the Bakken shale.

In our view, the convergence between Bitcoin mining and energy production will accelerate. Bitcoin’s ability to incentivize the discovery of cost-efficient energy sources, independent of location and consumer demand, has profound implications for energy infrastructure. Not only can Bitcoin mining capture natural gas that otherwise would be flared, but it also will encourage investment in intermittent energy systems like solar and wind, increasing the share of renewable energy provisioned to grids globally.