Is OpenAI worried about triggering a buyback?

Author: Jessica, Silicon Star Pro

The "repurchase agreement" has recently stirred the nerves of everyone in the Chinese venture capital circle.

From the dramatic farce of Luo Yonghao and Zheng Gang, to the overwhelming repurchase trend reflected in various data, to the voices of practitioners represented by Qiming Venture Partners' founding managing partner Kuang Ziping who believe that the repurchase clause has been abused, the discussion about "stampede-style" repurchase has spread.

According to the latest "VC/PE Fund Repurchase and Exit Analysis Report" released by Lifeng Law Firm, about 130,000 projects in the primary market are facing exit pressure one after another, and "tens of thousands of entrepreneurs may face repurchase risks of hundreds of millions of yuan."

In Kuang Ziping’s influential article, he mentioned the origin of the buyback clause, which is an “imported product”: “The early investments of US dollar funds in China were mainly based on Cayman mainframes, and generally there would be “redemption” or “repurchase” clauses, and occasionally performance ratchet clauses.”

But these imported products “have been abused by some peers over the past 10 years.”

Behind the discussion about the current buyback boom is the concern about whether the prosperity of the entire science and technology investment can continue. At the same time, a glaring contrast is that Silicon Valley start-ups, represented by AI companies, continue to raise large amounts of financing.

OpenAI is the most obvious representative, and the latest rumor is that it is trying to raise billions of dollars more. In addition to the previous investor Thrive Capital’s plan to invest $1 billion, Apple and Nvidia will also join the ranks of investors.

As the birthplace of Silicon Valley, what is the current situation of buybacks here? After Sam Altman took so much money from all walks of life, will he also lose sleep because of "buybacks"? We talked to local front-line investors.

1 "Be willing to gamble and accept defeat" guaranteed by sophisticated mechanisms

The bet and buyback clauses originated from the risk control needs of the US dollar capital market in global business practices, aiming to help investors lock in exit channels in a high-risk environment. Since the mid-20th century, these clauses have gradually become popular and standardized on Wall Street, and then promoted to the global financial market, becoming a common practice in the modern investment industry.

Specifically, the buyback clause stipulates that under certain specific conditions, the company needs to redeem the shares held by investors at an agreed price. These conditions may include the company's failure to go public within the promised time, the founder's major violations, or the company's acquisition.

For example, you are an ambitious founder who has received a 10 million yuan investment from an institution. When the two parties signed the contract, they agreed that if the company failed to successfully go public within five years, the investor would have the right to require you to repurchase the shares at 1.5 times the investment amount. As a result, five years later, due to various reasons, the company did not IPO. Then, according to the agreement, you must take out 15 million yuan to buy back the shares held by the investor.

The concept of performance ratchet is similar. You and the investor bet on a future performance target or milestone event, such as reaching 100 million yuan in revenue within five years. If you achieve it by then, you will be safe. If you fail to achieve it, you will have to compensate the other party with agreed personal equity or cash payment. In essence, it is equivalent to giving investors an insurance policy, which provides an extra way out when the invested company fails to grow as expected.

An investor who has worked for a leading US VC institution for many years told us that although the repurchase clause is an industry regulation and will be written into the agreement, it is almost never triggered in Silicon Valley, and even if it is triggered, the investor will not sue. If it is a company that they are optimistic about, they will stick to the principle of accepting the loss.

"Silicon Valley is a place that encourages entrepreneurship in terms of culture, atmosphere, and mentality. For example, Stanford VCs will support alumni when they see them starting a business. And many investors were entrepreneurs before they became investors, so they want to support and give back. For example, the first achievement of a16z founder Marc Andreessen was actually the founding of Netscape," he said.

"There is also Steve Chen, the co-founder of YouTube, who joined PayPal before graduating from his senior year and worked as a software engineer for five years. Later, when he was working on YouTube, he needed to spend a lot of money on copyrights and was in urgent need of financing. It was Roelof Botha, the former CFO of PayPal, one of the PayPal Mafia, who went to Sequoia as a partner, who invested in him.

How many successful giants came out of garages. If they had no money to save them in difficult times, they would have died countless times." - These investors have all suffered in the past, and after being caught in the rain, they also want to hold an umbrella for others.

In addition to the lack of support from the "ancestral" entrepreneurial culture in Silicon Valley, objectively speaking, the repurchase terms here are much more relaxed than those in China.

Because the technology companies invested by US dollar funds in the early stage are mainly based on the Cayman entity structure. According to Cayman law, repurchases must ensure that they do not affect the normal operation of the invested company before they can be implemented, and the responsible party is the company rather than the individual founder.

However, when this structure was localized by RMB funds, two changes gradually emerged: first, the premise of "not harming the normal operation of the company" was ignored, and the execution of the repurchase clause no longer considered the impact on the company's operations. Secondly, the "individual unlimited joint and several" liability was added, which transferred the repurchase obligation from the company to the founder. And it is "unlimited joint and several", no matter how much money you have in your account, you must return the promised amount in full. These two aspects together pierced the protective layer that was carefully laid out for the company and the founder in the original repurchase clause.

That's why some entrepreneurs lamented: "The repurchase clauses signed before were more like gentlemen's agreements, which were basically not triggered. Now signing repurchase clauses is like a sword of Damocles hanging over your head, and investors are serious with you."

Qiu Zhun, an overseas partner of Huaying Capital, explained that in addition to extremely loose repurchase clauses, Silicon Valley also has a complete system to provide support and security for entrepreneurs.

The first is not to enforce non-compete agreements, which is a typical California product. (For example, many technical backbones who left DeepMind to start large-scale model startups were not held accountable, but they may be sued in other states.)

The second point is the greatly simplified investment agreement (Term Sheet). The purpose is to avoid overly complicated legal details in the final signed investment agreement (SPA) through a concise and clear terms framework, thereby accelerating investment decisions and execution. Zhen Fund once launched a very popular "one-page" investment letter of intent, which actually evolved from the Silicon Valley Term Sheet. The Silicon Valley Term Sheet has been 1-2 pages from the beginning.

The third is that early-stage projects fully adopt the standard convertible bond SAFE agreement proposed by YC, allowing startups to obtain funds without immediately determining the company's valuation, and investors can convert it into equity in subsequent rounds.

The fourth is to change Wall Street's risk management mechanism and embrace risk (Venture).

This is also the fundamental difference between the two investment systems of Silicon Valley and Wall Street:

Wall Street's risk control mechanism is more conservative, relying on building models, financial forecasts, and strict terms to avoid risks and ensure investment security and stable returns, especially in the field of private equity (PE).

As a technology center, Silicon Valley emphasizes venture capital. VCs in Silicon Valley generally believe that once the growth of a start-up can be modeled and calculated, it loses its innovative value. They are willing to take higher risks to support innovative companies that may change the industry and bring huge returns, especially early technology companies. The idea of ​​Silicon Valley is that although most projects may fail, once a project succeeds, its returns will far make up for all risks. Therefore, Silicon Valley is more willing to "bet" on the disruptive nature of future technology, and to demonstrate its long-term trust in technological innovation with a higher risk tolerance.

"But this mechanism was not invented out of thin air, nor was it really an angel that descended from the sky." Qiu Zhun said, "VC also wants financial returns. From the perspective of traditional financial models, VC may seem a bit "alternative", but its true roots can be traced back to the development of the computer industry in Silicon Valley. The exponential growth and reinvestment characteristics of the computer industry have promoted the rise and value of VC."

It seems that to some extent, the investment style in China in recent years is closer to the Wall Street style, or it can be said that the high-risk role undertaken by Silicon Valley has been digested internally. Investors pay more attention to risk control and exit mechanisms, give priority to return guarantees, and pursue steady profits rather than winning in danger.

2 It is even used by entrepreneurs to "counter" investors

But no matter what, terms such as bet-back repurchases are also used to protect investors in the general consensus. But in Silicon Valley, it can also be a tool for startups to fight back against investors.

Last month, a Silicon Valley unicorn called Bolt staged such a drama: the company plans to complete a new round of financing of $450 million at a valuation of $14 billion. The CEO announced that if existing investors do not accept a price that is 30% higher than the previous round of valuation and continue to invest, Bolt will activate the "Pay to Play" clause and buy back their shares at an ultra-low price of 1 cent per share.

Pay to play is a special system in the US capital market that requires existing investors to continue to pay money in subsequent financing rounds to maintain their equity shares, otherwise they will lose anti-dilution rights, future financing participation rights, board seats and preferred stock status.

Bolt's announcement has put almost all old investors in distress: on the one hand, they do not want to continue to invest in Bolt at a high price of $14 billion, but on the other hand, they clearly know what it means if they do not comply. ——The equity buyback that has always protected investors has become a bargaining chip in the hands of entrepreneurs under the cover of Pay to play.

According to statistics from Cooley, a well-known American law firm, the proportion of investment agreements containing "Pay to play" clauses has increased year by year since the second quarter of 2021, reaching 8% in the first quarter of 2024. "This rare clause in the past is becoming more and more popular."

At first glance, Pay to play seems to be counterintuitive. Old investors provided early funds for the company, why are they so disloyal and want to be "washed out"? But when you return to the core of Silicon Valley's pursuit of risks, everything will become reasonable.

Because Pay to play is not simply to clean out certain investors, but to give them a choice: continue to invest and support the company, or give up subsequent participation. If investors are unwilling to continue to unite with startups and share risks, they must accept the loss of share dilution. On the other hand, from the perspective of some investors, they are no longer optimistic about the company's prospects, and the financial level has already defaulted on the failure of this investment. Share repurchases can also recover some funds, which is also a stop loss.

So, different soils will give birth to different tools to adapt to different rules and make everything run organically.

3 Some cracks appear under the bubble concerns

But is Silicon Valley so confident, self-consistent, and even a little utopian? Or has anyone begun to smell the hidden worries under the glamorous surface?

Behind OpenAI's new round of financing with a valuation of $100 billion, Silicon Valley is also quietly changing.

For a long time, the valuation and financing of emerging star companies such as OpenAI are not based on traditional profitability or financial performance, but on expectations for future AI progress, especially general artificial intelligence. OpenAI adopts a unique structure that combines non-profit organizations and for-profit subsidiaries to attract external capital to support high-cost AI research and development, while limiting investors' returns until the milestone of AGI is achieved.

However, this time OpenAI's actions exposed some uneasy signals. The Information recently pointed out in the article "Why OpenAI Needs an IPO" that despite OpenAI's impressive valuation and revenue, its profitability is not optimistic and it is expected to lose billions of dollars this year. Cloud services, GPUs, model training and employee costs are extremely high, and the company is unlikely to generate a lot of revenue in the short term, so going public may become their most urgent choice.

In addition, Thrive Capital has just led a round of employee share buybacks at a valuation of $86 billion, and then led a new round of financing. Typically, high-growth technology companies attract multiple investors, and the same investor's continuous lead may mean that other investors are cautious about the current valuation. Internal financing is often seen as a warning that the company may have difficulty attracting new investors. Although OpenAI is technologically advanced, the long-term viability of its business model and profitability remains unclear, and the market has therefore begun to re-evaluate this high-valuation, high-risk investment model.

The recent series of "reverse acquisitions" in Silicon Valley are also the product of this dilemma.

Many of the most glorious AI unicorn companies have been reversely acquired by large companies. Character.AI, Inflection, Adept and Covariant have all been recently acquired through technology licensing and talent "recruitment". Although these companies have high valuations, technology and products, investors are clearly impatient when they cannot find a stable way to make profits in the long term.

Silicon Valley is also aware of the bubble problem, but it seems that they think that if the music must stop, it may also be important to make it stop decently.