In May, a group of more than 20 finance whizzes from across the country descended on OpenAI's San Francisco HQ for what CEO Sam Altman cheekily called a "party." They sat through a presentation, mingled with researchers, and some received formal interview invites.
今年 5 月，来自全美的 20 多位金融精英涌入 OpenAI 位于旧金山的总部，参加首席执行官 Sam Altman 打趣所称的“一场派对”。他们听取了演示、与研究人员交流，部分人还收到了正式面试邀请。
A month later in New York City, Altman's team held another recruiting overture for quant trading professionals — the highly coveted mathematicians, physicists, data scientists, and engineers who power the world's top hedge funds and high-frequency trading firms.
一个月后，Altman 的团队在纽约市再次向量化交易专业人士发出招聘邀约——这些炙手可热的数学家、物理学家、数据科学家和工程师为全球顶级对冲基金及高频交易公司提供动力。
Altman's pitch: Forsake Wall Street stalwarts like Citadel, D.E. Shaw, and Jane Street and join his \$300 billion powerhouse's quest to build artificial general intelligence.
Altman 的招募宣言是：离开 Citadel、D.E. Shaw、Jane Street 等华尔街巨头，加入这家市值 3000 亿美元的公司，共同追寻打造通用人工智能的宏图。
Much attention this summer has surrounded Mark Zuckerberg's extravagant recruiting blitz for top AI researchers for Meta. Offers in the tens of millions — in rare cases, north of \$100 million — have stirred up FOMO across Silicon Valley.
今年夏天，Mark Zuckerberg 大手笔为 Meta 招揽顶尖 AI 研究人员备受关注。动辄数千万、个别甚至超过 1 亿美元的报价，让整个硅谷掀起了一股害怕错过的热潮。
AI labs, including OpenAI, Anthropic, xAI, and others, are also looking beyond competitors to staff up, increasingly hunting in Wall Street's backyard.
OpenAI、Anthropic、xAI 等 AI 实验室也在向竞争对手之外拓展招聘视野，越来越多地在华尔街腹地挖掘人才。
Poaching quants for Silicon Valley isn't new. But AI startups awash in cash can not only match but outbid Wall Street pay. Junior and midlevel traders at top high-speed trading firms are now fielding multimillion-dollar packages — up sharply from a year ago, quants and quant recruiters with direct knowledge of the offers told Business Insider.
从硅谷挖角量化人才并不新鲜。但手握巨资的 AI 初创公司不仅能匹配华尔街薪酬，甚至开出更高价。知情的量化从业者和招聘人员告诉《商业内幕》，顶级高频交易公司的初中级交易员如今也在收到数百万美元级别的薪酬方案，较一年前大幅攀升。
OpenAI has scooped up researchers, engineers, and senior recruiters from firms like Hudson River Trading and Citadel Securities. Last year, it hired HRT's longtime HR chief.
OpenAI 已从 Hudson River Trading、Citadel Securities 等公司挖走研究人员、工程师和高级招聘主管。去年，公司还聘用了 HRT 长期任职的人力资源主管。
Wall Street trading firms long vied for and often won over the brightest quantitative minds, dangling \$600,000 comp packages to fresh-faced grads and multimillion-dollar guarantees to seasoned pros.
华尔街交易公司长期争夺并频繁赢得最聪明的量化人才，向应届毕业生开出 60 万美元的薪酬包，向资深专家保证数百万美元的待遇。
Math olympiad champs have long felt the smart money — the low-risk, high-reward wager — was at a top-tier quant hedge fund or prop trading firm. Leaving would mean accepting a comparatively uninspiring Big Tech gig, often working on internal systems like improving advertising algorithms, or a high-risk gamble on a startup in bootstrap mode.
在许多数学奥赛冠军看来，最聪明的选择——低风险高回报——一直是在顶级量化对冲基金或自营交易公司任职。离开那里就意味着要么选择相对平淡的大型科技公司岗位，常常是改进广告算法等内部系统，要么在资金拮据的初创企业里冒高风险一搏。
As some AI valuations soar into the billions, though, the risk calculus is shifting for some quants.
然而，随着部分 AI 公司估值飙升至数十亿美元，一些量化人才的风险收益考量正在改变。
"Optimizing ads at Google gets boring," said Paul Carr, who recruited quants for years as a business development head at prop trading firm Tower Research. "This is different."
“在 Google 优化广告会变得乏味，”曾在 Tower Research 任业务拓展主管、多年负责招募量化人才的 Paul Carr 说，“这可不一样。”
For Wall Street's elite, prior threats from Silicon Valley never amounted to much, said Carr, who recently launched Harchester Research, a data firm that tracks early career AI professionals and researchers in academia who may be desirable hires for trading firms.
Carr 最近创立了数据公司 Harchester Research，追踪早期 AI 职业人士和学术研究者，以供交易公司招聘。他表示，对于华尔街精英来说，过去来自硅谷的威胁从未造成太大影响。
Now, they're on notice.
如今，他们已经警觉。
"If LLM research labs start aggressively poaching from trading firms, it could become a problem," Carr said.
Carr 说：“如果大型语言模型研究实验室开始大举从交易公司挖角，那可能会成为一个问题。”
After seeing Altman advertise his "parties" for quants, Johnny Ho, a cofounder of Perplexity, which operates a popular AI-powered search engine and a newly launched web browser, decided to throw a rival event for quants at an event space on Canal Street during NYC tech week.
看到 Altman 为量化人才宣传他的“派对”后，运营热门 AI 搜索引擎并新推出网页浏览器的 Perplexity 联合创始人 Johnny Ho 决定在纽约科技周期间，于运河街的一处活动空间举办一场针对量化人才的“对标”活动。
Raised in New York, Ho spent five years as a quantitative trader in the city after studying math and computer science and graduating from Harvard in 2017. He knew it could be a fruitful area from which to recruit new talent.
Ho 在纽约长大，自 2017 年自 Harvard 毕业并获得数学与计算机科学学位后，在这座城市担任了五年量化交易员。他深知这将是一个招募新人才的丰饶领域。
"The hot thing to do back in those years was to try and get a trading internship, and if you couldn't do it, you'd go into tech," said Ho, the chief strategy officer of Perplexity, which is valued at \$14 billion and focuses on refining and honing models from other AI labs for greater accuracy and performance.
“在那些年里，最热门的选择就是努力争取交易实习；如果做不到，就去科技行业，”Ho 说。他是估值 140 亿美元的 Perplexity 的首席战略官，该公司专注于从其他 AI 实验室提取、打磨模型，以提升精确度和性能。
"Obviously, the tables have turned now."
“显然，现在形势已经完全颠倒了。”
Altman's call-out to quants
Altman 对量化人才的公开召集
In January, the AI world fell off its axis when DeepSeek, a little-known Chinese firm, vaulted to No. 1 on Apple's US App Store with a free chatbot that rivaled ChatGPT. DeepSeek started as a side project of the hedge fund High-Flyer and was built at a fraction of the cost of mainstream AI labs that were spending billions on computing power. Markets lurched, and AI-related tech stocks plunged.
今年一月，AI 领域发生重大震动：一家名不见经传的中国公司 DeepSeek 凭借与 ChatGPT 匹敌的免费聊天机器人，一跃登上 Apple 美国 App Store 排名第一。DeepSeek 最初是对冲基金 High-Flyer 的副项目，成本仅为主流 AI 实验室斥资数十亿美元算力开销的零头。市场随之剧烈波动，AI 相关科技股大幅下挫。
It was a potent reminder that AI firepower was lurking inside secretive quantitative trading firms. A few months later, Sam Altman took to X, calling on high-frequency trading personnel to escape their "existential dread" of "shaving nanoseconds off latency" or "extracting bps from models," and help him build artificial general intelligence — autonomous AI systems that can outperform humans in a wide range of tasks.
这强烈提醒人们，隐秘的量化交易公司内部潜藏着强大的 AI 实力。几个月后，Sam Altman 在 X 上呼吁高频交易人员摆脱“为缩短延迟数纳秒而感到的存在性恐惧”或“从模型中榨取基点”的困境，加入他，共同打造可以在广泛任务上超越人类的人工通用智能（AGI）。
He followed up with a link to an application for the two recruiting events, with a more direct pitch on the "massive impact" quants could have in making AGI.
随后，他附上了两场招聘活动的申请链接，直接阐明量化人才在构建 AGI 方面可以产生的“巨大影响”。
While the DeepSeek episode amplified efforts to recruit quants, the pursuit was already well underway, industry experts say.
业内专家表示，尽管 DeepSeek 事件加大了对量化人才的招募力度，但这一追逐早已在进行中。
In 2024, OpenAI poached a handful of people from HRT's HR and recruiting teams, including Chief People Officer Julia Villagra. HRT, one of the industry's top proprietary trading firms, produced \$8 billion in net trading revenue in 2024, and Iain Dunning, an ex-DeepMind researcher, has led its AI lab since 2018. Hiring Villagra, who worked at HRT for 15 years, and her team isn't a coincidence, one quant recruiter said, requesting anonymity to protect client relationships.
2024 年，OpenAI 从 HRT 的人力资源与招聘团队挖走了数名人才，包括 Chief People Officer Julia Villagra。HRT 是业内顶尖自营交易公司之一，2024 年净交易收入达 80 亿美元，自 2018 年以来一直由前 DeepMind 研究员 Iain Dunning 领导其 AI 实验室。一位量化招聘人士在匿名接受采访时表示，OpenAI 招募在 HRT 工作长达 15 年的 Villagra 及其团队绝非偶然，此举意在保护客户关系。
"If you have somebody like Julia there, she was part of the whole ascension of Hudson River," he said, adding that she understands the talent flow and who built key systems. (Hudson River Trading did not respond to requests for comment.)
“如果你有像 Julia 这样的人在场，她见证了 Hudson River 的整个崛起过程，”他说，并补充道她了解人才流向以及是谁搭建了关键系统。(Hudson River Trading 未回应置评请求。)
An OpenAI spokesperson said several of its top-performing researchers over the years have had experience in quant finance, showing how successfully that skill set can translate. Chief Research Officer Mark Chen joined OpenAI in 2018 after nearly seven years in quant research roles.
OpenAI 的一位发言人表示，多年来公司表现最出色的几位研究员都拥有量化金融背景，证明这项技能可以成功迁移。首席研究官 Mark Chen 在从事近七年的量化研究工作后，于 2018 年加入 OpenAI。
Noam Brown, a top OpenAI researcher who joined in 2023, worked for a small quant-trading shop after graduating from Rutgers in 2008, "but didn't want my lifetime contribution to humanity to be making equity markets marginally more efficient," he wrote in a LinkedIn post promoting OpenAI's quant recruiting event. "Taking a pay cut to pursue AI research was my best life decision. Today, you don't even need to take a pay cut to do it."
OpenAI 顶尖研究员 Noam Brown 于 2023 年加入公司，他在 2008 年从 Rutgers 毕业后曾在一家小型量化交易公司工作，“但我不想我对人类的终身贡献只是让股票市场效率略有提升，”他在为 OpenAI 量化招聘活动宣传的 LinkedIn 帖子中写道。“为了从事 AI 研究而接受减薪，是我人生中最好的决定。如今，你甚至不必减薪就能做到这一点。”
Behind AI's multimillion-dollar offers
AI 数百万美元薪酬背后的故事
Less than two years ago, some AI labs were offering machine-learning researchers with a couple years of experience up to \$750,000, including equity, according to a researcher who recently left a high-frequency trading firm.
据一位最近离开高频交易公司的研究员透露，不到两年前，一些 AI 实验室就为拥有几年经验的机器学习研究员开出了最高 75 万美元的总薪酬（含股票）。
Those offers have now soared into the millions: The researcher says peers with similar work experience have received job offers from top AI labs ranging from roughly \$1.5 million to \$3 million, including equity.
如今，这些报价已飙升至数百万美元：该研究员表示，具有类似工作经验的同行已从顶级 AI 实验室获得大约 150 万至 300 万美元的录用邀约（含股票）。
The researcher is considering a job at an AI lab — not necessarily for the AGI mission Altman trumpets, but because the economics seem too good to pass up.
这位研究员正在考虑到一家 AI 实验室任职——并不完全是为了 Altman 所宣扬的 AGI 使命，而是因为经济回报好得难以拒绝。
"The offers have come up dramatically," said Matt Moye, a 20-year veteran of quant recruiting and the founder of Monochrome, a search firm.
“报价的确大幅上涨，”拥有 20 年量化招聘经验、猎头公司 Monochrome 创始人 Matt Moye 说。
Four Wall Street recruiters bemoaned shepherding quant candidates to final-stage interviews at top trading firms only to lose out to one of the buzzy top AI startups.
四位华尔街招聘人员抱怨，他们辛苦把量化候选人送到顶级交易公司的最终面试，却被炙手可热的顶级 AI 初创公司横刀夺爱。
One of the recruiters said candidates frequently juggle offers from both sides. In one case, the offers were financially comparable — both low seven figures — but Anthropic offered a more enticing and broader mandate that was tantamount to "Hey, come play in AI land," this person said. (Anthropic did not respond to requests to comment.)
其中一名招聘人员表示，候选人常常同时权衡双方的报价。某次，两份报价在薪酬上大体相当——都是七位数出头——但 Anthropic 提供了更具吸引力且范围更广的职位职责，相当于“嘿，来 AI 世界玩吧”，这位人士说道。(Anthropic 未回应置评请求。)
Another quant headhunter said that while he has firsthand knowledge of several quant candidates getting offers worth millions to leave for AI, he's also aware of tech candidates who have been getting the same — a point Zuckerberg's recent talent raid underscores.
另一位量化猎头表示，他亲眼见到多位量化候选人收到数百万美元的 AI 公司邀约，但他也知道技术候选人同样拿到过这类报价——Mark Zuckerberg 最近的抢人行动正印证了这一点。
"The quants aren't redefining the market; OpenAI redefined it before they went after quants," this headhunter said.
“重新定义市场的不是量化人才，而是 OpenAI 在招揽量化人才之前就已重新定义了市场。”这位猎头说道。
Why AI labs see quants as natural fits
为什么 AI 实验室认为量化交易人才天生适配
There's the obvious overlap in skill set that makes quants attractive to AI labs. Quants already excel at mining vast data sets and engineering systems to unearth insights. As Carr puts it, "In quantitative trading you need to be comfortable conducting research on large quantities of data, then you need to build a system that can take advantage of any insights from that research. Sound familiar?"
量化交易人才之所以受到 AI 实验室青睐，显而易见的原因是两者在技能上高度重叠。量化交易员本就擅长挖掘海量数据并构建工程系统以提炼洞见。正如 Carr 所说：“在量化交易中，你必须熟练处理海量数据的研究，然后搭建一个系统，利用由此得到的洞察。这听起来是不是很熟悉？”
Then there's the high-octane culture.
另一大因素是这两者都拥有高速运转的工作文化。
"The intensity of the work is similar," said Ho, the Perplexity cofounder. "Teams are very tight-knit, working super hard."
Perplexity 联合创始人 Ho 表示：“工作的强度非常相似。团队紧密协作，努力拼搏。”
The breakneck pace of AI fundraising and hiring mirrors the non-stop grind of elite trading desks. One consultant who's worked with both Citadel and OpenAI said, "The culture is not dissimilar. Citadel are famous for moving like a bullet train. Everyone is working every second."
AI 行业在融资和招聘上的疾速步伐，与顶级交易席位的连轴运转如出一辙。一位曾同时为 Citadel 和 OpenAI 提供咨询的顾问表示：“两者的文化并无二致。Citadel 以子弹列车般的速度著称，每个人都在每一秒高强度工作。”
In both worlds, slight edges translate to enormous returns. Trading strategies decay fast, and AI models need constant updates to stay ahead.
在这两个领域，微小的优势都能转化为巨大的回报。交易策略衰减迅速，而 AI 模型也必须持续更新才能保持领先。
"In trading, there is a lot of pressure to win, and you can see results directly tied to P\&L, so in many ways it's always a race," said Moye. "And that's what we're seeing with AI — everyone knows it's a race."
Moye 说：“在交易中，取胜压力巨大，结果直接体现在损益表上，所以在很多方面这始终是一场竞速。AI 领域同样如此——大家都明白这是一场竞赛。”
Another key component that makes quants attractive is the lengthy, ubiquitous noncompete clauses that systematic trading firms put in employees' contracts to prevent them from leaving for rivals with proprietary information. Because industry sit-outs can last two years, traders and researchers looking for a change have used the time to test drive a career in Big Tech. If they don't like it, they can return to finance once their noncompete has lapsed.
另一个吸引量化人才的关键因素是系统化交易公司无处不在的长周期竞业限制条款，旨在防止员工携带专有信息跳槽至竞争对手。由于行业常见的停工期可长达两年，想要转型的交易员和研究员可以利用这段时间尝试进军大型科技公司；如果不满意，待竞业期结束后，他们仍可回到金融业。
Or they've worked in a hot new startup sector — a potentially risky bet. In recent years, that was crypto.
他们也可能投身于炙手可热的新创赛道——这是一种潜在的高风险押注。近几年，这样的赛道是加密货币。
The new batch of AI startups offer a more compelling vision.
而如今，新一波 AI 初创公司则描绘出更具吸引力的愿景。
"The pitch is really net impact on the world," Ho said. "You can continue farming markets for small amounts of edge on every trade. Or you can actually make a difference and make improvements in people's lives."
Ho 表示：“真正的卖点在于对世界的净影响。你可以继续在每一笔交易中榨取微小优势，亦或真正改变世界、改善人们的生活。”
With AI products, "the direct impact on the user is so high," he added.
他补充道，在 AI 产品中，“对用户的直接影响极为显著”。
Don't count Wall Street out
别小看 Wall Street
The industry has been here before.
该行业以前也经历过类似情况。
Back in the early 1990s, long before Renaissance Technologies became the industry's most fabled quant hedge fund, founder Jim Simons was trying to lure a pair of IBM researchers who were working on nascent versions of large language models and speech recognition.
早在 1990 年代初，远在 Renaissance Technologies 成为业界最传奇的量化对冲基金之前，创始人 Jim Simons 就试图招揽两位 IBM 研究员，他们当时正从事大型语言模型和语音识别的早期版本研究。
Bob Mercer and Peter Brown would become invaluable leaders at Renaissance, but Brown, who was the first grad student to study under AI godfather Geoffrey Hinton, was initially uninterested in joining — until Simons offered to double his salary.
Bob Mercer 和 Peter Brown 最终成为 Renaissance 的核心领袖，但 Brown——首位师从 AI 教父 Geoffrey Hinton 的研究生——起初并不想加入，直到 Simons 提出将他的薪水翻倍。
"After that offer, I came home. I took one look at our newborn daughter and realized I had no choice in the matter," Brown, now CEO of Renaissance, said in a podcast interview. "So, the decision to leave computational linguistics for a small hedge fund that no one had ever heard of was made purely for financial reasons."
Brown（现任 Renaissance CEO）在一次播客访谈中说：“在那份报价之后，我回到家，看了看我们刚出生的女儿，意识到我别无选择。所以，放弃计算语言学，加入一个没人听说过的小型对冲基金的决定完全是出于财务原因。”
The anecdote underscores that while mission matters in recruiting, money talks. And it highlights an underappreciated truth about quants: They're often risk-averse. Instead of highly leveraged, gut-wrenching bets, they favor cool calculations and expected-value math.
这一轶事凸显了招聘中使命固然重要，但金钱更具说服力，也揭示了一个常被忽视的事实：量化人才往往厌恶风险。他们不喜欢高杠杆、惊心动魄的赌注，而倾向于冷静运算和期望值数学。
Now, it's firms like OpenAI, Anthropic, and Perplexity — the inheritors of that early IBM work — winning those expected-value calculations.
如今，OpenAI、Anthropic 和 Perplexity 等公司——那些早期 IBM 研究成果的继承者——正在赢得这些期望值计算。
At the moment, these AI giants may be able to outbid trading firms thanks to their lofty valuations. Carr cautions that Wall Street won't sit idle if a talent war erupts.
眼下，由于估值高企，这些 AI 巨头也许能够在出价上压过交易公司。Carr 警告说，如果人才大战爆发，Wall Street 不会坐视不管。
"Trading firms are run by competitive people with a habit of winning," he said. "They're used to resource constraints, and they've never had the option of hurling SoftBank's money at problems."
他表示：“交易公司由习惯于取胜的竞争型人士掌舵。他们习惯于资源受限，从未有机会像 SoftBank 那样用大笔资金砸向问题。”