OpenAI's $1.4 trillion ask would mark up its biggest backers' stakes by about two thirds

"Nearly double" depends on which earlier number you pick. Calcalist puts the last round at $110 billion on a $730 billion pre-money valuation in February ^[3]. Quartz describes a $122 billion round that closed in March at $852 billion post-money ^[5]. The two accounts fit: $730 billion plus $122 billion is $852 billion ^[1]. That is consistent with a February round that grew by $12 billion before it closed, though none of the reports says so. Pre-money to pre-money, $1.4 trillion is 92% higher ^[2]. Against the March post-money it is about 64% higher ^[6].

The new money is small next to the price. Quartz describes the $1.4 trillion as pre-money ^[2]. On that basis, $30 billion buys about 2.1% of a company worth roughly $1.43 trillion once the cash arrives ^[3]. The whole round is the size of the single check SoftBank wrote in February, and about a quarter of the March close ^[4]. Eighteen months ago OpenAI raised more, $40 billion, at a $300 billion valuation ^[7].

Holders of the other 98% feel the price most. SoftBank and Nvidia put in $30 billion each and Amazon $50 billion in the earlier round ^[4]. Marked from $852 billion post-money to $1.4 trillion pre-money, each $30 billion stake would be carried at about $49 billion and Amazon's at about $82 billion ^[5], before any dilution from shares issued in between. None of that is cash. The reports do not say whether any of the three is investing again.

Revenue under the price is in dispute. Reuters reported annualized recurring revenue approaching $70 billion, with enterprise sales more than doubling since July ^[8]. Bloomberg put annualized revenue past $40 billion, up 70% since July ^[9]. At $1.4 trillion, the price is 20 times one run rate or 35 times the other ^[6]. Calcalist notes that a run rate is not revenue actually generated over a full year ^[10]. OpenAI is also working on price: it trimmed usage allowances on its $200 plan and added a $500 tier ^[11].

Earlier this year OpenAI missed its own targets for revenue and new users, and CFO Sarah Friar warned colleagues that sluggish revenue growth could leave it unable to honor future computing contracts, The Wall Street Journal reported ^[12]. Sam Altman said going public in 2026 would be "ill-advised" ^[13]. The listing is now 2027 at the earliest, despite a confidential filing with the SEC in June ^[14]. Bloomberg describes the round as runway while the IPO waits ^[15].

Calcalist renders the Bloomberg target as up to $30 billion ^[1]. Quartz says at least $30 billion, in talks that are early and could change ^[16], and one person told Bloomberg that investors are driving the interest ^[17]. Bloomberg had earlier reported a round under consideration at $1.2 trillion ^[18]. The new price would pass Anthropic's latest private valuation, according to Bloomberg ^[20], but Anthropic plans an IPO this year that Cointelegraph says could value it above $2 trillion ^[19].

I think the $1.4 trillion mostly measures how little OpenAI stock is for sale. It is a 2% slice priced at 20 to 35 times a run rate, sold while the company puts off the market that would price all of it every day. The counter-case is Friar's warning. If revenue cannot cover the computing contracts, the $30 billion is a funding need and the mark is secondary. The view is wrong if the round closes well above $30 billion at this price. Buyers taking a large block at $1.4 trillion would be paying for the business itself.