news

if we only consider the pros and cons from the capital level, then openai is still the most popular artificial intelligence large model super unicorn, not one of them.

at the end of august, major financial media broke the news that openai, the number one player in artificial intelligence, was about to complete its latest round of financing. the valuation of this round will be significantly higher than the previous round of us$86 billion. during our national day, financing quietly landed. openai raised us$6.6 billion (approximately 46 billion yuan) with a valuation of us$157 billion (approximately 1.225 billion yuan).

the leading investor this time is thrive capital, an old shareholder, with a leading investment amount of us$1.3 billion, and the option to continue investing another us$1 billion at this valuation in 2025. includemicrosoft、nvidia、softbank, tiger global, etc. have also joined the investment, including microsoft once again investing us$750 million.

once the deal is completed, this figure will directly break the record set by musk's xai earlier this year and become the largest venture investment in history. if we put this number ten years ago (2014), it would be equivalent to the total venture capital investment in new york, florida, and texas in the united states.

the second half of the title is a collection of news, because apart from the capital level,OpenAIthere has been a lot of news recently, and it can even be said that it has been deeply involved in the whirlpool of public opinion.

rumors appeared in august, and the specific transaction plan was exposed at the end of september. it is expected to be officially launched in the first week of october. from a time perspective, everything is going well for this largest venture capital investment in history. however, at such a critical juncture, openai suddenly fell into a series of "trouble".

exitedappleand bubbles in history

in the initial report, in addition to thrive capital and microsoft, openai’s investors also included tiger, cotu, apple and nvidia. among the old shareholders, only sequoia has chosen to suspend its support for openai because it invested in safe superintelligence, a startup company founded by ilya sutskever, the former chief scientist of openai.

it can be said that thisfinancingit can be called "the game of giants".

however, on september 27, an unexpected thing happened: apple withdrew from investment negotiations at the last minute, which will lead to certain changes in the final financing amount and investor lineup of openai’s current round.

and because apple did not announce a specific reason for its withdrawal, it did not immediately respond to requests for comment from the media. in the days that followed, there was a lively discussion about why apple withdrew its investment.

the mainstream view is that the real reason why apple gave up investment this time is most likely related to the cooperation they reached with openai in june this year. at wwdc in june, apple officially announced a "highly anticipated partnership" with openai. in the future, chatgpt will be integrated into the experience of ios, ipados and macos. at the same time, siri will also be integrated with chatgpt so that apple users can flexibly use chatgpt's functions (including image and document understanding) without having to switch between tools.

according to the final reports conducted by the media at the time, this cooperation will not incur any fees, no one will make money, and no one will pay. more importantly, multiple sources pointed out that apple believes that through system integration, it is equivalent to a promotion of hundreds of millions of openai units, and the exposure it brings "has a high cash value."

with this background, coupled with apple's historical lack of interest in venture capital, many believe that exiting the investment is actually a reasonable outcome. existaiin terms of business, the operation that is most in line with apple's tone should be to increase internal support for apple intelligence, or directly spend money to acquire an artificial intelligence company that is the first echelon in the industry.

a considerable number of onlookers also pointed out that "cash transactions" were the biggest obstacle. the reason is that although openai completed a round of financing with a total scale of us$10 billion from microsoft in january last year, most of the us$10 billion is the converted value of basic services such as computing power, not real money. according to the latest news in early september, although openai’s revenue this year is expected to reach us$3.7 billion, it still has to bear us$5 billion. and this us$5 billion only calculates operating costs, personnel salaries, and office space rentals, and does not include other potential expenses such as equity compensation.

the core of the liberal market economy is "it's hard for me to do things for you with such a small amount of money." it is not difficult to imagine that openai, which has single-handedly started an artificial intelligence arms race, is facing a huge cash gap. considering that openai’s valuation has exceeded us$100 billion and microsoft’s eye-catching 49% share in the shareholder structure, it can be said that this round of “cash is king” financing can easily appear to be “not cost-effective.”

there is also a theistic view that everyone should be wary of the "$150 billion" figure. because in 1999, the total amount of funds managed by american venture capital institutions was us$150 billion. this money directly promoted the formation of the internet bubble the next year, and then came the devastated memories.

don’t feel far-fetched, either. we must know that "overheated investment in artificial intelligence conceals the true situation of the market" has long ago changed from the experience of working in the past to the mathematical result of "real and verifiable results." for example, according to data compiled by crunchbase in august, more than 35.5 billion of the total global financing of 79 billion us dollars was invested in the artificial intelligence track. there were 35 in the united states alone, with a single investment amount of more than 100 million us dollars. artificial intelligence trading case. the excessive drama directly stimulated the editor of crunchbase to solemnly write a headline: "artificial intelligence start-up deals are gaining momentum, but is this a good thing?" 》

in short, theistic conversion to theism is quite popular in terms of popularity.

exodus executives and the openai gang

of course, apple's withdrawal is just a sideways glance at openai's hidden worries. the "trouble" that really hits the soul is the official resignation of openai chief technology officer mira murati.

mira murati joined openai in 2018 and was promoted to chief technology officer in 2022. during this period, she led the company to develop the famous chatgpt, vincent graph large model dall-e and code generation system codex. besides, she's still hereteslahe worked for three years as a senior product manager for model x. during this period, tesla developed an early version of autopilot, the artificial intelligence driving assistance system.

in her personal resignation statement released on september 25, local time, mira murati said that "although there is no perfect time," she believed that "it is appropriate to leave at this moment" because openai has "fundamentally changed artificial intelligence reasoning and learning." way” and she needed time and space to explore herself. sam altman responded immediately, "i am personally very grateful for her support and love during difficult times."

it seemed like a decent breakup. but just that week, openai contributed three other headlines:

• openai is rumored to be discussing providing sam altman with a 7% stake and reorganizing it into a for-profit business.

• sam altman published an unusually high-profile blog post, declaring that “ai can create anything we can imagine,” “through the ai ​​industry, we can achieve mutual prosperity to levels unimaginable today” and “we may super artificial intelligence within a few thousand days."

• the internal atmosphere at openai is not relaxed, and overtime culture is prevalent. especially in may this year, in order to avoid being in the limelight,googletaking away the annual i/o developer conference in may, openai chose to rush to release the latest flagship model gpt-4o, leaving only 9 days for testers.

all of this has led people to speculate whether mira murati marks openai's complete transformation into "sam altman's company," and sam altman's transformation from a future technology navigator back to a "businessman." further discussion is whether mira murati will follow in the footsteps of ilya sutskever and continue the story of openai mafia.

the term "openai gang" was modified from "PayPal"paypal mafia" (paypal mafia), the latter is used to describe paypal that has provided a large number of entrepreneurs to silicon valley, and venture capital institutions are also willing to send investments that are much higher than the average standard because of this status. people see similar potential in openai , at least according to current data, nearly 75 employees have left openai to choose to start a business, bringing more than 30 ai startups to silicon valley. among them, xai has the highest valuation, reaching us$21 billion, and anthropic has a valuation of us$18 billion. can only settle for second place.

ilya sutskever joined the ranks in may this year, creating safe superintelligence (safe superintelligence), which is almost tit-for-tat with sam altman, and received its first financing in september. leading players such as a16z, sequoia capital, and dst have valued ssi at as high as $5 billion.

will mira murati make the same choice? of course, there are no obvious signs yet. it can only be said that since pure r&d personnel like ilya have become bosses, mira murati, who has more experience in engineering management, obviously has a greater chance. it’s quite irritating to think about idealists collectively challenging “teammates who betrayed their ideals”.

however, if the plot really happens like this, perhaps even the most optimistic investors will have to wonder: is our current asset concentration a bit too high?