TIME
1993年5月，徐直军即将完成博士学业，到深圳找工作。他在朋友家翻看《深圳特区报》（Shenzhen Special Zone Daily）时，看到华为的一则整版招聘广告。广告特别写明要招硕士和博士。
徐直军认为这有点反常。深圳当时是中国改革开放的前沿，也是加工贸易的聚集地，很少有企业招高学历人才。出于好奇，他第二天就去面试，并拿到了录用通知。
临近毕业时，他手里已有好几个录用通知。华为的人力资源经理每周都给他打电话，说一切准备就绪，就等他过去。华为承诺的薪水是他博士生导师的5倍。他说，这份尊重和急切让选择毫无悬念，他选择了华为。
他8月入职后发现，公司600名员工中有200多人是研发工程师，还有一大批研究生。同年入职的博士有5位。当时全国每年毕业的博士只有约2000人。他认为，华为在创业初期就在人才上下重注，表明了它对创新的执着和面向未来的投入。
当时研发团队在做数字程控交换机。中国经济高速增长，家庭和办公电话需求远超供给，装一部电话既难又贵。西方厂商的交换机在中国也能买到，但标准五花八门、互不兼容，每线价格约3000元人民币，在当时是一笔巨款。
中心城市周边散布着大量乡镇和村庄。每个点都建一套完整交换机不划算，用铜缆连接偏远地区成本高，还容易造成信号衰减。华为因此研发出光纤连接的远端交换模块，用一套集中式系统为整个县提供网络覆盖。
客户认可这一方案。华为由此在与西方老牌厂商的正面竞争中快速做大业务，而这些厂商无一例外都与华为的客户建有合资企业。华为的产品每线约1000元，利润率超过20%，为公司带来了急需的现金流。在1993—2002年的10年间，华为员工增长到22000人，其中研发工程师超过10000人，收入逼近40亿美元。
徐直军写道，许多美国领先企业都是从小做起，靠卖产品赚钱，再把钱投回研发，是「自力更生、白手起家」精神的写照，华为同样如此。
2002年，徐直军意外被安排负责「集成产品开发（IPD）」变革项目。2005年，他被任命为投资评审委员会（IRB）主任，负责审视公司所有产品投资。2011年出任轮值CEO后，他卸下了IPD相关职责，但至今仍投入大量时间推动IPD与时俱进。
IPD是创始人任正非（Ren Zhengfei）的构想。1997年，任正非走访几家美国科技公司，受到它们产品开发方式的启发，回国后请IBM担任顾问，帮助华为在研发投资决策流程、组织和文化上进行全面变革。
变革于1999年启动。徐直军起初只是众多参与者之一，2002年起主导变革，与IBM顾问一起把IPD流程推广到整个研发体系。
徐直军表示，变革前华为的研发成功高度依赖个别优秀人才，能否做出高质量产品带有很大偶然性。10年间，华为建立起研发投资决策流程、集成产品开发流程及相应的组织架构、能力和IT系统，统称IPD管理体系。这套体系让华为管理的研发队伍从当时的10000人增长到如今的110000人，并持续交付满足客户需求的产品。
1997年到2006年，华为向IBM支付了近6亿美元咨询费。徐直军称，这很可能是中国历史上规模最大的单笔企业咨询项目。
2003年，徐直军被任命为无线产品线总裁。该部门当时有2000多人，正在研发3G产品。他的使命是推动华为从固定网络设备供应商转变为同时提供固定和移动网络设备的供应商。由于中国一再推迟发放3G牌照，华为转向海外市场。
当时负责WCDMA产品的余承东（Richard Yu）与荷兰运营商Telfort查看基站机房时发现，欧洲站点获取成本极高，且没有空间新建基站。他从欧洲打电话给徐直军，提议把基站拆成两部分：一部分装在铁塔上，另一部分做成小盒子塞进现有机柜。两人决定研发分布式基站。
这一方案曾遭到许多运营商和设备商抵触，但最终成为如今3G、4G、5G基站的标准形态。余承东还在与西班牙电信德国公司（Telefónica Germany）讨论时提出把2G和3G集成到同一基站。此后华为与沃达丰（Vodafone）合作推出SingleRAN，把2G、3G、4G、5G在物理上集成到同一基站，成为全球基站的事实标准。徐直军说，这两个方案确立了华为在移动通信领域的全球领先地位。
为在数字信息洪流中找到「诺亚方舟」，华为加强了基础研究。2011年，华为把研究团队整合为2012实验室（2012 Labs），名字取自电影《2012》，涵盖基础研究、核心技术和工程能力。实验室资金一部分来自公司，其余来自各业务单元。徐直军从第一天起就参与塑造实验室。自动驾驶、5G和昇腾（Ascend）都在这里孵化。
华为2009年启动5G研究，最终成为5G标准的主要贡献者和全球最大的5G专利持有者之一。2012年，徐直军主导在2012实验室下成立车联网实验室（Internet of Vehicles Lab），催生了华为的自动驾驶解决方案。2016年，实验室开始研究人工智能。
徐直军写道，他职业生涯的绝大部分时间都在见证、参与并引领华为的创新之路，这条路由研发团队数十万人共同走过。
文章介绍，华为通过利益分享机制，让几乎每一位符合条件的员工都能获得公司股权。华为每年把年收入的10%以上投入研发，2007年至2025年累计研发投入超过2700亿美元（1.8万亿元人民币）。截至目前，华为已获得授权专利245000件。
文章署名为徐直军副董事长、轮值董事长，落款日期为2026年9月29日。
英文原文：
Back in May 1993, as I was wrapping up my Ph.D., I took a trip to Shenzhen to look for work. While at a friend’s house, I was flipping through theShenzhen Special Zone Dailyand was immediately drawn to a full-page recruitment ad from Huawei. It specifically highlighted that they were looking for people with a master’s degree or doctorate. It was odd. As the frontier of China’s reform and opening up, Shenzhen was a hub for processing and trade, and very few companies were hiring people with higher education. Out of curiosity, the next day I went for an interview and ended up getting offered a job.
As the day of my graduation neared, I had already received several offers but the one from Huawei stood out. Their HR manager would call me every week, saying everything was ready and they were expecting me. The respect they showed me, the urgency of their calls, and the promise of a salary five times more than that of my Ph.D. supervisor made the decision easy. I chose Huawei.
When I came onboard in August, I realized that, of the company’s 600 employees, more than 200 were R&D engineers, as well as a ton of postgraduates. Five Ph.D.s joined that same year. It was unimaginable anywhere in China, especially in Shenzhen, because the entire country was producing only about 2,000 Ph.D. graduates a year. Huawei’s huge bets on talent in its early years were a clear sign of its commitment to innovation and forward-looking investment.
Back then, the R&D team was working on digital stored program control switches. As China’s economy soared, demand for household and office phones far outstripped supply. For years, getting a phone line installed was difficult and expensive. Switches from Western vendors were available in China, but they used a wide range of incompatible standards and went for about 3,000 Chinese yuan per line. That was a fortune at the time.
This presented an opportunity. Back then, major urban centers were surrounded by geographically dispersed townships and villages. Setting up a complete switch for each location didn’t make economic sense, and connecting distant communities with copper was both expensive and prone to signal loss. So, we came up with a fiber-connected remote switching module, which allowed for a single centralized system to provide network coverage for an entire county.
Our customers liked it. And that’s how we managed to rapidly grow our business in a head-to-head race with established Western vendors, all of whom had joint ventures with our customers. Our product was cheaper, around 1,000 yuan per line, with a profit margin higher than 20%. This helped sustain Huawei’s rapid growth and gave us much-needed cash flow. In a short span of ten years (1993–2002), our headcount climbed to 22,000, including over 10,000 R&D engineers, and revenue was closing in on $4 billion.
Many leading U.S. companies started out small. They struck out on their own, probing forward one step at a time, selling products to make money, and investing it back into R&D. They epitomized the ethos of “pulling yourself up by your bootstraps.” And the same can be said for Huawei.
In 2002, I was unexpectedly put in charge of a major transformation program called Integrated Product Development (IPD). Then in 2005, I was appointed the chair of the Investment Review Board (IRB), a role that oversees all of the company’s product investments. When I became rotating CEO in 2011, I handed over the IPD-related responsibilities, but even now I still devote a lot of my time to keeping IPD up-to-date and driving Huawei’s business development.
IPD was a game changer in how we developed products and made R&D investment decisions. It also completely reshaped our R&D structure. Prior to this transformation, our R&D successes were highly dependent on individual outperformers, and high-quality products were a matter of chance. The IPD transformation put systems in place for us to reliably deliver competitive, quality products that meet our customers’ needs.
IPD was the brainchild of our founder Ren Zhengfei. In 1997, he visited several U.S. tech companies and was inspired by the way they developed their products. Following this visit, he brought on IBM as a consultant to help us design and execute a sweeping transformation across our R&D investment decision processes, organization, and culture.
The transformation kicked off in 1999. I was one of many participants in the program until 2002, when I was given the reins to work alongside IBM consultants and drive the transformation forward, rolling out the IPD process across all of R&D.
Over the course of a decade, we developed a more efficient decision-making process for R&D investment, a process for integrated product development, and the corresponding organizational structures, capabilities, and IT systems – collectively called the IPD Management System. It’s this system that has enabled us to not only manage an ever-expanding R&D team (from 10,000 members then to 110,000 today) but keep delivering innovative products that meet our customers’ evolving needs.
From 1997 to 2006, Huawei paid IBM nearly $600 million in consulting fees. This may very well be the largest single corporate consulting engagement in China’s history.
In 2003, I was appointed president of the Wireless Product Line. At the time, the division had just over 2,000 people, and was working on 3G products. My mission was to transform Huawei from a fixed-network equipment supplier to a provider of both fixed and mobile network equipment. As China kept delaying the issuance of 3G licenses, we were forced to pursue market expansion overseas.
Back then, Richard Yu was responsible for WCDMA products, one of the main technologies that powers 3G networks. While visiting a base station equipment room with the Dutch carrier Telfort, he discovered that site acquisition costs in Europe were exceptionally high, and there wasn’t really any room for new base stations anyway. He called me from Europe to discuss whether we could split a base station into two parts—one part placed on the cell tower, and the other into a small box that could be installed inside the existing base station’s cabinet. We both thought it was a plausible solution, so we decided to create a distributed base station.
The solution was met with resistance from many carriers and equipment vendors, but it eventually became standard for the 3G, 4G, and 5G base stations we see today. Richard also proposed integrating 2G and 3G into a single base station during a discussion with Telefónica Germany. Later, in collaboration with Vodafone, we launched SingleRAN, a solution that physically integrates 2G, 3G, 4G, and 5G into a single base station. This, too, became the de facto standard for base stations around the world. These two groundbreaking solutions—the distributed base station and SingleRAN—established Huawei’s global leadership in mobile communications.
To find our Noah’s Ark amid the coming flood of digital information, we strengthened our basic research efforts. In 2011, we consolidated the company’s research teams into a new organization—2012 Labs, named after the film2012: Doomsday. It would serve as a platform for investment in sustained innovation, including basic research, core technologies, and engineering capabilities. At the same time, we changed our R&D investment model: Part of 2012 Labs’ funding would come from the company, and the rest from its business units. From day one, I have been shaping the Labs’ organization, team, capabilities, and future-oriented research projects. Autonomous driving, 5G, and Ascend were all incubated there.
We began researching 5G as early as 2009, ultimately becoming a major contributor to 5G standards and one of the world’s largest holders of 5G patents, cementing Huawei’s global leadership in the field. In 2012, I led the establishment of the Internet of Vehicles Lab under 2012 Labs to conduct research on future automotive technologies, a move that led to our autonomous driving solutions, and laid the foundation for our leadership in that domain. In 2016, the Labs began researching AI, securing us a place in the race to come.
Thanks to the Labs’ pioneering basic research, technology readiness, and engineering capabilities, Huawei has been able to continuously launch innovative products and solutions that keep ahead of market demand.
I joined Huawei in 1993 to work in R&D, and have spent the vast majority of my career witnessing, taking part in, and leading Huawei’s innovation journey—one shared by hundreds of thousands of people in our R&D team.
Our commitment to talent is backed by a benefit-sharing mechanism that gives virtually every qualified employee an equity stake. We invest more than 10% of our annual revenue back into R&D, with a total R&D spend of more than $270 billion (1.8 trillion yuan) from 2007 to 2025. We have the utmost respect for others’ intellectual property while protecting our own, with 245,000 patents granted to Huawei to date. Our system for effectively managing R&D investments has helped us bring together a large swathe of incredible engineers, laying a solid foundation for sustained innovation.
All of these factors have enabled Huawei to keep pushing the limits of our products, launching one competitive offering after another to grow and thrive in highly competitive markets.
Eric Xu
Deputy Chairman, Rotating Chairman
September 29, 2026
