一轮融资，两次砍价
据彭博和 Financial Times 报道，OpenAI 本轮融资目标从 300 亿美元缩至约 200 亿美元，缩水 33%，估值目标从 1.4 万亿降至 1.25 万亿美…
这不是一个突兀的决策。砍价分两步走。
第一步发生在估值。
9 月底，OpenAI 对外释放的信号是：以 1.4 万亿美元投前估值融资 300 亿美元，并把这个定位成明年 IPO 前最后的上车机会。当时 SF 科技基金 Dragoneer 拿着这个报价去推私人财富客户。反应如何？知情人士原话是「完全没兴趣」。
估值先让了一步。10 月 22 日 Bloomberg 确认谈判进入尾声时，估值已经降到 1.25 万亿美元。
第二步发生在金额。
估值调完之后，认购承诺没有跟上预期。部分投资者削减了出资规模，截止日到期的首批认购资金预计不足 200 亿美元。规模从 300 亿缩到约 200 亿，缩水三分之一。
更狠的是，这次砍价发生在意向书阶段之后、正式签署之前。这意味着谈判桌上已经出现了一次明确的重新定价。
私人财富客户说没兴趣的时候，场面已经很清楚了
1.25万亿到底值不值
1.25 万亿美元这个数，听起来天文数字，放到具体语境里再算一遍。
OpenAI 今年 3 月的上一轮融资是 1220 亿美元，投后估值 8520 亿美元。半年多时间，估值从 8520 亿跳到 1.25 万亿，涨幅约 47%。
这 47% 靠什么支撑？OpenAI 的说法是 7 月以来营收快速增长。这个理由够不够硬，取决于你用什么口径看。
47%的涨幅，投资人要看到可核验的数据
横向对比一下可比公司的定价。
Meta 本周自由现金流转负，CDS 创纪录；CoreWeave 这样的纯算力公司股价已经从高点大幅回撤。二级市场对这些 AI 重资产公司的定价逻辑，已经从「奖励大额投入」转向追问自由现金流和债务风险。
一级市场的私募估值没有每天报价，它参照的是可比上市公司的价格。当二级市场价格下行时，一级市场不可能长期维持独立定价。
这就是这次砍价背后的核心机制：一级市场开始向二级市场的定价靠拢。OpenAI 只是第一个明确降价的超级标的。
纵向对比同赛道的竞争者。
Anthropic 本周提交 IPO 注册声明，最快 11 月中旬上市，市场定价区间预期从约 2 万亿降到 1.5 万亿。对手上市估值下移，OpenAI 这轮融资的定价空间同时被压缩。
融资窗口与IPO时间线
这轮融资最准确的定性是「opportunistic bridge」——为可能到来的 IPO 铺路的过渡性安排。
OpenAI 在 6 月秘密向 SEC 提交了招股书，8 月 CFO Sarah Friar 在内部告诉员工公司将在 2027 年上市。紧接着 9 月，Sam Altman 改口说现在上市不是一个明智的时机。
同一家公司，两套时间表，指向同一个判断：私有市场给的价格比公开市场目前愿意给的价格更高。
这也是为什么估值即使下调到 1.25 万亿仍高于半年前的 8520 亿——私有市场在用溢价换取流动性窗口，而公开市场的 AI 板块正在经历估值重塑。
私有定价和公开定价之间的套利窗口
对AI基础设施的传导链条
融资收缩意味着什么？逻辑链很直接。
模型公司融资承压，新增算力采购就会更谨慎。下游扩张一旦放慢，上游已经投入的机房、设备和利息不会跟着停下来。
尤其对那些客户集中、负债高、还在继续借钱扩建的算力公司来说，订单签了多少只是第一步。客户能否按期付款、收入能否覆盖折旧和利息，才决定项目能不能撑下去。
这不是危言耸听。业内已有算力公司开始面临客户账期延长、收入确认延迟的问题。如果 OpenAI 这类头部玩家的资本开支节奏放缓，传导到上游会有滞后效应，但不会消失。
一句话结论
连 OpenAI 都需要调低价格、缩小融资规模，靠持续融资支撑算力扩张的模式确实该重新审视了。一级市场向二级市场定价看齐的信号已经明确——下一个降价的不一定是 OpenAI，但一定会有人跟进。
这不是简单的数字游戏。背后是科技股遭遇猛烈抛售、市场对 AI 巨额资本开支回报率的重新评估，以及一级市场定价逻辑向二级市场靠拢的信号。
从300亿到200亿的缩水过程
更狠的是，这次砍价不是同时发生的——是先砍估值，再砍规模。
10月22日，彭博先报道估值从1.4万亿降到1.25万亿。一周后，10月31日《金融时报》确认融资规模也从300亿缩到约200亿。两个动作间隔不到十天，说明投资者对价格的抵触，比预期更强烈。
认购截止日是10月30日。截止日一过，首批到账资金预计不足200亿美元，因为部分投资者削减了认购额。
这轮融资原本被称为"上市前最后上车机会"。OpenAI 6月秘密提交了IPO文件，8月CFO Sarah Friar向员工透露上市窗口预计在2027年。9月Sam Altman又改口说现在上市不是明智时机。同一家公司，两套时间表，指向同一个判断：私有市场给出的价格，比公开市场目前愿意给的价格更高。
Coatue 洽谈联合领投，阿联酋多家投资基金最高拟投100亿美元，贝莱德也在洽谈。有消息称部分出资与一份为 OpenAI 追加算力供应的协议绑定。这不是纯粹的股权交易，更像是有附加条件的过桥融资——业内称为 opportunistic bridge。
估值为什么从1.4万亿掉到1.25万亿
1500亿的估值折让，对应着10月初爆掉的一个财务口径。
OpenAI 以7月以来营收快速增长为由，将估值从8520亿美元近乎翻倍。但彭博随后援引知情人士报道，OpenAI 的真实营收比内部宣传少了约200亿美元。这个数字直接影响了私募定价。
机制是这样运作的：私募估值本身没有每日报价，它要参照可比上市公司。本周 Meta 自由现金流转负、CDS 创纪录，CoreWeave 股价暴跌，整个 AI 基础设施板块都在重估。一级市场开始向二级市场的定价逻辑靠拢，OpenAI 只是第一个明确降价的超级标的。
一级市场向二级市场定价靠拢的传导路径
从3月那轮8520亿美元投后估值算起，即使下调后的1.25万亿投前，新估值仍比半年前高约47%。这个数字意味着什么：OpenAI 需要用远超当前营收增速的资本投入，来支撑账面估值的增长逻辑。
更准确地说，如果融资节奏跟不上资本开支的速度，每年能省下一个亿的利息负担。
这个数字不是拍脑袋定的。降价发生在认购意向书签署之后、正式交割之前——说明私人财富渠道的买家在 1.4 万亿的报价面前，把脚踩住了。
私人财富客户的拒绝
摩根士丹利与旧金山科技基金 Dragoneer 在 9 月底面向私人财富客户做推介时，用的就是 1.4 万亿报价。参与推介流程的知情人士反馈很直接：价格偏高。
这不是第一次有人公开表示阻力。早在 10 月初就有报道指出，OpenAI 把估值从 1.4 万亿下调到 1.25 万亿，正是因为私人财富客户的认购意愿低于预期。
认购截止日是 10 月 30 日。截止日一过，首批到账资金确认不足 200 亿美元。规模从 300 亿降到 200 亿，是投资者削减承诺额的直接结果，而不是公司主动放弃需求。
更准确地说，这轮调降分两步走。先降估值，再缩规模。一级市场定价对二级市场波动越来越敏感，这条传导链在 OpenAI 身上最先被验证。
领投方与潜在出资结构
领投方浮出水面。据 CNBC 和 Bloomberg 报道，Coatue 正洽谈联合领投位置，阿联酋的多家投资基金组成组团，最高可投 100 亿美元，BlackRock 也在参与谈判。
这里有个关键细节：阿联酋资金的进入并非纯粹的财务投资。多位知情人士指出，这轮融资可能与一份追加算力供给的商业协议挂钩。换句话说，钱不是白给的，后面绑着一张算力采购单。
OpenAI本轮融资出资结构示意
Coatue 作为长期深耕科技赛道的基金，愿意在联合领投位置上谈，说明这个价格区间至少在部分机构眼里是可接受的。但「谈」这个字很关键，意味着条款还在博弈，尚未落定。
阿联酋基金的 100 亿是最确定的部分，其余靠 Coating 和贝莱德补位。这个结构决定了即使以 1.25 万亿估值落地，融资规模也可能不到最初目标的 70%。
融资与算力协议的捆绑
这笔钱的性质比表面数字更重要。报道明确使用了「opportunistic bridge」这个表述——过桥融资，为 IPO 铺路。
OpenAI 在 6 月已经秘密向 SEC 提交了招股书，8 月 CFO Sarah Friar 内部告知员工公司将在 2027 年上市。但紧接着 9 月 Sam Altman 又改口称现在上市「不是一个明智的时机」。
两套时间表指向同一个判断：私有市场愿意给出的价格，比公开市场当前愿意给的价格更高。过桥融资的本质就是用更高估值换时间，等二级市场定价收敛。
阿联酋资金的绑定协议是一个信号。算力供应合同意味着这笔钱有一部分将直接转化为基础设施投入，而不是纯粹的股权溢价。
更狠的是，如果这笔算力协议的价格条款不够优惠，后续融资成本会被进一步放大。
这轮交易对AI基建的传导
融资规模缩水 33% 的连锁反应不止于 OpenAI 自身。
模型公司收紧采购预算 → 上游算力公司新增订单放缓 → 已经投入的机房、设备和债务利息不会跟着停下来。
尤其是那些客户集中度高、负债率高、还在继续借债扩建的新云服务公司。订单签了多少只是第一步，客户能否按期付款、收入能否覆盖折旧和利息，才决定项目能不能撑下去。
马斯克此前预估将有 200 多家新云服务公司被整合或淘汰。这一轮融资调整，加速了那个判断的兑现节奏。
私募估值的参照锚
私募估值本身没有每日报价，它要找一个参照锚。这个锚通常是可比上市公司的市值，再叠一层流动性折价。
Meta 自由现金流转负、CDS 价差创纪录，CoreWeave 订单节奏放缓——这些信号不会自动传导到一级市场，但只要锚点在动，折价率就会重新算一遍。
这次下调最准确的说法不是「估值泡沫破了」，而是一级市场开始向二级市场的定价看齐。更狠的是，每年能省下的不是几个亿，而是一整条靠融资滚算力的增长曲线。
估值系统过载
Anthropic上市：竞争者也是定价者
Anthropic 预计11月中旬上市，市场给的定价区间在1.5万亿到2万亿美元之间。这个数字落下来的时候，OpenAI 的估值锚点也跟着晃了一下。
两家公司的业务高度重叠，用户迁移成本极低，投资者很难在两个标的之间做估值区分。Anthropic 上市后如果表现不及预期，OpenAI 的融资窗口只会更紧。
``mermaid
OpenAI vs Anthropic 估值传导链路
Anthropic 不只是竞争对手，更是定价者。当市场上同时出现两个万亿级AI标的，资本必须在两者之间做选择，这就意味着谁先上市、谁先给出业绩，谁就有机会定下整个板块的价格基准。
从IPO前最后机会到2027年
这轮融资原本被定位为「上市前最后上车机会」。OpenAI CFO Sarah Friar 此前告诉员工，IPO窗口预计在2027年。
但Sam Altman 在9月改口说现在上市不是一个明智的时机。同一家公司，两套时间表，指向的是同一个判断：私有市场给出的价格比公开市场目前愿意给的价格更高。
推迟IPO不是信心不足，是定价权还在私有市场手里。等到2027年，要么公开市场接受了AI支出的长期回报率假设，要么OpenAI需要用更低的价格去公开市场找接盘方。
推迟的风险在于， Anthropic 先在公开市场定价，把整个板块的估值天花板钉死，到时候OpenAI反而失去了溢价空间。
``mermaid
IPO时间表选择的两种路径
算力公司的账本
模型公司融资承压，新增算力采购就可能更谨慎。下游扩张一旦放慢，上游已经投入的机房、设备和利息，却不会跟着停下来。
尤其是对那些客户集中、负债高、还得继续借钱扩建的算力公司——订单签了多少只是第一步，客户能否按期付款、收入能否覆盖折旧和利息，才决定项目能不能撑下去。
OpenAI 这轮融资从300亿砍到200亿，哪怕只是意向阶段的信息泄露，也足够让几家依赖头部客户的云服务商重新评估自己的营收预期。
参考文献
[1] 1.4 万亿没谈拢，OpenAI 把新一轮融资目标砍到约 200 亿美元 - Net.Coffee. ip.net.coffee/gpt/news/20… [2] OpenAI 300 亿美元融资估值下调：从 1.4 万亿降到 1.25 万亿 - AI工具导航. toolnavs.com/zh/article/… [3] OpenAI 砍掉 100 亿美元融资目标：估值降至 1.25 万亿，华尔街开始用现金流定价 - 玉米小站. www.yumiok.com/archives/69… [4] 一片下跌声中，OpenAI大幅削减融资金额，从300亿降到200亿 - 华尔街见闻. wallstreetcn.com/articles/37… [5] OpenAI融资目标骤降100亿美元，估值同步下调至1.25万亿美元 — Noticias de TradingView. es.tradingview.com/news/panews… [6] 老猿说OldApe on X: "10 月 20 日我们写过，OpenAI 在推进一轮约 300 亿美元的融资，当时谈的估值可能超过 1.4 万亿美元。三天后，这个数字降了下来。 据彭博援引知情人士，OpenAI 在谈判最后阶段把这轮融资的目标估值定在 1.25 万亿美元，比此前讨论的 1.4 万亿美元低了 1500 亿美元；融资规模仍是约 300 亿美元。CNBC 较早报道了 1.25 万亿美元这一估值。据彭博，Coatue 正洽谈联合领投，阿联酋的多家投资基金和贝莱德等也在谈，预计下周末前敲定投资承诺；谈判仍在进行，细节可能变化。 作为对照，据彭博，OpenAI 今年 3 月完成了… / X. x.com/OldApeTalk/… [7] 老猿说OldApe on X: "10 月 20 日我们写过，OpenAI 在推进一轮约 300 亿美元的融资，当时谈的估值可能超过 1.4 万亿美元。三天后，这个数字降了下来。 作为对照，据彭博，OpenAI 今年 3 月完成了… / X. x.com/OldApeTalk/… [8] OpenAI 拟融资 300 亿美元：1.25 万亿估值背后的算力账单|AI自学笔记. fluxbbs.com/openai-30b-…
延伸入口
- 原文归档：tobemagic.github.io/ai-magician…
- 公众号：计算机魔术师
本文由 AI图鉴自动抓取并翻译，版权归原作者所有；如需查看原文请点下方链接。
A round of financing, two cuts.
According to Bloomberg and Financial Times, the current financing target for Open AI has been reduced from $30 billion to about $20 billion, with 33 per cent shrinking and a valuation target from 1.4 trillion to $1.25 trillion...
This is not a sudden decision. Cut the price in two steps.
The first step took place in the valuation.
At the end of September, Open AI released the signal that it would finance $30 billion with a pre-investment valuation of 1.4 trillion dollars and position this as the last opportunity to get on board before next year ' s IPO. At the time SF Science and Technology Foundation Dragoneer took this offer to push private wealth clients. How'd it go? The source of information says "no interest at all".
The valuation was preceded by a step. On 22 October, Bloomberg confirmed that the valuation had fallen to $1.25 trillion at the end of the negotiations.
The second step is in the amount.
After the valuation was revised, the commitment did not keep up with expectations. Some investors reduced the scale of their contributions and the first subscriptions due by the cut-off date were expected to be less than $20 billion. The scale has shrunk from 30 billion to about 20 billion, a third.
Worse still, the cut occurred after the letter of intent phase and before it was formally signed. This means that there has been a clear re-pricing at the negotiating table.
When private wealth clients say they're not interested, the scene is clear.
$1.25 trillion is not worth it.
The sum of $1.25 trillion sounds astronomical and goes back to the context.
Open AI ' s last round of financing in March this year was $12.2 billion, with a post-investment valuation of $85.2 billion. Over more than six months, the valuation jumped from 85.2 billion to 12.5 trillion, an increase of approximately 47 per cent.
This, 47%, on what? Open AI says there has been a rapid increase in revenue since July. That's hard enough, depending on what you're looking at.
47% increase, investors need to see verifiable data.
A horizontal comparison of prices of comparable companies.
Meta's free cash flow has turned negative this week, with a record CDS; pure-calculations like CoreWeave's stock price has been withdrawn from high points. The pricing logic of these AI heavy asset companies in secondary markets has shifted from "rewarding large investments" to asking about free cash flows and debt risks.
Private valuations for the first-tier market do not offer daily quotations and are based on prices comparable to those of listed companies. When secondary market prices go down, it is not possible to maintain independent pricing in the first market over the long term.
This is the central mechanism behind this cut: first-tier markets are beginning to converge towards second-tier market pricing. Open AI is just the first super-standard to make a clear price reduction.
Longitudinal comparison of competitors on the same track.
Anthropic submitted an IPO registration declaration this week, which was listed at the earliest in mid-November, with market pricing areas expected to drop from about 2 trillion to 1.5 trillion. The hands market valuation moved down, and Open AI ' s pricing space for this round of finance was also compressed.
Finance window and IPO time line
The most accurate characterization of this round of financing is “opportunistic Bridge”, a transitional arrangement to pave the way for possible IPOs.
In June, Open AI filed a secret equity book with the SEC, and in August, CFO Sarah Friar informed the employee that it would be listed in 2027. And then in September, Sam Altman changed to say that it's not a good time to be listed.
The same company, with two sets of schedules, points to the same judgement: private markets offer higher prices than the open market currently offers.
This is why the valuation, even down to 125 trillion, is still higher than the 852.2 billion that was six months ago — private markets are trading premiums for liquidity windows, while open-market AI blocks are undergoing valuation re-engineering.
arbitrage window between private and open pricing
Transmitting chain to AI infrastructure
What does financial contraction mean? The logical chain is straightforward.
Model companies are under pressure to finance, and new calculus purchases are more prudent. When downstream expansion slows down, the aircraft rooms, equipment and interest already invested upstream do not stop.
In particular, the number of purchase orders signed is only the first step for those powerful companies that have concentrated clients, are highly indebted and continue to borrow money to expand. The sustainability of the project is determined by whether the client is able to pay on time, whether the income covers depreciation and interest.
It's not alarmist. Business-based firms have begun to face problems with extended customer accounting periods and delays in revenue recognition. If the pace of capital spending by such head players as Open AI slows, the carry-over effect will be delayed but will not disappear.
One conclusion:
Even Open AI needs to lower prices and scale down finance, and it is time to re-examine the model that sustains the expansion of computing power with continued financing. The signal that first-tier markets are priced in line with secondary markets is clear — the next price reduction is not necessarily Open AI, but someone will follow.
It's not a simple number game. Behind these are the strong sales of the technology unit, the market's reassessment of the return on AI's huge capital expenditure and the signal that the first-market pricing logic converges with the second-tier market.
Between 30 billion and 20 billion.
Worse still, this cut-off did not take place at the same time — first, in valuation, then in scale.
On 22 October, Bloomberg first reported that the valuation had fallen from $1.4 trillion to $1.25 trillion. A week later, the Financial Times confirmed that the scale of financing had also shrunk from $30 billion to about $20 billion on 31 October. The fact that the two movements were less than 10 days apart suggested that investors were in conflict with prices, which was stronger than expected.
The date of acceptance is 30 October. As of the end of the day, the initial funds received were expected to be less than $20 billion, as some investors reduced their subscriptions.
This round of finance was originally called "the last chance to get in the car before the market." Open AI submitted the IPO document in secret in June, and CFO Sarah Friar disclosed to its employees in August the listed window expected in 2027. In September, Sam Altman changed to say that it was not the right time to be listed. The same company, with two sets of schedules, points to the same judgement: the price given by the private market is higher than the price currently offered by the open market.
The Coatue negotiations led to a proposed investment of up to $10 billion in several United Arab Emirates investment funds, and were being negotiated by Belet. There are reports that part of the funding was tied to an agreement to supply additional computing power to Open AI. This is not a purely equity deal, more like conditional bridge financing — what the industry calls "opportunistic Bridge."
Why did the valuation fall from 1.4 trillion to 12.5 trillion?
The $150 billion valuation concession corresponds to a financial calibre that exploded in early October.
Open AI almost doubled the estimate from $85.2 billion, citing the rapid growth in revenue collection since July. But Bloomberg then quoted informed sources to report that Open AI's real profits were about $20 billion less than internal publicity. This figure has a direct impact on private fundraising pricing.
The mechanism operates as follows: the private collection valuation itself does not have a daily quote, and it refers to comparable listed companies. This week Meta Free Cash Flows turned negative, the CDS hit record, the CoreWeave stock fell, and the entire AI infrastructure block is being revalued. Level-I markets are beginning to converge with the price-fixing logic of secondary markets, and Open AI is only the first super-standard to clearly reduce prices.
Guided path of first-tier market towards second-tier market pricing
From the March round of $852 billion, the new valuation was approximately 47 per cent higher than it had been six months earlier, even before the downturn of $1.25 trillion. What this figure means is that Open AI needs to support the growth logic of book valuation with capital inputs that are much faster than current base rates.
More precisely, if the pace of financing does not keep pace with capital spending, it will save up to a billion interest burdens a year.
That's not the number for the headshot. The price reduction occurred after the signing of the letter of intent to purchase and before the formal delivery - the buyer indicating the channel of private wealth stepped on his foot in front of a 1.4 trillion offer.
The refusal of private wealth clients
Morgan Stanley and the San Francisco Science and Technology Foundation Dragoneer used 1.4 trillion quotations at the end of September for their presentation to private wealth clients. The feedback from informed participants in the referral process is direct: high prices.
This is not the first time that resistance has been publicly expressed. As reported as early as October, Open AI reduced the valuation from 1.4 trillion to 12.5 trillion, precisely because private wealth clients were less willing to subscribe than expected.
The subscription deadline is October 30. As at the end of the day, the first funds received were recognized to be less than $20 billion. The fall in size from $30 billion to $20 billion is a direct result of the reduction of commitments by investors rather than the voluntary abandonment of demand by companies.
More precisely, the rotation was reduced by two steps. Lower valuation before scaling down. Level-I market pricing is becoming increasingly sensitive to secondary market fluctuations, and this transmission chain is first validated on Open AI.
Lead investors and potential funding structures
The lead party surfaced. According to CNBC and Bloomberg, Coatue is in the process of negotiating a joint lead position, with several investment funds in the United Arab Emirates forming a consortium that can invest up to $10 billion, and BlackRock is also participating in the negotiations.
Here is a key detail: the entry of UAE funds is not a purely financial investment. According to several sources, this round of financing may be linked to a commercial agreement for additional calculus provision. In other words, the money was not paid for, and a credit card was attached.
Indicative funding structure of Open AI for the current round
As a fund for the long-term deep-tilled technology track, Coatue is willing to talk about the joint lead position, suggesting that this price range is acceptable to at least some agencies. But the word "talk" is crucial, meaning that the terms are still in play and have not yet been finalized.
The 10 billion United Arab Emirates funds are the most identified component, with Coating and Belet being replaced. This structure determines the scale of the financing that may be less than 70 per cent of the original target, even if it were to land with a 1.2 trillion valuation.
Financing and the binding of power agreements
The nature of the money is more important than the apparent number. The report explicitly uses the expression “opportunistic Bridge” - bridge financing to pave the way for IPO.
In June, Open AI submitted a secret offer to the SEC, and in August, CFO Sarah Friar informed the employee that it would be listed in 2027. But then in September, Sam Altman changed his name to "it's not a wise time."
The two sets of schedules point to the same judgement: the price the private market is willing to offer is higher than the price currently offered in the open market. The essence of cross-bridge financing is to switch time to higher valuations, and wait for secondary market pricing to shrink.
The United Arab Emirates funding binding agreement was a signal. Computer supply contracts mean that part of the money will be converted directly into infrastructure inputs rather than a purely equity premium.
Worse still, the cost of follow-up financing would be further magnified if the price terms of the power agreement were not sufficiently favourable.
This round of transactions is a guide to AI infrastructure.
The funding scale shrinks by 33% and the chain reaction is more than Open AI itself.
The model company tightens the procurement budget and slows down the new orders of upstream computing firms and does not stop interest on aircraft, equipment and debt already invested.
In particular, new cloud service companies with high customer concentration, high debt ratios and continued borrowing and expansion. The amount of purchase orders issued is only the first step in determining the sustainability of the project if the client is able to pay on schedule, if the income covers depreciation and interest.
Mask predicted that more than 200 new cloud service companies would be integrated or eliminated. This round of financing adjustments has accelerated the pace of delivery of that judgement.
Private valuation reference anchor
The private valuation itself did not offer a daily quote and it sought a reference anchor. This anchor is usually more than the market value of the listed company, with an additional discount on liquidity.
Meta Free Cash Flows, the record spread of CDS, and the slow pace of CoreWeave Orders — these signals are not automatically transmitted to the first market, but the discount rate is recalculated as long as the anchor is moving.
The most accurate version of this downward adjustment is not "valuing bubble broke", but rather that the first-tier market has begun to align with the second-tier market. Worse still, it is not hundreds of millions of dollars that can be saved each year, but a whole rolling growth curve based on financing.
Overloading of valuation system
Anthropic listing: competitors are pricers too
Anthropic expects to be on the market by mid-November, with market-based pricing ranges between $1.5 trillion and $2 trillion. When that figure fell, Open AI's valuation anchor went off.
The high level of overlap between the two companies and the extremely low cost of user migration made it difficult for investors to make a valuation distinction between the two targets. Open AI's financing window will only be tighter if it fails to meet expectations when Anthropic is listed.
``mermaid
Open AI vs Anthropic Valuation Transmission
Anthropic is not just a rival, but a pricer. When two trillion-class AI targets appear in the market at the same time, capital must choose between the two, which means that whoever is listed first and who gives performance first has the opportunity to benchmark the price of the entire block.
From last chance before IPO to 2027
This round of financing was originally identified as "the last chance to get on the bus before the market". Open AI CFO Sarah Friar had previously told employees that the IPO window was expected in 2027.
But Sam Altman changed his voice in September saying that it's not a wise time to be listed. The same company, with two sets of schedules, points to the same judgement: private markets offer higher prices than are currently offered by open markets.
Delaying IPOs is not a lack of confidence, but rather the fact that pricing rights are still in the hands of the private market. By 2027, either the open market accepts the long-term return assumption for AI expenditures or Open AI needs to find a match in the open market at lower prices.
The risk of delay is that Anthropic first priced the entire valuation ceiling on the open market, and then Open AI lost the margin space.
``mermaid
Two paths selected for the IPO timetable
The books of the firm.
Model companies may be more cautious when under pressure to finance new calculators. When downstream expansion slows down, the aircraft rooms, equipment and interest already invested upstream do not stop.
In particular, those companies that are focused, highly indebted and have to continue to borrow money to expand -- • The number of orders signed is only the first step in determining whether the project will survive, whether the client will be able to pay on time, whether the income will cover depreciation and interest.
Open AI, from $30 billion to $20 billion in this round of financing, would be enough to allow cloud service providers who depend on their head clients to reassess their collection expectations, even if the information leaked at the indicative stage.
References
[1] 1.4 trillion dollars is missing, and Open AI has cut the new financing target to about $20 billion - Net.Coffee. ip.net.coffee/gpt/news/20.[2] Open AI's $30 billion financing valuation is down: from 1.4 trillion to 1.2 trillion - AI's navigation of tools. Three days later, that figure dropped. According to information sources quoted by Bloomberg, in the final phase of the negotiations, Open AI had set the target for this round of financing at $1.25 trillion, $150 billion below the $1.4 trillion previously discussed; the scale of financing remained at about $30 billion. CNBC had earlier reported the valuation of $1.25 trillion. According to Bloomberg, Coatue is negotiating joint offers, as are several investment funds in the United Arab Emirates and Belet, which are expected to finalize investment commitments by the end of next week; negotiations are ongoing and the details may change. By contrast, according to Bloomberg, Open AI completed.../X.x.com/OldApeTalk/...[7] The apes said OldApe on X: "On October 20, we wrote that Open AI was pushing for a round of about $30 billion in financing, and the valuations were likely to exceed $1.4 trillion. Three days later, that figure dropped. By contrast, according to Bloomberg, Open AI completed in March this year... / X. x.com/OldApeTalk/...[8] Open AI proposed to finance $30 billion: 125 trillion in self-study notes.
Extension entrance
- Original archive: tobemagic.github.io/ai-magician...
- WeChat Official Account: Computer magician
Auto-fetched and translated by AI Atlas. All rights belong to the original author.