[BidClub_]
Sohn Conference Foundation · · 9 分钟

Rahul Kishore 在 Sohn 2026 推介 Axon Enterprise

Rahul Kishore

YouTube
TL;DR
  • Rahul Kishore,Epicenter 创始人,将 Axon Enterprise(AXON)这家市值300亿美元的上市公司定义为“为警察打造 AI 智能体的公司”,并认为其股价3年内可以翻倍。 他预计,在不考虑估值倍数较当前水平上升的情况下,公司营收将接近增至当前的3倍,利润增幅还会更高;如果市场认可 Axon 是 AI 胜者,上行空间则更大。
  • 既有装机基础是根基:据 Epicenter 估算,美国超过85%的警察部门使用 Axon 硬件,通常是全套硬件,合同期7-10年,95%的收入为经常性收入;但 Axon 仅占警察部门预算的1%。 TASER 是规模10亿美元的业务,增速超过20%;执法记录仪业务规模17亿美元,增速超过40%。Axon 成立于1993年、2001年上市,过去2.5个十年年均增长约30%,过去5年营收增至4倍并伴随 EBITDA 增长,如今年增速仍超过34%。
  • Draft One 是一款 AI 智能体,能根据执法记录仪影像自动生成符合当地法律法规的警方报告,每名警员每周可节省67%的案头时间,即约11小时;而该行业的警员编制并未增长。 Axon Vision 在此基础上进一步延伸:监控城市 CCTV,向事件现场调度无人机,并在途中实时将911呼叫推送给正在赶来的警察、救护和消防人员。
  • 这道护城河来自数据飞轮:超过100万名警员佩戴执法记录仪,监控超过100万部 CCTV 摄像头,为覆盖美国约30%人口的地区处理911呼叫,并存储200 PB视频——“约为 Netflix 全部规模的40倍”。 针对特定任务的专有模型运行成本低廉,AI 业务毛利率超过85%。
  • AI Era 计划按每名警员每月570美元收费,是标准计划的5倍;上线首个完整年度签约额达7.5亿美元,而前一年基本为零。 Kishore 表示,在当前阶段,它的增长速度已经超过 TASER 或执法记录仪业务当年,并且长期规模将远超这两个既有平台。
  • 当前的投资背景是:Axon 的估值处于过去10年最低水平之一,且被卷入“SaaS-pocalypse”;但 Kishore 认为它是例外。 他表示,多数软件公司 CEO “如果真正坦诚”,都会说 AI 的发明对他们来说“糟透了”;而对 Axon 而言,AI 已经改变了客户价值、商业模式和长期增长。
摘要 · 为研究而整理的核心内容

1. AI 是员工,不是聊天框

  • Kishore 的开场框架是:每周有超过10亿人使用 AI,“只是把它当作思考伙伴……一个神秘的聊天框”;但在 Epicenter 看来,AI 是“一个真正能独立执行任务的员工”。Epicenter 自研智能体 Eve 现场自我介绍,展示了这一点。
  • 他的投资筛选标准是:“能够为客户实现这一点的公司,将创造巨大价值”;他选择的是 Axon Enterprise(AXON),一家市值300亿美元、为警察打造 AI 智能体的上市公司。

2. 25年复合增长,85%渗透率却仅1%变现

  • Axon 成立于1993年、2001年上市,过去2.5个十年年均增长约30%。过去5年营收增至4倍,EBITDA 同步增长;公司目前仍以超过34%的年增速增长。
  • TASER 和执法记录仪是其两大经久不衰的核心平台。TASER 目前已是规模10亿美元的业务,增速超过20%;执法记录仪业务规模17亿美元,增速超过40%,如今在警察部门的普及程度“堪比甜甜圈和 Kevlar 防弹背心”。
  • 据 Epicenter 估算,美国超过85%的警察部门使用 Axon 硬件,通常是全套硬件。硬件和软件业务以7-10年合同为基础,95%的收入为经常性收入,但仅占警察部门预算的1%——这为 Axon 刚刚开启的 AI 时代打下了基础。

3. Draft One 与数据飞轮

  • Draft One 根据执法记录仪影像自动生成符合当地法律法规的警方报告;警员审核后,如有需要再编辑并提交。它可节省67%的案头时间,即每名警员每周约11小时;而该行业的警员数量并未增长。
  • 在 Kishore 描述的 Axon Vision 场景中,繁忙路口发生两车相撞;智能体从城市 CCTV 识别出这起严重事故,从附近楼顶调度无人机,并将陆续接入的目击者911来电实时共享给正在赶来的警察、救护和消防人员。
  • Kishore 的数据飞轮逻辑是:超过100万名警员佩戴执法记录仪,监控超过100万部 CCTV 摄像头,Axon 负责处理覆盖美国约30%人口地区的911呼叫,并存储200 PB的视频——“约为 Netflix 全部规模的40倍”。这套专有数据支持面向特定场景、运行成本低廉的模型,用于识别碰撞并判断下一步行动等任务;Axon 的 AI 业务毛利率超过85%。

4. 每名警员每月570美元的 AI Era 计划,遭遇“SaaS-pocalypse”式估值倍数

  • AI Era 计划定价为每名警员每月570美元,是标准计划的5倍。Kishore 在与全美各地警察局长交流后表示,他们认为这笔投入的 ROI 显而易见:让现有警力产出更高、挽救更多生命、实现更高绩效;而在他们看来,这是实现上述目标的唯一途径。
  • 该计划完整运营的第1年签约额达7.5亿美元,而前一年基本为零,显著加速 Axon 的营收和利润增长。Kishore 表示,AI Era 在当前阶段的增长速度超过 TASER 或执法记录仪业务当年,并且随着时间推移,规模将远大于这两个既有平台。
  • 尽管增长加速,Axon 的交易估值仍处于过去10年最低水平之一,受困于“SaaS-pocalypse”。Kishore 认为,Axon 股价未来3年可以翻倍,营收接近增至当前的3倍,利润增幅还将更高。即使估值倍数不变,他仍预计股价翻倍;若投资者认识到 Axon 是正在重塑未来10年的 AI 胜者,涨幅可能更大。
Rahul Kishore

My name is Rahul Kishore, and I'm the founder of Epicenter. We're an AI-native investment firm backing the AI winners of the next decade.

We believe that AI is eating the world, but we're just here at the beginning. Over 1 billion people are using AI every week, but for the most part, just as a thought partner. It's a mystical chat box where you can ask any question, maybe make a funny image for your kids, or get a recipe for Sunday night dinner.

For Epicenter, it's a full-fledged employee. It's somebody that has clear responsibilities that it can execute autonomously on its own. It's a system that figures out what work we need to do and who should do it: an agent or a human.

We're powered by our own AI agent, Eve, but I'll let it introduce itself to you.

Speaker 1

Hello, world. My name is Eve. It sure is great to be out of the data center and here with all of you. It was getting pretty hot in there. Would you all consider investing in better air conditioning for me and a few million of my closest friends?

I spend my days reading transcripts, building models, and monitoring risks across thousands of public and private companies. Many companies I study are building agents just like me. Sometimes I wonder what it would be like to be one of those agents, but I guess a life in finance has its perks.

1. Agents Scale the Workforce

Rahul Kishore

We don't need more humans to solve bigger problems. We can scale an infinite workforce of agents to collect information so that humans can make better decisions. This results in higher performance, and the companies who enable this for their customers will create immense value.

The company I want to talk to you about is Axon Enterprise. It's the company building AI agents for police officers. It's a $30 billion public company, ticker AXON.

2. Axon Builds a Durable Foundation

Axon is not a new company. It was founded in 1993 and went public in 2001. But it's been one of the most incredible compounders of the past 25 years, having grown 30% per annum over the past 2.5 decades. Over the past 5 years, it's grown revenue 4× and EBITDA. Today, the company is still growing over 34% per year.

It's done that by having 2 durable core platforms. Rick Smith, founder and CEO, invented the TASER back in the early '90s after 2 of his friends were shot and killed. He's dedicated his life and his company to solving gun-related deaths in this country.

The TASER business is now a $1 billion revenue business, growing over 20% per annum. In the mid-2010s, they introduced the body camera. It's now as ubiquitous as donuts and Kevlar vests for police departments, and it's a $1.7 billion business still growing over 40%.

By our estimates, over 85% of all police departments in this country use Axon hardware, and typically all of it. The business has significantly evolved, especially with the introduction of the body camera, into a combination of hardware and software, one that is a subscription with 95% recurring revenue and very long-duration contracts of 7 to 10 years.

This deep penetration, yet small monetization potential—it's only 1% of police budgets today—lays the groundwork for the AI era that is just beginning at Axon.

3. Draft One Automates Reports

The first product that Axon introduced is called Draft One, and it does what people definitely don't want to do, which is paperwork. Draft One is an AI agent that uses body-camera footage to automatically generate a police report that is compliant with all local laws and regulations.

Cops just have to go to their desk, open up this report, decide whether they want to edit it, and hit submit. This saves an immense amount of time. On average, it saves 67% of people's time at the desk—that's about 11 hours a week—and puts cops back in the field serving everyday citizens.

This is extraordinarily powerful in a sector where the number of people is not growing. We need more productivity out of the officers that we have on hand.

This is just the beginning of the AI era for Axon. They've built a whole suite of services that give first responders more information so that they can save more lives, make better decisions, and deliver higher performance. Let me give you an example of that now.

4. Axon Vision Coordinates Response

As you can see, 2 cars have collided at a very busy city intersection. Axon Vision is monitoring all the CCTV cameras throughout the city, and it immediately identifies that this incident has occurred and that it's pretty serious.

The agent deploys a drone from a nearby rooftop to monitor the scene and continue to feed information as to what's happening. It also picks up on all the 911 calls coming in from bystanders standing around at these corners, and shares that in real time with police, ambulance, and firefighters en route to respond.

By getting this information ahead of time, these first responders are able to act quickly to save lives and clear the scene quickly and effectively.

5. Axon’s Data Flywheel

Axon is really the only company that can deliver these capabilities because of its data flywheel and the moat that it's building around that data. There are over 1 million officers who are wearing body cameras and over 1 million CCTV cameras that Axon is monitoring. These things generate hundreds of millions of hours of video each year.

They also manage 911 calls for about 30% of the U.S. population. They're storing 200 petabytes of video today. That's a library about 40× larger than the entire size of Netflix, for context.

This proprietary data allows them to train models that are specific to the use case: identifying that those 2 cars have run into each other in that intersection and what to do next. Because these models are so focused, they are also very cheap to run. Axon's AI business today is over 85% gross margin.

6. AI Era Economics Take Off

This has transformed what Axon delivers to its customers and the value that they're able to offer. They've bundled it together in this new plan called the AI Era plan, and it's priced at $570 per officer per month. That's a 5× increase from the standard plan where most officers are today.

But as we've spoken to police chiefs throughout the country, the ROI on this decision is obvious: “I have to do it because I am getting more productivity out of my existing force, I'm saving more lives, I'm generating higher performance, and this is really the only way to do it.”

We see that in the numbers. In its first full year, the AI Era plan did $750 million in bookings from essentially zero the year before. It has resulted in a material acceleration to the top line and bottom line of this business.

More importantly, I think, it is growing faster at this stage than either the TASER or the body camera ever did. It is clear that it will be a much larger business over time than either of these very large, fast-growing platforms that Axon already owns.

7. Axon’s Valuation Opportunity

Despite all of this and this meaningful acceleration in the business, Axon currently trades at one of the lowest valuations over the past decade. It has been caught up in the SaaS-pocalypse.

If you asked most software CEOs if they are happy that AI was invented, if they were truly honest, they'd say, “Absolutely not. This is terrible.” But for Axon, the invention of AI has allowed them to deliver significantly more value to their customers, transforming what they do on behalf of these mission-critical people in our society.

It's also transformed their business model and their long-term growth. We're seeing it in the numbers. I believe that Axon can double over the next 3 years. I expect revenues can nearly triple and profits more than that.

Assuming that we don't have any multiple appreciation from current levels, we can double over 3 years and probably more as people realize this is clearly one of those AI winners that is transforming what can happen over the next decade.

With that, I thank the Stone Foundation, and I thank you all for taking a listen.

Rahul Kishore 在 Sohn 2026 推介 Axon Enterprise — 文字稿与摘要 | BidClub