从绿拖拉机到农业AI先锋:John Deere CEO 详解智能工业战略与精准农业 Norges Bank Investment Management 2026-09-02

农业的经验局限与技术变革

John May: 我自己也有一个农场。我自己打干草捆,而我能产出的草捆质量,完全取决于我自己的经验水平。所以,当我判断干草已经到了可以收割的状态时,我坐进割草压扁机(moco)进行切割,然后坐在拖拉机里打捆。我一边仔细听着机器的动静,一边盯着电脑屏幕看数据。但最终成品的品质,完全受限于我个人的技术水平。

Original English

John May: I have a farm myself, and I bale hay, and the quality of the bale that I can produce is only as good as my experience set. So, you know, when I think the hay is at a point where it can be harvested, I get in the moco, I cut it, and then I sit in the tractor and I'm baling it. I'm listening for things, I'm looking at the computer, but the end product is going to be only as good as I am.

Nicolai Tangen: 大家好,我是挪威主权财富基金(Norwegian Sovereign Wealth Fund)首席执行官 Nicolai Tangen。今天与我对话的是 John Deere 的首席执行官 John May

提到 John Deere,你可能会想到标志性的绿色拖拉机——屏幕上 John 身后确实就停着一辆。但这家公司如今已经成为农业领域最先进的技术公司之一:能够自动驾驶的拖拉机、能区分杂草与作物的人工智能算法……而 John 正是推动这一巨大转变的领军人物。他几乎在 Deere 奉献了整个职业生涯,如今正带领这家传统农机巨头蜕变为一家真正的科技公司。

John,热烈欢迎你的到来!

Original English

Nicolai Tangen: Hi everybody. I'm Nicolai Tangen, the CEO of the Norwegian Sovereign Wealth Fund, and today I'm joined by John May, the CEO of John Deere. Now when you think of John Deere, you may think of green tractors, and indeed there is a green tractor standing behind John here on the screen. But the company has become one of the most advanced technology companies in farming: tractors that drive themselves, AI that tells a weed from a crop, and John is the man behind the shift. He spent almost his entire career at Deere, and now he is trying to turn the maker of farm machines into a real technology company. John, warm welcome.

John May: 谢谢你,Nicolai。非常荣幸能来到这里,感谢邀请我参加你的播客节目。

Original English

John May: Thank you. It's great to be here, and thanks for including me on your podcast.


智能工业战略与生产系统重构

Nicolai Tangen: 太棒了。John,你在 2019 年出任 CEO,随后很快便推出了所谓的“智能工业战略”(Smart Industrial Strategy)。这项战略的核心是什么?

Original English

Nicolai Tangen: Fantastic. Now, John, you became CEO in 2019, and then very quickly after that, you launched what you call Smart Industrial Strategy. What is that?

John May: 这项战略的实质,是将我们的关注点更深度地聚焦在客户本身以及客户所执行的具体作业流程上。过去,我们的思维模式往往是制造出世界上最好的播种机;而现在,我们转向利用技术赋能,帮助客户完成前所未有高水平的播种作业

因此,我们围绕客户的完整“生产系统”(Production Systems)重组了业务架构,并大规模重金投入技术研发,旨在让客户获得更高的生产力、更强的盈利能力以及更好的可持续性

Original English

John May: Well, really what we wanted to do is get more focused on the customer itself and the jobs that our customer do. You know, in the past, what we would do is focus on building the best planter, and we wanted to shift to helping our customers plant better than they ever had in the past by leveraging technology. So we realigned around production systems, and then really invested heavily in technology that would make our customers more productive, more profitable, and more sustainable.

Nicolai Tangen: 这与 Deere 以往的运营模式有何本质不同?

Original English

Nicolai Tangen: And how is it different from how Deere used to run?

John May: 过去我们并不一定专注于客户面临的最核心瓶颈。举例来说,农户往往会在特定时间节点遭遇严重劳动力短缺,或者因为无法精准控制化肥投入量而蒙受损失。如果我们仅把注意力放在机器本身的硬件性能上,就很容易忽略整个作业周期中的系统性痛点。

重新聚焦生产系统后,我们能够统筹考量从土壤准备、精准播种、作物养护到最终收获的每一个环节,将农机硬件与先进的感知传感、软件算法全面打通。

Original English

John May: You know, in the past we didn't focus on necessarily the biggest challenges that our customer had. So for example, labor availability at certain times of the year, or input costs where they're over-applying fertilizer. By looking at the machine instead of the job, we missed opportunities to solve the real bottlenecks. Realigning around production systems allowed us to connect hardware and software across the entire lifecycle—from tillage and planting to spraying and harvesting.


计算机视觉与精准喷药技术:See & Spray

Nicolai Tangen: 谈到技术创新,你们推出的 See & Spray 精准喷雾技术令人瞩目。这项技术在实际农田作业中是如何工作的?

Original English

Nicolai Tangen: Speaking of technology, your See & Spray technology is quite remarkable. How does it actually work out in the field?

John May: See & Spray 是我们利用计算机视觉机器学习解决农业痛点的典型范例。当大型喷雾机以每小时 15 到 20 英里的速度穿过农田时,喷杆上安装的 36 个高帧率摄像头以每秒数十次的速度拍摄地面图像。

机载 AI 计算机在毫秒级时间内对每一帧图像进行推理计算,能够清晰区分绿色作物(如玉米或大豆)与夹杂在其中的杂草。一旦识别到杂草,喷嘴会以毫秒级的响应速度精准喷出微量除草剂,实现“只喷杂草,不碰庄稼”。这直接帮农户减少了高达 70% 到 80% 的非残留性除草剂使用量。

Original English

John May: See & Spray is a prime example of applying computer vision and machine learning to farming. As a 120-foot boom moves across a field at 15 to 20 miles per hour, 36 cameras along the boom capture images at high frame rates. Onboard AI processors classify every pixel in milliseconds, distinguishing a weed from a crop plant like corn or soybeans. The moment a weed is detected, a nozzle fires precisely on target. It reduces herbicide usage by up to 70 to 80 percent, which delivers huge cost savings to farmers and massive environmental benefits.

Nicolai Tangen: 这不仅大幅降低了农户的化学品采购成本,对生态环境保护也是巨大的贡献。除了除草,这项技术还能扩展到化肥喷施上吗?

Original English

Nicolai Tangen: That's an enormous cost reduction and a major win for the environment. Beyond weed management, can this precision spraying approach be applied to fertilizer application?

John May: 没错,我们将其延伸到了播种阶段的施肥场景,推出了 ExactShot 技术。以往农户在播种时,施肥管道通常是持续不断地向地沟里流淌液态启动肥。

但借助 ExactShot 传感器和高速电磁阀,系统能在每颗种子被精确植入土壤的瞬间,仅向该种子周围喷射一小滴肥料,而在种子之间的空隙停止喷施。单是这一项技术,就能帮助农户减少高达 60% 的初期化肥用量,同时确保作物根系获得最充分的养分供应。

Original English

John May: Absolutely. We applied that exact same principle to starter fertilizer with a technology called ExactShot. Traditionally, farmers apply a continuous stream of liquid fertilizer in the furrow during planting. With ExactShot, sensors detect when a individual seed is dropping, and high-speed valves pulse a precise dose of fertilizer directly onto the seed, pausing in between. That reduces starter fertilizer use by up to 60 percent while still giving the plant everything it needs to emerge and thrive.


全自主作业拖拉机与农业自动驾驶

Nicolai Tangen: 自动驾驶在汽车领域谈论了很多年,但你们已经在农田里交付了全自主作业拖拉机(Autonomous Tractor)。农田场景的自动驾驶与城市道路有何不同?

Original English

Nicolai Tangen: Autonomous driving has been talked about for years in automotive, but you've already delivered fully autonomous tractors to production agriculture. How is autonomous operation in a field different from city streets?

John May: 农田环境表面看起来没有红绿灯和行车标线,似乎比城市简单,但实际上它的工程挑战非常独特。农田里有强烈的扬尘、飞石、剧烈的颠簸振动、复杂的地形起伏以及各种不可预测的障碍物(如倒下的树枝、沟壑、野生动物等)。

我们的自主拖拉机配备了 6 对立体双目相机,配合 360 度环视感知系统和深度学习神经网络。农民只需将拖拉机运送到田边,在移动端应用(Operations Center Mobile)中划定作业边界并点击启动,就可以离开农田去处理其他业务。机器会自动规划路径、避障、自主掉头,并在遇到无法确认的复杂情况时向农民手机发送警报,等待远程确认。

Original English

John May: While farm fields don't have pedestrians or traffic lights, the operational environment presents unique challenges: extreme dust, harsh vibration, changing lighting conditions, and unpredictable obstacles like sinkholes, stray animals, or fallen tree limbs. Our fully autonomous 8R tractor uses six pairs of stereo cameras providing a 360-degree field of view with advanced deep neural networks. A farmer simply transports the machine to the field, configures the boundary via the Operations Center app, swipes to start, and can walk away. The machine executes the job, detects obstacles, and alerts the farmer remotely if it encounters something ambiguous.

Nicolai Tangen: 农民对把价值数十万美元的重型机械交给 AI 自主掌控接受度如何?

Original English

Nicolai Tangen: How receptive have farmers been to handing over a half-million-dollar piece of heavy machinery entirely to AI control?

John May: 接受度非常积极,核心原因在于劳动力短缺。现代农业面临的最严峻挑战之一就是找不到足够的技术工人。在秋收或春播的关键窗口期,可能只有短短几天适耕天气,机器必须连轴运转 24 小时。如果找不到司机,作物的减产损失是不可挽回的。全自主作业解决了“用工荒”,让农户从繁重的舱内重复劳动中解脱出来,转型成为管理全局的农场运营主管。

Original English

John May: The adoption curve has been driven by a massive structural problem: agricultural labor scarcity. Finding skilled operators who are willing to sit in a cab for 16 to 18 hours a day during narrow planting or harvesting windows is nearly impossible in many rural areas. Autonomous equipment doesn't get fatigued, runs around the clock when weather windows open up, and elevates the farmer from a machine operator to a farm fleet manager.


数据飞轮与商业模式演进:软硬件一体化生态

Nicolai Tangen: 这一切技术背后都依赖海量的数据与算力。John Deere 是如何构建自己的数据生态系统的?

Original English

Nicolai Tangen: All of this capability relies heavily on data and compute. How has John Deere built out its connected data ecosystem?

John May: 我们的核心枢纽是 John Deere Operations Center。目前全球有数亿英亩的耕地数据连接在我们的云平台上。互联机器在田间作业时,实时上传土壤湿度、播种深度、喷药速率、产量分布等几十种遥测与农艺数据。

这构成了一个强大的飞轮效应:机器越智能,收集的数据越精准;农户在云端获得的决策分析就越深入;而反过来,这些数据模型训练又让下一代农机作业更具自适应能力。客户拥有自己数据的所有权,并能无缝将数据授权分享给农艺师、种子商或金融伙伴。

Original English

John May: The foundational layer is the John Deere Operations Center. Today, hundreds of millions of engaged acres around the globe are connected to our cloud platform. Every connected machine in the field streams real-time telemetry and agronomic data: soil conditions, seed placement depth, application rates, moisture levels, and yield maps. This creates a powerful flywheel: smarter machines capture higher-fidelity data, which feeds predictive agronomic models, enabling even higher operational efficiency. Farmers own their data and can securely collaborate with agronomists, input suppliers, and retail partners.

Nicolai Tangen: 软件与订阅服务在你们的整体商业模式中占比如何变化?你们是否正在转型为类似 SaaS 的商业模式?

Original English

Nicolai Tangen: How is the contribution of software and recurring revenue changing your business model? Are you shifting towards a SaaS-style model for equipment?

John May: 我们正在推行“按需解锁解决方案”(Solutions-as-a-Service)模式。传统的农业装备购买是一次性的大额资本支出,农户承担了所有前期硬件风险。

现在,我们将先进的传感器和计算硬件作为标准配置内置于机器中,然后通过软件订阅模式提供 See & Spray 或自主导航等高级功能。农户可以根据实际作业亩数或季节性需求按需付费。这种方式降低了农户尝试新技术的资金门槛,同时也为 Deere 创造了高毛利、可预测的经常性软件服务收入(Recurring Revenue)。

Original English

John May: We are evolving towards a Solutions-as-a-Service model. Traditionally, agricultural machinery was purely an upfront capital expenditure, placing the full financial risk on the producer. By standardizing advanced hardware kits on machines and offering high-end capabilities like See & Spray or autonomous operation through usage-based or annual software licenses, we lower upfront adoption friction. Farmers pay relative to the value and acres treated, while Deere builds a durable, recurring revenue stream.


新能源动力与可持续农业未来

Nicolai Tangen: 谈到可持续性与碳中和,重型农机设备的电动化进展如何?电池能否取代大型柴油拖拉机?

Original English

Nicolai Tangen: When we talk about sustainability and decarbonization, what does the transition look like for heavy farm machinery? Can batteries actually replace diesel in high-horsepower equipment?

John May: 这是一个需要因地制宜、分场景推进的技术路线图。对于中小型设备(如草坪修剪机、紧凑型多功能拖拉机、果园专用机),纯电电池动力已经非常成熟且极具经济性,我们在这些领域推出了全电动产品。

但在大马力重载作业场景(比如一台 600 马力拖拉机在泥泞农田中连续拉扯重载犁具 14 个小时),目前的锂电池能量密度会导致电池包极其庞大沉重,反而会严重压实土壤,破坏土壤通透性。因此对于大马力装备,我们正多管齐下探索可再生柴油生物乙醇氢内燃机以及混合动力架构,确保在减排的同时不牺牲作业功率与连续出勤率。

Original English

John May: We approach powertrain electrification with a nuanced, workload-specific roadmap. For compact utility tractors, turf care, and specialized orchard equipment, pure battery-electric powertrains are viable and economically compelling today. However, for large production agriculture—such as a 600-horsepower tractor pulling heavy tillage tools for 14 continuous hours in high-draft conditions—the current energy density of batteries would add prohibitive weight, causing severe soil compaction. For those heavy-duty applications, we are investing in renewable biofuels, ethanol, e-fuels, hydrogen combustion, and advanced hybrid architectures.

Nicolai Tangen: 全球粮食安全是未来几十年的重大议题。在耕地面积不增加甚至减少的情况下,科技如何支撑全球人口增长的粮食需求?

Original English

Nicolai Tangen: Global food security is one of the defining challenges of our time. With arable land remaining flat or declining, how does technology bridge the gap to feed a growing global population?

John May: 预计到 2050 年,全球人口将接近 100 亿,全球粮食产能需要提升近 50%。与此同时,耕地面积、水资源在减少,化肥与化学品监管日益严苛,极端天气频发。

唯一的解题钥匙就是单产效率提升单株精细化管理(Plant-by-Plant Management)。以往农业是以“田块”为单位进行粗放式管理,而现在的 AI 与传感技术让机器能够以毫米级精度对待每一颗种子、每一株幼苗。通过精准施肥、精准除草、最佳时机播收,我们在显著减少资源消耗的同时,大幅推高单产上限。这是可持续农业的必然路径。

Original English

John May: By 2050, the global population will approach 10 billion people, requiring an estimated 50 percent increase in agricultural output. Yet arable land is constrained, fresh water is scarce, and regulatory requirements on chemical inputs are tightening. The only solution is unlocking yield gains through plant-by-plant precision management. Historically, farming treated fields as broad, uniform units. Today's technology enables sensing and action at the individual plant level. By optimizing every single seed's environment while drastically reducing total chemical inputs, we can sustainably meet global demand.


企业文化、技术人才与领导力哲学

Nicolai Tangen: John,将一家拥有近两百年历史的传统制造巨头改造成软件和 AI 驱动的科技先锋,组织文化上面临的最大挑战是什么?

Original English

Nicolai Tangen: John, transforming a nearly 190-year-old manufacturing legacy into an AI- and software-driven tech powerhouse must have met internal friction. What was the biggest cultural hurdle?

John May: 最大的挑战在于打通业务壁垒与思维惯性。Deere 过去是由各个独立的产品线部门(如拖拉机部门、联合收割机部门、播种机部门)各自为政运作的,大家各做各的优秀硬件。

转向智能工业战略意味着我们必须打破部门孤岛,建立统一的通用技术架构平台。我们在旧金山、奥斯汀等地设立了先进技术研发中心,吸引了全球顶尖的计算机视觉、机器学习和云原生软件工程师。更关键的是,让硅谷的技术极客与深谙农田实际操作的农业机械工程师深度融合,建立互相尊重的跨界合作文化。

Original English

John May: The greatest challenge was breaking down silos. Historically, Deere was structured around isolated product divisions—combine division, tractor division, seeding division—each optimizing their own hardware roadmap. Our Smart Industrial Strategy required unifying into enterprise-wide technology platforms. We opened engineering hubs in places like San Francisco and Austin to attract top-tier machine learning, computer vision, and cloud engineering talent. The magic happens when you bring brilliant software engineers together with ag engineers who have dirt under their fingernails and deep domain expertise.

Nicolai Tangen: 你在 Deere 奉献了几乎整个职业生涯,并在不同地区、不同部门轮岗工作过。这段经历对你的领导风格产生了怎样的塑造?

Original English

Nicolai Tangen: You've spent almost your entire career at Deere, working across different divisions and living in multiple countries. How did that broad trajectory shape your leadership philosophy?

John May: 在不同部门、不同国家和文化背景下的工作经历,是我职业生涯中最宝贵的财富。走出熟悉的舒适区、前往海外不同市场实地考察,会让你深刻意识到各地客户的需求差异有多大,营商环境与团队文化有多么多元。

这造就了我核心的领导力准则:包容多重视角,倾听每个人的声音。在管理团队中,我鼓励每一位成员充分表达观点,因为每个人都拥有我所不具备的独特经验与视角。唯有汇聚所有人的智慧,企业才能在剧烈的技术变革周期中保持敏锐与前瞻。

Original English

John May: Having opportunities in different business units, functional areas, and living internationally outside the United States was truly transformative for me as a leader. Operating in diverse geographies gives you firsthand visibility into different cultures, customer pain points, and commercial practices. It reinforces my core leadership conviction around genuinely valuing different perspectives. In our leadership team, I want every voice to be heard because each person brings distinct insights and experiences that I simply don't have.

Nicolai Tangen: John,这是一场极其精彩且富有洞见的深度对话。我敢保证,从此以后大家再看到绿色拖拉机时,眼光将彻底不同。

Original English

Nicolai Tangen: John, this has been a fantastic conversation. I can assure you that when people see a green tractor from now on, they will view it in a completely new light.

John May: 谢谢你,Nicolai!非常感谢你的邀请与交流时间。

Original English

John May: Thank you, Nicolai. I really appreciate the invitation and your time.

Nicolai Tangen: 感谢大家的收听。如果你喜欢本期播客,请点赞、订阅并开启通知,这能帮助我们持续邀请更多杰出的行业领袖做客节目!

Original English

Nicolai Tangen: Thank you all for listening. If you enjoyed this episode, please like, subscribe, and turn on notifications. It really helps us bring on more great guests.

📌 文中提及的人物和组织

关键字: precision-agriculture autonomous-machinery computer-vision smart-manufacturing business-model