600亿美元。
8月11日，SEC文件证实SpaceX与Cursor开发商Anysphere签下了约束性收购协议——全股票交易，Cursor估值600亿美元。这个消息在AI圈炸开了锅，因为就在几个月前，Cursor正准备以500亿美元估值从a16z、Thrive和Nvidia那里融资20亿美元。SpaceX直接横插一脚，把这笔交易变成了自己口袋里的一部分。
这不是那种"科技公司收购小工具"的新闻。这是AI编程史上最大的一笔并购——没有之一。更讽刺的是，被收购的那家公司CEO Michael Truell，今年才25岁。2022年创立公司，2026年被600亿美金买走，四年时间，5800倍的增长倍数。如果这个生意做得成，Truell会一下子成为最年轻的亿万富翁之一。
为什么是SpaceX？为什么是Cursor？
SpaceX收购Cursor的逻辑，放在马斯克整个AI版图里看，其实非常清晰。
2026年2月，SpaceX吸收合并了马斯克的另一个AI公司xAI——也就是说，SpaceX现在同时拥有火箭、卫星、AI模型（Grok）和AI编程工具（Cursor，即将并入）。这三者之间的关系是什么？Grok是通用模型，但它在代码生成能力上一直落后于OpenAI和Anthropic；Cursor是增长最快的AI编程工具，年化收入约26亿美元，但Cursor本身没有自己的大模型，用的是OpenAI、Anthropic和Google的API。现在SpaceX一手包办——有了Grok做大模型基座，有了Cursor做编程入口和数据源，有了自己的算力和火箭发射能力做基础设施。
马斯克已经在X平台上说过：Grok的新版本用了大量Cursor数据训练，效果显著改善。这意味着什么？意味着Cursor的用户代码数据、调试日志、交互模式——这些高价值的编程上下文——现在成为了Grok模型的训练燃料。反过来，Grok变强了又提升了Cursor的代码生成质量。这是一个正反馈循环。
"SpaceX收购Cursor不是买了一个代码编辑器，是买了一个实时产生的高质量编程数据集——以及数千万开发者的注意力。"—— 一位科技行业分析师的判断
当然，这笔交易对SpaceX股东来说也是个大胆的下注。Morgan Stanley估计，Cursor在2026年可以为SpaceX贡献约25亿美元收入，到2027年可能达到130亿美元。如果这个预测成立，Cursor的收入规模甚至超过SpaceX火箭业务的大部分年份。一个编程工具的年收入，快比得上一个火箭公司的营收了——这就是AI时代的经济学。
500亿到600亿：一张期权如何改变命运
这个交易最精彩的部分在于它的历史纵深。2026年4月，SpaceX首次披露了一份收购Cursor的期权——给SpaceX一个选择权：可以在未来以600亿美元收购Cursor，或者支付100亿美元分手费放弃收购。当时业界普遍认为这更像是SpaceX在向Cursor施压，逼他们选择合作而不是融资。
Cursor确实正在准备融资。根据此前报道，他们计划以500亿美元估值融资20亿美元，投资方包括a16z、Thrive和Nvidia。这是一个相当合理的估值——Cursor的ARR已经达到26亿美元，PS约19倍，在SaaS领域不算离谱。但SpaceX的期权直接跳过了这一轮融资，把Cursor从"即将融资的高速增长公司"变成了"SpaceX的全资子公司"。
100亿美元的分手费条款很有意思。如果SpaceX最终选择不收购Cursor，他们需要支付100亿——这个金额足够让任何一家公司认真考虑收购条件。可以认为这份期权本身就是一道护城河：它阻止了其他竞购者，也给了SpaceX极大的谈判杠杆。
交易结构方面，SpaceX将通过其子公司X67 Inc.合并进入Cursor，Cursor作为全资子公司存续。Cursor的股份将转换为SpaceX A类股票，价格基于交割前七个交易日的平均股价。交割预计在今年第三季度完成，仍需监管批准。
AI编程格局的彻底重构
SpaceX收购Cursor之后，AI编程工具的市场格局会发生什么变化？
最直接的影响是：Cursor可能会从"多模型可选"转向"优先推荐Grok"。目前Cursor支持OpenAI、Anthropic、Google等多个模型的接入，用户可以根据自己的偏好选择。一旦成为SpaceX子公司，产品路线必然调整——Grok会被深度集成，其他模型的权重可能下降。对于依赖OpenAI或Anthropic模型的用户来说，这意味着什么？意味着他们的"主力编程AI"正在慢慢变成一个竞争对手的产品。
更深远的影响在于数据。Cursor每天处理数百万开发者的代码交互——这是世界上最丰富的编程行为数据集之一。SpaceX拥有这些数据之后，训练出的代码模型将比OpenAI和Anthropic的同类模型更有针对性。这不仅仅是"更多数据"的问题，而是"更贴近真实开发场景的数据"的问题。
对于整个AI编程赛道来说，SpaceX的入局意味着：
第一，竞争维度从模型能力扩展到生态整合。以前比的是谁生成的代码更准确、更快。以后比的是谁能把模型、工具链、算力、数据闭环做完整。Cursor+Grok+Starlink+SpaceX计算的组合，是目前市场上最完整的AI编程生态系统——没有之一。
第二，开源与闭源的博弈加剧。Cursor本质上是闭源的（虽然前端部分开源），而GitHub Copilot也是闭源的。SpaceX收购Cursor后，这两个最大的闭源AI编程工具都掌握在一个公司手里。这对开源编程生态（如Continue、Codeium等）构成了更大的压力。
第三，定价策略可能发生变化。SpaceX不是一家追求短期利润的公司——它的战略逻辑是长期生态控制。这意味着Cursor的定价策略可能不会遵循传统的SaaS涨价路径，而是继续用低价甚至免费策略扩大用户基数，用数据飞轮巩固护城河。
600亿贵不贵？三个维度看
回到估值问题。600亿美元买一个25岁CEO的编程工具公司，贵吗？
从PS（市销率）角度看：Cursor年化收入约26亿美元，600亿估值对应PS约23倍。在SaaS领域，这个估值偏高但不在离谱区间——参考Snowflake上市时的PS超过30倍，Snowflake的收入增速也没有Cursor快。如果Cursor保持当前的增长势头，23倍PS其实是合理的。
从战略协同角度看：Cursor带来的Grok代码能力提升、数据积累、开发者生态锁定，这些价值很难用传统的PS倍数来衡量。SpaceX为这些战略价值支付的溢价，可能比26亿美元的账面收入更有意义。
从历史对比角度看：AI编程工具历史上最大的一笔融资是Cursor在此之前的估值——500亿美元。SpaceX用600亿拿下，说明市场认可这个赛道的稀缺性和成长性。更重要的是，这是AI编程领域有史以来最大的一笔并购，没有 precedent 可以参考——所以600亿这个数字本身，就是在定义新的标准。
"当SpaceX用火箭的钱来买编程工具的时候，你就知道AI时代的权力结构正在发生根本性的转移。"—— 一位硅谷风投合伙人的评论
当然，交易还未完成。监管审批、股东投票、整合规划——每一步都可能出问题。而且SpaceX作为一家上市公司（刚完成史上最大IPO），需要向股东证明这笔600亿的交易物有所值。如果Cursor的收入增长放缓，或者Grok的代码能力没有如预期提升，这笔交易可能会被质疑。
但无论如何，8月11日的SEC文件已经写清楚了：SpaceX要买Cursor，600亿美元，全股票交易。AI编程史上最大的并购，正式进入倒计时。当火箭公司开始收购编程工具公司的时候，我们可以确定一件事：AI的权力结构，正在被重新洗牌。
明天见。
$60 billion.
On August 11, SEC filings confirmed that SpaceX signed a binding acquisition agreement with Cursor developer Anysphere — an all-stock deal valuing Cursor at $60 billion. The news sent shockwaves through the AI industry, because just months earlier, Cursor had been preparing to raise $2 billion at a $50 billion valuation from a16z, Thrive, and NVIDIA. SpaceX stepped in and turned what was supposed to be a funding round into a wholly-owned subsidiary.
This isn't the kind of headline where "a tech company acquires a small tool." This is the largest M&A deal in AI coding history — by far. The more ironic detail? The CEO of the acquired company, Michael Truell, is only 25 years old. Founded the company in 2022, sold it for $60 billion in 2026 — four years, 5,800× growth. If this deal closes, Truell becomes one of the youngest self-made billionaires in history.
Why SpaceX? Why Cursor?
The logic behind SpaceX acquiring Cursor, when placed in Musk's broader AI vision, is actually quite clear.
In February 2026, SpaceX absorbed Musk's other AI company, xAI — meaning SpaceX now simultaneously owns rockets, satellites, an AI model (Grok), and an AI coding tool (Cursor, incoming). What's the relationship between these three? Grok is a general-purpose model, but it has consistently lagged behind OpenAI and Anthropic in code generation capability. Cursor is the fastest-growing AI coding tool, with approximately $2.6 billion in annualized revenue, but Cursor itself doesn't have its own foundation model — it uses APIs from OpenAI, Anthropic, and Google. Now SpaceX controls all three: Grok as the model foundation, Cursor as the coding entry point and data source, and its own compute and rocket launch capability as infrastructure.
Musk has already said on X that newer versions of Grok were trained on a significant amount of Cursor data with noticeable improvements. What does that mean? It means Cursor users' code interactions, debugging logs, and conversation patterns — the richest programming behavior dataset on Earth — are now training fuel for Grok. In reverse, as Grok improves, Cursor's code generation quality rises too. It's a positive feedback loop.
"SpaceX acquiring Cursor isn't buying a code editor — it's buying a real-time, high-quality programming dataset and tens of millions of developers' attention." — A Tech Industry Analyst
That said, this is a bold bet for SpaceX shareholders. Morgan Stanley estimates Cursor could contribute roughly $2.5 billion to SpaceX revenue in 2026 and up to $13 billion by 2027. If those projections hold, Cursor's revenue scale would approach most of SpaceX's rocket business in any given year. A coding tool generating nearly as much as a rocket company — that's the economics of the AI era.
From $50B to $60B: How an Option Changed Everything
The most fascinating part of this deal is its historical depth. In April 2026, SpaceX first disclosed an option to acquire Cursor — giving SpaceX the right either to buy Cursor for $60 billion or pay a $10 billion breakup fee if it walked away. At the time, many in the industry read this as SpaceX pressuring Cursor to choose partnership over fundraising.
Cursor was indeed preparing to raise. According to prior reports, they planned to raise $2 billion at a $50 billion valuation from a16z, Thrive, and NVIDIA — a reasonable valuation: with $2.6B ARR, the PS ratio would be about 19×, not outrageous for SaaS. But SpaceX's option bypassed that entire funding round and turned Cursor from a "high-growth company about to raise" into a "SpaceX wholly-owned subsidiary."
The $10 billion breakup fee clause is interesting. If SpaceX ultimately chooses not to acquire Cursor, it must pay $10 billion — an amount sufficient to make any company carefully consider the acquisition terms. You could say the option itself was a moat: it blocked other bidders and gave SpaceX enormous negotiating leverage.
Structurally, SpaceX will merge into Cursor through its subsidiary X67 Inc., with Cursor surviving as a wholly-owned subsidiary. Cursor shares will convert to SpaceX Class A stock, priced based on the seven-trading-day average before closing. Expected Q3 2026 closing, subject to regulatory approval.
Complete Restructuring of AI Programming
What happens to the AI programming tool landscape after SpaceX acquires Cursor?
The most direct impact: Cursor may shift from "multi-model optional" to "Grok-first." Currently, Cursor supports multiple models — OpenAI, Anthropic, Google — letting users choose their preference. Once a SpaceX subsidiary, product strategy will necessarily pivot: Grok will be deeply integrated, while other models' weighting may decline. For users who rely on OpenAI or Anthropic models, this means their "primary coding AI" is slowly becoming a competitor's product.
The deeper impact lies in data. Cursor processes millions of developer code interactions daily — arguably the richest programming behavior dataset in existence. With this data, SpaceX can train code models far more targeted than what OpenAI and Anthropic can offer. This isn't just "more data" — it's "data closer to real development scenarios."
For the entire AI programming track, SpaceX's entry means:
First, competition shifts from model capability to ecosystem integration. Before, it was about who generates code more accurately and quickly. Now it's about who can build the most complete loop of model + toolchain + compute + data. Cursor + Grok + Starlink + SpaceX compute is currently the most complete AI coding ecosystem on the market — by a wide margin.
Second, open-source vs. closed-source dynamics intensify. Cursor is fundamentally closed-source (though the frontend is partially open), and GitHub Copilot is also closed. With both the largest closed-source AI coding tools in one company's hands, the pressure on open-source alternatives (Continue, Codeium, etc.) increases significantly.
Third, pricing strategy may change. SpaceX isn't a short-term profit-maximizing company — its strategic logic is long-term ecosystem control. This means Cursor's pricing may not follow traditional SaaS escalation paths; instead, it may continue using low-cost or even free strategies to expand its user base, using the data flywheel to deepen its moat.
Is $60B Expensive? Three Dimensions
Back to the valuation question. $60 billion for a 25-year-old's coding tool company — expensive?
By PS (price-to-sales): Cursor's ARR is ~$2.6B, $60B valuation implies a PS of ~23×. In SaaS, this is on the high side but not absurd — Snowflake traded above 30× PS at IPO, and its revenue growth wasn't faster than Cursor's. If Cursor maintains its current trajectory, 23× PS is reasonable.
By strategic synergy: The value of Cursor bringing Grok's code improvement, data accumulation, and developer ecosystem lock-in can't be measured by traditional PS multiples. The premium SpaceX pays for these strategic assets may matter more than the $2.6 billion in book revenue.
By historical comparison: The previous high-water mark for AI coding tool financing was Cursor's own $50B valuation. SpaceX paying $60B confirms the market recognizes the track's scarcity and growth potential. More importantly, this is the largest M&A deal in AI coding history — no precedent exists to compare against. The $60 billion number itself is establishing a new benchmark.
"When a rocket company buys a coding tool, you know the power structure of the AI era is being fundamentally rewritten." — A Silicon Valley VC Partner
Of course, the deal isn't closed yet. Regulatory approval, shareholder votes, integration planning — every step could run into problems. And as a newly public company (having just completed the largest IPO in history), SpaceX needs to justify this $60B bet to shareholders. If Cursor's revenue growth slows or Grok's code capability doesn't improve as expected, the deal could face scrutiny.
But regardless, the August 11 SEC filing is clear: SpaceX is buying Cursor, $60 billion, all-stock. The largest M&A in AI coding history is officially counting down. When a rocket company starts acquiring coding tools, one thing is certain: the power structure of AI is being reshuffled.
See you tomorrow.
