如何赢得交易、做大2500亿美元业务|Nikesh Arora,Palo Alto Networks
Palo Alto 的扩张,建立在把获客成本摊薄到更广泛的产品组合之上。 Arora 接手时,公司市值为180亿美元,他的目标是把市值推到“1000亿美元以上”;8年后,他谈到公司已有70多款产品。他的经济逻辑是:将销售、市场和支持费用占成本的比例从约60%降至20%左右,从而为40%的营业利润率腾出空间——向已经赢得信任的客户销售更多产品。
Arora 通过收购完整团队,换回多年产品研发时间。 他认为,打造一款优秀的科技产品需要4到7年;而初创公司持续迭代时,大公司挑战者还得从招人开始。“买的是团队”,意味着保留团队领导权,包括让被收购公司的创始人领导内部那个没能击败他们的高管。
并购谈判的关键,是签约前敲定双方认可的产品路线图。 Arora 要求在最终协议签署前的30天内制定联合路线图;Palo Alto 曾在这一过程中放弃2笔交易。创始人会重新获得为期3年的股权归属安排,理想情况下是4年,因为被收购产品通常还需要数年稳定执行。
估值更高,不代表战略上必要的收购就该叫停。 Arora 表示,如今同类标的的价格是2年前的3到4倍,所处阶段“可能还是那么早”。他的判断标准是资产在未来5到10年的潜在贡献:“并购不是战略”;如果把现有产品组合做得更好能给客户创造更大价值,公司也可以停止收购。
Arora 认为,AI 将重塑企业组织,也会扩大网络安全需求。 他说,“我相当确定”,2年内人们会像信任某一领域的人类专家一样信任 AI;5年内,公司组织结构将截然不同。如果助手能完成专业工作中90%的任务,1个人就可能覆盖目前分散在8个团队中的职能;与此同时,连接增多也意味着攻击面扩大。
他为 Palo Alto 商业模式辩护的核心是:按 token 计价的 AI,无法低成本取代持续的安全检测。 他举例称,公司每台笔记本每年收费25美元,每天检测约150 MB流量,边际检测成本为零。AI 可以提升分析能力,但安全仍需要数据接入、基础设施、专业知识和运营“框架”;他的结论毫不含糊:“我们的业务没有风险。”
他的经营哲学,是激进执行,同时接受结果存在不确定性。 Arora 说,这种从容是在过去5年逐渐形成的。他举例说,发现新的漏洞后,应检查系统、封堵暴露面、连夜打补丁、发布防护方案并联系客户,因为“焦虑不会改变结果”。
1. 关系的价值,往往在交易之外显现
Arora 在 Google 欧洲区任职时,与 Silver Lake 的 Egon Durban 建立了联系。他拒绝了正式顾问安排:参加为期2天的闭门会议、做演示,还得“喝免费咖啡”。Durban 提议改成非正式交流,由他继续负责投资决策。Arora 说,这样的合作持续了约15年;后来,Silver Lake 投资 William Morris 后,他和 Marc Andreessen 也为其提供科技领域的建议。
有人脉,不代表交易一定能落地。一次与 Andreessen 共进晚餐后,Arora 开始推动投资 Oculus,与 Brendan Iribe 见面,还带妻子去洛杉矶郊外的工业园看演示。他和 Durban 都同意投资,但交易文件迟迟没有送到。随后 Oculus 被 Facebook 收购:他们最终没能把支票投进公司。
Senra 说,Ari Emanuel 打电话通常只有约25秒,连再见都不说。Arora 熟悉这种以交易为导向的沟通方式,但也讲了另一面:他离开 SoftBank 后,Emanuel 邀请他聊聊未来。Arora 原本只给这次会面留了半小时,之后要去机场,结果两人聊了90分钟。“当一件事对他真的很重要时,他会花很多时间。”
2. 从一开始,多产品分销就是增长逻辑
留在同一家公司,不等于重复做同一份工作。Arora 在 Google 任职期间,公司员工从2000人增至6万人,营收从20亿美元升至620亿美元。在 Palo Alto,8年间业务从防火墙扩展到70多款产品。他的标准是:平台能否持续带来足够有挑战性的问题。
Arora 加入 Palo Alto 时,带来的是企业业务经验,而非网络安全专长。他最初的目标,是把180亿美元市值的公司带到“1000亿美元以上”。他观察到,产品开发通常占成本的15%–18%,行政管理占6%–10%;其余大头主要是销售、市场和支持费用。若把这部分占比从60%降至20%左右,营业利润率就有望达到40%。
具体做法,是让更多产品共用同一笔获客成本。如果只有一个产品,敲开20家公司的门,可能已有15家拥有所需产品。有了“身后的一整车产品”,销售对话就变成“我还能为你提供什么?”围绕同一段昂贵的客户关系,邻近产品也能创造更多销售机会。
他的目标客户是全球排名前10000的企业,只有在赢得客户信任后,才逐步扩大其支出。产品线够宽还不够:产品必须好用,还得做到一流。“没人想买网络安全领域的第二名产品。”这也直接引出了并购问题。
3. 并购买来研发时间,也买来团队默契
Arora 的判断很明确:“做出一款优秀的科技产品需要4到7年。”加入 Palo Alto 时,公司缺少一条足够丰富的产品管线,因此面临时间上的难题。收购公司可以加快产品扩张,但前提是避开那些让并购史变得惨淡的执行错误。
一家初创公司已有一支50至75人的成熟团队,可能领先18个月,而且成员已经积累了协作经验。大公司若打算新招50名工程师,得先把人招齐、组织起来;等到第一个版本推出,竞争对手可能已经领先2年。Arora 对比说,一支默契团队可以每周工作80小时;大公司的招聘却要经过人力资源和4轮面试。
他以 Google 曾讨论收购 Netflix 为例。内部反对意见是,YouTube 已经有视频分发基础设施:只要加上专业内容、1000万美元和20名工程师就行。“事情从来不是这么简单。”Senra 问这类错误是否源于 CEO 的自负;Arora 更强调执行失误。“没有哪个 CEO 会站出来说:‘我今天早上就想试着犯个错。’”
4. 被收购的赢家,必须在大公司里继续赢
Arora 不赞成把收购交给内部那个曾在该品类中输给对手的管理者。初创公司创始人资源更少却赢了,就该由他领导,即便原有员工需要向他汇报。把这项“宝贵资产”奖励给失败的内部团队,可能会毁掉它。他用一句话讽刺并购中的傲慢:“我的创新是天才之作。他们的成功纯属走运。”
品类排名很重要,因为第1名和第2名能拿走不成比例的利润,但如何判断排名,需要经验。同类公司可能 ARR 相近,客户只有30至40家,最初的产品优先级却不同。Arora 承认,Palo Alto 曾收购自己认为排名第1的公司,后来又因执行问题丢掉领先地位。35笔收购下来,“并非每件事都按计划发展”。
Senra 追问,看似相似的竞争者之间究竟差在哪儿,并引用网球教练的观点:顶尖选手的差距在心理层面。Arora 认可创始人的意志力是因素之一,但更强调适应能力:被收购的创始人需要从管理8名销售人员,转向管理6000人。这些人必须成为“我的放大器”,而不是障碍。做过弱势方,不代表就能自然适应成为“大赢家”后的角色。
Senra 提到 Hock Tan 的建议:被收购员工应该意识到“征服者”和“被征服者”的区别。Arora 采取相反做法,谈判时保持尊重,因为对方之后会成为同事。“如果身边20个人觉得自己没有赢,我就赢不了。”他们必须觉得自己是平等参与者,自己的成功同样重要。
5. 先对齐产品,再付钱
有一笔早期收购,交易完成后双方花了6个月争论产品方向。创始人仍想推进原来的愿景;Arora 则认为,卖掉公司就意味着接受共同方向。这段经历改变了他的做法。现在,他会直接与创始人谈,而不是让银行家挡在这些未来必须一起工作的人中间。
商业条款谈妥后,财务和法务团队负责文件,创始人则与 Arora 及 Palo Alto 的产品负责人一起工作。他们有30天时间设计“我们共同的产品蓝图”。如果没有签好的联合路线图,他就不会签最终协议:“如果你不想做我想做的产品,那我们就不该做这笔交易。”
Senra 问创始人是否抗拒失去自主权。Arora 说,支票到账之前,达成共识容易得多;当然,创始人仍可以退出。没有创始人退出过,但 Palo Alto 自己放弃了2笔交易。这个流程暴露出真正做到意见一致有多难;其中一笔交易还在团队深入研究架构时,显露出技术底子没有预想的扎实。
留住创始人,才能保障产品继续走完研发之路。Palo Alto 通常收购已开发2到4年的产品,后续还要再做2到3年,并为创始人重新安排股权归属,通常为3年,理想情况下为4年。如果创始人已经打算离开,Arora 就不会收购。商业出身的创始人后续可能没那么关键;产品和工程出身的创始人,则是收购本身的核心。
6. 战略决定何时收购、出什么价,也决定何时停手
Arora 承认,基于知识产权细分市场的收购 Aprimo 没有成功:预期中的市场并未出现,Palo Alto 只能止损。时机也可能过早。Evident.io 在他加入之前被收购,产品支持 Amazon;后来收购的 RedLock 则支持三大云平台。他的判断框架包含“价格、阶段和信念”,而不是把抢先入场视为必然优势。
当前融资环境让这套判断更复杂。类似公司如果在2年前收购,价格可能只有现在的三分之一或四分之一,所处阶段“可能还是那么早”。Senra 问他会怎么做,Arora 回答:“只要符合战略,就买。”他评估的是资产在未来5到10年能发展到什么程度,而不是把过去更便宜的融资环境当成决定性参照。
Senra 提到 Henry Singleton 后来约15年不再收购。Arora 承认,Palo Alto 也可能走到这一步:“并购不是战略。”当内部团队还没有进入某个位置时,并购只是执行战略的一种手段。如果让现有产品更好地协同能为客户创造更多价值,那就可能是更值得投入管理精力的方向。
7. AI 改变企业形态,也扩大安全边界
Arora 认为,Google 让“知识或信息普及化”,AI 则在“让智能普及化”。他说,“我相当确定”,2年内人们会像信任某个领域的人类专家一样信任 AI。他设想,用户可以选择大众化建议,或者承担一定风险、获取前沿建议,而不是只能接受一种固定水准的判断。
现在的产品开发会把工作分给约8类专业岗位,包括编程、测试、文档、产品和 UI。因为个人无法样样精通,才有了这种分工。如果能力足够强的助手能完成每个专业岗位90%的工作,1个人就可能覆盖这些职能。Arora 预计,5年后“我们今天所运行的组织形态”将截然不同。
技术和连接越多,安全需求也越多。Arora 提到,Facebook 每天有36.5亿用户,相当于基础设施每天面对数十亿个潜在入口;他又把这一逻辑扩展到消费应用。AI 还带来一个不断变化的安全边界:“用于保护 AI 的产品,目前还没有完全成形。”他的需求判断,建立在需要保护的连接数量和种类不断增加之上。
8. AI 会强化攻击者,但不代表防守方会被取代
Arora 把网络安全归纳为3项工作:约一半是在边界阻止威胁;10%–20%用于内部安全卫生和分析;剩余20%–30%用于发现入侵者并在造成损害前将其清除。这样的拆分,区分了 AI 发现漏洞的能力与防守系统所需的运营工作。
他表示,OpenAI 的新模型和 GLM-5.3 能发现漏洞,攻击者也可以“低价租用”。Palo Alto 的应对方式包括“虚拟补丁”:用特征规则为暴露的系统临时加一道防护,等待客户修复底层软件。Arora 说:“自从 Mythos 出现以来,我们的业务增长了。”他把漏洞发现与客户立即寻求防护联系起来。
在终端安全上,他说公司每台设备每年收费25美元,每天检测约150 MB流量,“上下浮动”,而流量检测的边际成本为零。按 token 计价的 AI,成本结构则不同。他举出的替代方案日成本——5000美元、500美元,最终降到50美元——只是示意:“随便挑个数字。”他的论点在于成本结构,而不是说这些数字代表某项 AI 服务的报价。
在网络内部,AI 可以加快对异常行为的分析,但仍需要“框架”、网络安全专业知识、数据接入和基础设施。Arora 的结论毫不含糊:“我们的业务没有风险。”Senra 说,这和他之前的外部印象一致,同时仍追问威胁环境变化有多快。
9. 执行力消除摩擦,从容接受无法掌控的部分
Senra 问,Google 是否让人觉得这是千载难逢、压力巨大的机会。Arora 说,他34岁时觉得一切都很正常:搜索量和广告主需求持续增长,工作就是消除摩擦。Google 在奥地利的收入已达1000万美元,却因为不接受当地货币而受阻;广告主还需要一张美国信用卡。“看看能不能把这些障碍清掉。”
销售 Google 广告时,团队既不决定价格,也不决定流量:竞价者决定价格,搜索量决定广告库存;法国下雨,收入都可能增加。因此,Arora 的团队会和客户一起研究转化率与线索经济性,确定最高出价。之后,广告主便可以尽量获取符合回报要求的现有流量。
他的平静是最近才形成的——过去5年逐渐培养出来,并非一直如此。Senra 将其描述为印度教式的宿命论;Arora 谈的是不确定性,以及可控的努力与不可控的结果之间的区别。Senra 追问这种想法在实践中有多难。Arora 的回答既容纳充分的风险控制,也接受最终结果:“到了某个时候,你会说:‘我已经尽力了。’”
他举了一个涉及2250万个位置上的 PHP 的漏洞案例。团队检查 Palo Alto 自身的暴露情况、封堵风险、修补系统,连夜为客户开发防护方案,发布补丁并联系受影响的客户。“焦虑不会改变结果。”他的做法是排查所有可能的漏洞,然后“把剩下的交给命运”。
10. 孩子需要自己的目标,而不是继承父母的压力
Arora 讲述的是自我驱动的成就,而非父母施压。他父亲不知道他如何申请上大学、又如何进入美国商学院。Arora 当时要2000美元,父亲一开始只能拿出200美元;他带着这笔钱出发,之后父亲又给了他2000美元学费。他也说,自己不会逼孩子复制他的职业道路。
女儿 Aisha 曾在 Brex 做市场,之后加入 Bill Maris 的 Section 32,后来创办了自己的基金。她请 Arora 给建议,意外带来了另一层收获:两人每天会通话1到2次,讨论公司和投资决策。他称这段关系是“一份祝福”,强调的是共同兴趣,而不是把女儿的道路说成自己替她安排的结果。
Senra 提出创始人常见的两难:自己是贫穷家庭的孩子,养的却是富裕家庭的孩子。Arora 不赞成反复强调这种反差。比他更穷的人,早就会把他的父母视作富人。“哥们,放下吧。”孩子无法选择出生环境;难点在于帮他们建立抱负,同时不把现有条件视为理所当然。
他的儿子热爱篮球,小女儿则投入体操,训练时很有毅力。儿子成为篮球运动员或评论员都可以,只要愿意努力,并从中获得快乐。目标是培养出“健康、快乐、均衡”的孩子。他还说,孩子会密切观察父母的行为,但强迫他们照着父母的样子生活,未必能带来想要的结果。
11. 持续参与对话,才能不断学习
Arora 每周至少花4到5小时和 AI 创业者交流,每月还会接触大约5到10家网络安全初创公司,“也许15家”。他通过交谈、阅读和刷 Twitter 学习,同时承认自己吸收得有限。“没有真正的专家”:人们了解自己正在做的事,但对其更广泛影响的理解仍在逐步建立。
Harry 曾提醒 Arora,与其等着别人通过播客向他提问,不如自己建立受众;之后,他开始更多使用 X。坚持每天发帖一段时间后,他改为有话才说:“我不想把它变成一份工作。”帖子都是他自己写的;有一次他用 Claude 帮忙改语法,遭到批评后,说自己不会再试了。
完整逐字稿
本集逐字稿暂无译文,当前显示原文。
Did you read Ari Emanuel’s new book?
No.
You’ve heard that you’re in it, though?
No, I didn’t know that.
Okay. He says the way he met Egon Durban was that he got introduced to Marc Andreessen, and Marc told Ari that he, Nikesh, and Egon had this affiliated group of angel investors who were working together. Tell me a little bit about that.
Yes. I met Egon when he had just come to Europe. He was a Silver Lake partner there, and he reached out to a mutual friend saying, “Listen, I want to hang out with some people, get to understand the tech scene over here, and see if I can get an advisor to Silver Lake.”
I was in my early 30s. I used to run Google Europe. Egon and I had lunch, and I said, “Listen, why are you so reluctant about this conversation? I’ve tried to do this. They really don’t want advice. They make you come for a two-day offsite. They present a whole bunch of stuff to you. The last thing I need is more PowerPoint in my life and free coffee.”
He said, “You know what? That resonates with me. I’m going to call you, and if you tell me that’s a good idea, I’ll independently consider whether it’s a good idea or not, because I have to make the decision. If you tell me it’s a bad idea, I’ll think about it similarly. But don’t worry, you don’t add as much value by telling me what to invest in and what not to invest in. So, can we have fun?”
As it turned out, I did that for about 15 years. That’s when I met Egon, when he was in Europe.
You were just investing your personal money?
Yeah, of course. But I’ll tell you a funny, interesting incident. One time, I flew back from CES with Marc Andreessen. Marc rarely travels, but when he does, he goes to CES and hosts a dinner. He had me come, and we spoke. I spoke a little bit there. Then I flew back and sent him an email saying, “Hey, Marc, is there any cool company you’re working on right now that I could put some money in?”
He said, “The guy who spoke at dinner?” I said, “Yes.” He goes, “Oculus.”
Palmer Luckey?
Yes. Well, Brendan, to be fair. Brendan Iribe and Palmer Luckey, of course. I said, “Great.” They gave me a number, so I texted Brendan and said, “Hey, we met at the dinner. I’d love to make an investment.”
He said, “Great, but you have to come see this stuff. It’s really cool.” I said, “Okay.” I was in L.A., and I said to Brendan, “I’m coming to L.A. Can we take a look at the technology?” It was me and my wife. He called and said, “Yeah, but you have to drive an hour outside of L.A. because we have this industrial park. We’re going to go to this park.”
So I took a car and went to the industrial park. The lights were out. He said, “This isn’t going to work.” We were hanging out, waiting for the lights to come back on. We went in, and he put this contraption on our faces. It had this little virtual reality type of situation going on. I said, “This is really cool. This is very cool. What’s going to happen?” He said it was going to be cool because of the promise of visualization, gaming, and so on.
I said, “Great. Can I write a check?” He said, “Sure. It’s you, Blake Krikorian, and Vivi Nevo, and I have room for one more.” I said, “What? You’re going to let us do this together?” Marc was already in the company.
Mhm.
So Egon and I agreed to do the deal. He sent me some documents—an email saying, “I’m working on the docs.” He didn’t respond. Egon said, “What’s going on? Why haven’t we heard from him?” So I sent him an email. He said, “The docs are just coming. They’ll be there with you any time soon.” No docs. Then boom, you wake up one morning and you’ve just been sold to Facebook.
Before you were able to invest?
[Laughter.]
All right. So, this group that Ari writes about—
Well, we ended up being the technology advisory board to William Morris when it was acquired by Silver Lake.
Okay. So—
Marc and I.
Okay, well, I don’t even know about this. What were you guys advising them on?
Whenever we had a question on technology, the advisory around Egon and Ari was very straightforward. Ari does the 15-minute phone call. You can keep Ari—
Fifteen? I did the episode with him, and now his book is called “Rolling the Calls.”
Yeah.
Now he does it to me. I get a call and it’s like, “Hold for Ari Emanuel.” I’m like, “You’re calling me?” And they’re like 25 seconds, and there’s no goodbye.
Yeah, but what’s fascinating is that we would do that when we had something to talk about. I understand his shorthand. He’s trying to get something done. He’s a man on a mission.
1. Selling Google ads without knowing the price
There was a time when I had left SoftBank and he reached out to say, “What are you doing now?” I said, “Nothing. I’m just trying to figure out what I’m going to do.” He said, “Why don’t you come by and see me, and then we can talk about it.” I said, “Okay, I happen to be in L.A. I’ll come see you.”
My assistant said, “Ari’s office says you’re having lunch. What time should I schedule your departure?” I said, “Listen, it’s probably going to be 15 minutes. Why don’t you schedule half an hour, and then we’ll get me to the airport and get me on a flight back?” That was the only time I spent an hour and a half with Ari. You can tell that he actually wanted to get into what I was going to do next, and he invested time in that conversation. So he does have a soul. When it’s really important to you, he’s going to spend a lot of time, but most times he’s transactional when he wants to get something done.
We were talking a little bit about some of the people we either know jointly or know of. You were saying some people jump around a lot. They have an idea, work on it for a little bit, and then they move on. It’s almost like they have ADHD. How do you think about yourself? You seem to have these jobs for a long time. You were at Google for 10 years. How long have you been at Palo Alto Networks now?
Eight.
Eight. I like to focus on one thing for a long period of time. How do you work?
The good news is that even at Palo Alto Networks, or when I was at SoftBank or Google, there was a lot changing. Google went from 2,000 employees to 60,000 employees when I was there. It went from $2 billion in revenue to $62 billion in revenue.
Even though you’re in one place, you’re doing so many different things. At that point in time, you went from search and YouTube and DoubleClick to buying tons of businesses and thinking about all kinds of stuff. So I think it’s not what you do at one point in time; it’s how much your brain can get challenged. How many different problems can you get involved in? You can swap from problem to problem and solve many things in the same space, as long as you get the right environmental structure.
If the opportunity is big enough at Palo Alto, when I joined eight years ago, we were a firewall company. Today, we have more than 70 products. We’re in every aspect of cybersecurity. We’re all over AI, trying to make sure we pivot hard on the AI stuff. You can always keep yourself challenged as long as you get the right platform in front of you.
I just talked to Nick from Revolut, and that was his thing.
2. I thought cybersecurity was two words
He’s like, “There’s no way I was going to run a company with one product. That would have been boring.” He’s running 20 different experiments at any given time. I think he’s in 30 or 40 countries. He wants to get to 100 different countries. When you got recruited to become the CEO of Palo Alto Networks, they essentially had one product, right?
Yes, the firewall.
Part, yeah. Okay, was this expansion that’s occurred over the last 8 years something you came in with a plan for? Or what was the conversation?
I had a plan when I came in.
You had a hilarious line where you said, “I thought cybersecurity was two words,” or something.
Yeah. Honestly, in my interview, I ended up talking a lot more about executive security. I thought that was an important topic, and then I realized it was a very small market. I was talking about how each of us needs to be secure. I was like, “No, that’s a consumer business. You don’t want to get involved in that. That’s a lot of marketing, a lot of R&D.”
So, I came in with more of an enterprise framework study than a cybersecurity framework study. I sat there and looked and said, “Which are the largest companies that exist in enterprise?” You look at them and say, I don’t know, Cisco or Salesforce at that point in time. These were hundreds of billions of dollars in market cap. We were an $18 billion market cap company. I’m like, “How do I go from an $18 billion market cap company to north of $100 billion?” That was the original expectation.
Then I said, “If you look at the P&L of any enterprise company—”
Mm-hmm.
Most companies spend 15% to 18% on product development. That’s where R&D sticks around. Everything else, administrative, is 6% to 10%. So the entire difference between our operating margin and gross margin is sales and marketing and customer support. If you can take 60% of your costs, which typically live in sales, marketing, and customer support, and bring them down to 20%, you can be a 40% operating margin business. That was my insight.
I’m like, “How do I get sales and marketing amortized over time?” You figure out that the biggest expense in enterprise is customer acquisition. That’s the maximum. That’s what costs the most. If you have a low-revenue product that you can only sell once to a customer, it’s really hard to build a big business. If you look at the largest businesses—Microsoft’s enterprise business or Salesforce, or take your pick of the largest businesses—they sell a lot to the same customer. That was the insight.
I’m like, “Well, if you’re going to be big at Palo Alto, we’ve got to find a way of selling a lot to a customer. You only have one thing to sell.” So you knock on 20 doors, spend all that time, and 15 people say, “I’m sorry, I don’t need it. I already have it.” You’re stuck with five.
You knock on the door with a truck behind you saying, “What do you want? You want this? You don’t want this? You want this? You want this?” It’s like, “The world is your oyster. What can I get you?” I had that insight. I’m like, “We need to find adjacencies to our business that customers will trust us to sell, so we can keep expanding what we sell to customers. We can serve the top 10,000 customers in the world and increase the amount of money we get from them.”
You only get that if you earn their trust. You only get that if your shit works. You only get that if it’s best in class, because nobody wants to buy the second-best product in cybersecurity. It kind of sucks. You want to buy the best, so our need was to be the best.
3. Why he buys instead of builds
That’s one part of the insight. The other part, which is still true, is that it takes 4 to 7 years to build a good technology product. It doesn’t matter who you are, what you are, or where you are. You look back and say, “This thing is not fully ready. It’s been 7 years. This thing looks closer to being ready.” Whether it’s Uber, YouTube, Facebook, or Google Search, it took 4 to 7 years, sometimes longer, to build a useful product that customers found exciting and interesting. That’s true in enterprise as well.
Sometimes I say, “Holy shit, I just joined this company. You don’t have a large product pipeline. It’s going to take me 4 to 7 years to build real products. What am I going to do?” You’re going to acquire them. That’s right, but you have to be careful because history is littered with bad acquisitions, and companies destroy acquisitions for the most part, whether in enterprise or consumer markets. So you have to figure out what makes for a good acquisition and what makes for a bad acquisition. I spent the next part of my life studying that.
What were you studying? How were you learning about this?
You go back and look at patterns. How many companies did company X buy? How many did company Y buy? Why did they work? Why did they not work? Which ones worked? What are the key characteristics? Was it a failure of the market? Was it a bad idea, something you shouldn’t have done? Was it bad execution or bad leadership?
You figure out, “Great, these are the 7 reasons why they failed. How can I minimize the risk of failure?” You minimize that by saying, “Okay, let me control 5 variables in a different way. Then I only have 2 risks. I have the risk that I picked the wrong idea. I bought the right company, but the market didn’t show up.”
We had one of those.
What was that company and how much did you pay for it?
The company is called Aprimo. We thought the world was going to go to IP-based segmentation, but it never went there. We cut our losses quickly once we decided it wasn’t going to go there, and we moved on. It’s almost like a portfolio business. It’s like marginal venture capital. The pharma companies do it, right? They try a whole bunch of drugs, take them to a certain phase, and say, “Cut bait. It’s not going to work.” You have to do the same thing.
I’ve been in companies where somebody says, “Okay, we should do this. We should build this. We should buy this.” I’m like, “No, we can’t buy it. We can build it.” I was at Google when they said, “Let’s go buy Netflix.” I was like, “No, we have YouTube. We can build the same capabilities. The video-serving platform is there to get professional content instead of consumer content. That’s easy to do technically.”
“Great. Give us $10 million and 20 engineers. We’ll solve it.” It never works like that. I had the same thing at Palo Alto. If I tell my team of engineers, “Hey, we need to buy this,” they say, “No, we can build it.” Once you build it, you have 2 or 3 problems. One, the people we’re seeing succeed are already a year and a half ahead of us. They have a functioning team of 50 to 75 people working on it. You want to go hire 50 engineers. By the time you hire the right people, get your shit together, start building, and have version one, they’ll be 2 years ahead.
4. Why the acquired founder becomes the boss
Why do you think so many other CEOs make that mistake? Is it an ego thing, like, “No, I can build this”? What is it?
I don’t think it’s an ego thing. I don’t think CEOs set out to make mistakes. No CEO stands up saying, “Let me try and make a mistake this morning.” They believe whatever they’re doing is right, and they most likely identify the right company.
I think M&A often fails because of bad execution. Your tendency is to be too busy, doing a lot of stuff. You buy the company and say, “Great, I have a guy or gal who manages this topic. This person does video streaming. He or she has been trying to do video streaming for the last 5 years. I want to buy a video-streaming company and give it to him or her.”
And you instead put the founder of the company—
Yeah, like, “Whoa, whoa, whoa, wait a minute. This person just got their ass kicked with all the resources, all the money we had, by that guy who got funded by venture capitalists and went out and lived in a crummy warehouse, hired people, and, through blood, sweat, and tears, built something that has product-market fit. And he or she should work for the empire? That’s not going to work. He should be part of the empire.”
So I switch it. I say, “Okay, this person has done really well. This person is now going to be the head of video streaming. Everyone can learn from him. You can work for him.” That sometimes doesn’t go down well with some of the people who have to end up working with—
—the existing organization.
Of course they—
They want to end up leaving or whatever.
But it doesn’t matter, right? They were the ones who were chartered with winning. They had access to resources, and they didn’t. You can’t reward that behavior by giving them a prized asset and running the risk that they destroy it.
That’s one failure mode. The other failure mode is when people say, “My innovation is my genius. Theirs is pure dumb luck.” That means, “I’m a founder. I created something cool. These guys are not smart. I’m just going to buy number two or number three, spin and shine it, make it like new, and kick their ass.” That never works. I want to buy number one.
5. Why he only buys number one or number two
You’ve talked about this a lot. Why are number one and number two so important to you?
If you look at most markets, number one and number two end up with a disproportionate share of the profits in the category.
This is what you were trying to get out of me before we started recording.
Which we'll talk about after.
Okay, fine.
That's exactly what's happening right now.
Right. Number one and number two end up with a disproportionate share of the profits. Number three gets close sometimes, but never. If you're going to win, you want to be number one or two in the long term. If you're not going to be number one or two, you have a problem.
If you buy number two or number three, remember that number one still exists. More than likely, you're going to slow down whatever you buy because you're still a large enterprise and you can't control every variable. There'll be somebody from finance saying, “I'm here to help.” Somebody from HR saying, “I'm here to help.” Some poor guy says, “Dude, let me go build the product.” Then you have this whole onslaught of people showing up saying, “I'm from the mothership, and I'm going to help you.” So you have to figure out how to maintain momentum for number one.
You don't have an acquisition team, right? I read somewhere that it's just you.
Well, no. There are people who are very good at what they do, and they work with me. It's not like I work alone, but eventually every acquisition ends up with me being involved.
Let's say you decide, “Okay, we want to extend into this product.” You look at both number one and number two, right? From the outside, maybe number two looks similar to number one. When you get deeper and closer, do you notice a difference? Can you explain what you typically find?
Sometimes there are small differences between number one and number two. Sometimes we've had a situation where we bought what we thought was number one, and number two went around us and became number one because we screwed up the execution. That happens, too. We buy 35 companies; not everything works out as planned.
Typically, both number one and number two have product-market fit. If you think about product development and say, “This is what the end product will look like in four years,” number one could have started with this as their primary thrust, and number two could have started with that part as their primary thrust. They both look like they're winning in the product category.
Sometimes it becomes apparent later that you should have started here because that's the primary need, and the customer wants this to work really well before they take the second half. But it's hard to tell in the early days, when both of them have 30 or 40 customers that effectively prove product-market fit.
You'll find it's a strategy question, an execution question, a product-market-fit question, and a business-model question. All those things are somewhat different, but it's hard to project how they're going to pan out over the next two to three years. You have to sit back and look at both their positions and say, “Which one is more likely to succeed?” Then you have to add to that, “Which one is more likely to be aligned with our product direction?”
I made a mistake early on, with the first or second company I bought. They spent six months trying to debate product direction after we bought them. I said, “Wait a minute, this is a problem.” These guys started a company, and rightfully, the founder was like, “I started a company with this vision in mind.” He sold the company, but that was still his vision.
Wait, what did you think the money was for?
Exactly. “No, I sold the company, but I still want to build what I wanted to build.” I'm like, “It doesn't work like that.” But it took me a while to realize it doesn't work like that. I felt like, “Oh my God, I'm going to [bleep] a founder.” I just didn't know. I can't do it; it doesn't work like that.
6. The 30-day test before every acquisition
So then we changed our model. Now, when I sit down with a founder, I agree on a deal, and I insist on reading the deal with the founder. I don't involve bankers in this because I'm like, “You're going to work with me for the next four years. Do you really want somebody else in the middle negotiating on our behalf, so that eventually it's, ‘But that's not what was sold to me’? We should have that conversation.”
The moment we agree to the deal, I tell the founder, “Listen, let the finance guys and the legal guys figure out all the paperwork. Your job is now to sit down with me and my product guys and design what our joint house is going to look like. We have 30 days to do it.” So we literally go back and forth on product direction.
Before the acquisition?
Before the acquisition closes. We don't sign the definitive term sheet—
I don't think I've ever heard that idea before.
I don't sign the definitive term sheet until I get the founder to sign on to a joint product roadmap. I'm like, “If you don't want to build what I want to build, then we shouldn't do this.”
It's natural for a founder to want autonomy and to think they have the right answers. So how difficult are these companies to find?
When there's a large check hanging on the other side of 30 days, founders are much more amenable to aligning their product vision than after they've got the proceeds. It's unbelievable. It's night and day.
By the way, they have the choice to pull the ripcord. They don't have to do it.
How many have backed out at that stage?
Nobody backed out. We backed out twice.
Why?
Because in the process, we figured out it would be hard to get consistent alignment on the product. We figured out it would be hard to manage the founding team afterward because they had very strong opinions about where they wanted to go. We could sense that they were playing ball because they knew they had to play ball to close the deal.
In one case, as we kept digging deeper and pushing harder over those 30 days, we figured out that the substance wasn't as strong as we'd sensed in our early diligence. As you dig deeper, you start looking at architectures, and you have to combine architectures, and you have to—
Do the founders typically stay on long after?
7. Why nothing at work stresses him out
Some do. One of the other things we do is, when you buy something like this, you're buying a team. You're buying a team, the vision, and the execution. So it's very hard when a founder says, “I'm out.” We don't buy companies when founders say, “I'm out.” If you're out, you should go, which means you could kill the company anyway if you're done. You have to want to be around.
Typically, there are two or three founders. One will slack off because they're doing biz dev. Maybe you don't need biz dev, or you can do better biz dev. One's doing go-to-market, and you don't need that. You'll find founders fall off depending on the function they bring in, but it's never the product or engineering guys who fall off, because that's what we're buying. We need them.
We still have many founders who work at Palo Alto. Some have converted to just continuing to work at Palo Alto. Some have the founding itch and leave after three years. But the other thing we always do is, when we buy a company, we re-vest the founders for three years, ideally four.
Why is that so important?
We want continuity. We want continuity in the execution of the vision, because they built something. I've said it takes four to seven years to build a great product. Typically, we buy these product companies two years in, or four years in. We think there's another two or three years of consistent product development to be done, and we want the founder and the founding team to keep driving that with continuity.
If they're not there to lead it, disruption could cause us to lose the lead. If they're number one in the category, we want them to continue executing in that direction.
I want to go back to this idea of looking at number one and number two, and then getting really close to them. Is there anything else that surprises you? They may look similar from the outside, but the more you dig in, you see why number one is in the position they're in.
What's interesting is that the hardest thing to get your arms around as a larger company is that sometimes number ones also win because of the sheer will of the founder.
But some founders are just—what do you call them?—alpha killers or whatever they are.
What are you talking about? We need context here.
Yes.
Because you don’t want to be—okay, good.
No, no, no. We’ve run into each other a few times, and I’ve watched a lot of your interviews. I just think there’s no bullshit. The reason I asked you to be on the show is that I like these founders who seem to have minimal bullshit. We’re here to fucking win.
Yes.
And that’s what we’re doing.
Yes. And this is something you’ve already alluded to a few times. I was shocked, but when we were talking before we started recording, you were playing that down a little bit. I was like, “No, no, no. I’m pretty sure I know what’s sitting across from me.” I was naming some other founders who remind me of you in the way that you operate.
It’s one thing to be that way as the leader of the entity. It’s another thing to have people within your team who are that way. If you acquire a company and the founder is like you described, they’re going to break glass. That’s okay, but they also have to find a way of coexisting with the larger entity.
When you’re a founder, you might have eight people, and I have 6,000 people in the field. You have to find a way to leverage those 6,000 people. They’re not all bad. They’re amazing, right? Sometimes, when you’re a founder, you take the point of view that everybody’s stupid, I’m smart, and I’m going to get this shit done. Anybody who gets in my way is slowing me down and isn’t competent enough.
You need a mental reset if you’re going to survive and succeed in a larger ecosystem. You have to be able to say, “I get it. I have 6,000 salespeople. My job is to motivate them to go out there and sell my product, because if they do, my vision of making this an amazing product will be realized. Those 6,000 people will be my amplifiers.”
Sometimes founders who are used to working with small teams and being the underdog don’t realize they’ve gone from being the underdog to being part of the overdog. That’s the risk you have to watch out for.
Tie what you’re saying to the will of the founder. You’re saying the will of the founder helped them succeed in a smaller company, but they have to modulate their behavior. Is that right?
You have to make sure they can modulate their behavior and succeed in your environment.
Got it.
If they can’t modulate their behavior and succeed in the environment, then I lose their leadership on the product side because they’re going to rub people the wrong way and get frustrated. That’s what I have to watch out for. Sometimes number one and number two have to be adaptable to figure out who’s going to be a better fit in our environment.
I’m personally obsessed with this idea that there’s usually a gap between number one and number two—
Yes, I heard that the third time.
But keep going. You can ask it again.
No, no, no. This whole show is for my personal education.
I can see that. But remember, number one and number two are also a judgment call. You could have one person call A number one and B number two, and I could have B as number one and A as number two because I’m looking at a different set of criteria. There’s no obvious criteria for deciding who’s number one and number two. They could have 30 customers and the same amount of ARR. Which one is number one and which one is number two?
Are you a tennis fan?
Yeah.
I’m not really, but—
But sure. Why don’t you use a tennis analogy?
I’ve got to tell you, I had dinner one time with a mental coach to some of the top tennis players. When he says “mental coach,” he means he talks to them about what’s going on in their minds. Other people help them with their swings.
He said something fascinating. Take the top 200 tennis players. The difference between number one and number three is negligible. The difference between number four and number 200 is smaller than the difference between number four and number three.
What appears to me from the outside—“That person is number four, that person is number three or two”—seems like a small gap. He said it’s actually a lot bigger when you look deeper and understand what’s going on in their minds. It’s not that number three and number four are physically that different. A lot of it is in their minds.
I’m fascinated by this in sports and business: from the outside, people can look similar, and then you spend time with the founders and see why the gap between one and their next competitor is so large.
As you said, the criteria differ. Some companies have clear attributes that make them number one or number two. One might have twice the ARR of the other, and then it’s easy to call that one number one. Sometimes they have similar starting points, or they started in the same place and one came at it from a different angle. Then it comes down to which founder is going to lead the company to success. That’s where you get into the mindset of the founders and the leadership team.
I put it differently. My job is to fund once we identify number one and number two. Remember, they also have to be willing to sell.
The people we were talking about before we started recording aren’t selling for any price.
That’s right. Yes, there’s also that.
Yeah.
You’ve got to reanalyze that from the vantage point that these could be alpha founders. They could be winners, but they’re also sending you a signal by saying they’re willing to sell.
Have you talked to Hock Tan about this?
Have I talked to Hock Tan about selling?
About acquisitions. He does a lot of acquisitions.
He does it very differently.
I know, but have you talked to him about this? His mindset is a little extreme. You and I were together—
Educate me. Yes.
You and I were together at Dell’s house. That’s where I heard the story about him advising a younger founder who was also there. He said, “Just make sure that on day one of your acquisition, the acquired company knows that you’re the conqueror and they’re the conquered.”
Then I asked him—
That’s a different philosophy.
I know you do. I know you do. Then I asked him, “Hock, why don’t you do any podcasts? Do you want to do my podcast?” He said, “Oh, no. My wife forbids me to do podcasts.” I asked why, and he said, “She said I’d get canceled.”
I say the opposite. I tell my team that when you’re starting a negotiation with a company, remember they’re going to be colleagues. Don’t treat them with disrespect.
He’s obviously an extreme character. I wasn’t taking it as disrespect, but—
So which one’s right? Which one’s number one and which one’s number two in your mind?
I obviously know that. Think about it. There’s no way you could think about it the way he does, because you just said, “This guy is coming in and now he’s your boss.”
No, but look, my point is that I see it differently. I cannot win if 20 people around me don’t feel they’re winning. For them to believe they’re winning, they have to believe they’re first-class citizens. They have to believe they’re part of the team and that their winning matters. Their winning matters as much to them as it does to me.
Have there been any acquisitions you did too early?
When I walked in, we had just bought a company called Evident.io. Palo Alto bought it before I got there. Then we ended up buying a company called RedLock, which was in the same space. The difference was that one only worked on Amazon, while the other worked on all three clouds. I suspect the first one was bought too early. My predecessor bought it, but it didn’t matter. The company bought it too early, and then we bought the second one in the same space, which was further along.
Sometimes you can buy things too early. It becomes a question of price, stage, and conviction. You have to marry those three. Am I convinced this is going to be a big space in the future? If so, I need a footprint in that space. Sometimes I can spin up my own team, charge ahead, and make it work, or I can go out and acquire something.
If I acquire something, I have to ask how long they’ve been in that space and how expensive it’s going to get. I think the current times are the clearest example of the risk. Today, companies are getting funded at valuations I don’t understand. Imagine if I wanted to acquire a company today: the same company would have been one-third or one-quarter the price two years ago, in a different funding environment. They’re probably just as early, but three or four times more expensive today.
So, what do you do?
You buy them if it makes sense for your strategy. Look, M&A is not a strategy. M&A is a mechanism to validate, amplify, or execute your strategy. People always ask me, “What’s your M&A strategy?” I don’t have one. I buy companies when I believe I need to win in a category and my team isn’t where I’d like it to be at that point in time. I look for another team that’s got it figured out and has the right people in place. So, you go buy it. If something’s worth $200, it’s worth $800, because my job is to take the $200 asset and turn it into a $5 to $10 billion asset.
If something’s worth $200, it’s worth $800. What does that mean?
As a $300 billion company, I should be buying things that I believe will be worth $5 to $10 billion for us in the next 5 to 10 years.
8. Egon Durban and the 29-round cricket auction
So, you’re not going to remember that other $500 million. This is Egon Durban. In Ari Emanuel’s book, one of the most memorable parts is when they’re buying WME-IMG. They wanted to pay $1.9 billion or $1.8 billion, and it went to $2.2 billion, something like that. I think maybe Ari’s team was kind of fucking up the deal. Egon comes in, with way more experience than Ari, and says, “You’re never going to remember the $400 million.”
Yeah.
He’s like, “Is this an asset we want or not? You’re going to remember fucking up the deal, but you’re not going to remember the $400 million.”
Yeah.
And they obviously closed the deal, and it’s been way more valuable to them.
Egon and I bought a cricket team together in London.
I thought you did that with Devesh.
No, no. Egon and I are equal owners. He and I came up with this together. He called me when I was about to go play golf at the AT&T. He said, “How’s the auction going?” It was about 6:00 in the morning because it was in London. I said, “The last auction went for 2 rounds. We’re at round 24.”
Wait, how many more rounds? You were at 24. How many did you have to go?
We had about 29 rounds.
Well, once he said that, you were like, “Fuck it, we’re getting this.”
Because he was like, “Yo, what’s the worst we’re paying? Like, half of it? It doesn’t matter, because either this team is going to be a huge success or it’s going to be a failure.”
So, he did it to you, too.
Yeah, but I did the same thing. I was trying to be cute on the other one. He just reminded me, “You know, you say that in your professional life. Think about your personal life, too.” So, that’s what we did. I saw him at the range at 7:00, and he said, “Do we own the team?” I said, “We own the team now.” It happened. So, yeah, he’s consistent in his point of view.
I’ve heard this phrase differently before, and I’m curious if you agree with this characterization: You’re not buying companies, you’re buying founders.
You’re buying teams. You’re buying teams. You’re buying their point of view. You’re buying their strategy. You’re buying their execution, right? A company is just a construct around the team that allows us to be held in a legal entity, protect IP, and eventually turn that into go-to-market capability. But effectively, what is a company? When you come up with an idea, your first instinct isn’t, “I’m going to start a company.” Your first instinct is, “I’m going to go build something amazing.” When you go build, you call yourself a builder. You find like-minded people who share the vision. Then you’re the most charismatic person in the room because you’re trying to convince people to get on board with your story, because you had the idea. Typically, it’s 1, 2, or 3 people, so you’re doing everything. You’re a builder. You’re building. You’re validating a strategy every time you hire the next person. You look them in the eye and tell them, “This is what we’re going to build.” The person gets excited. You’re hiring smart people, and they also have to buy into it.
So, you’re basically building a team of like-minded people. You’re bashing out the edges, aligning the strategy, and getting to a point where you say, “Okay, now we all agree we’re going to build this. Now let’s go to warp speed and build the ship, because you’re going to beat everybody else to the punch.” That’s what you’re doing. The company is just a little legal entity. Think about most of these guys: They probably have 1 administrative person who does HR, finance, facilities, and brings coffee. The company is a construct in which you organize. But you’re actually buying a well-functioning, cohesive team. That’s what you test for. Can this team execute better than other teams out there?
When you think about it like that, you say, “I’m going to compete with that team that spent a year and a half getting together and training together. They’re on the field every day. They’re living in the same house. They work 80 hours a week.” You say, “I’m going to beat them by hiring a bunch of corporate people, going through my HR team, going through 4 interviews, and going through internal processes.” It’s like, “Holy shit, you’re never going to beat them. They’re far ahead.”
A couple of things are popping into my mind while I’m listening to you. Have you ever heard of this guy named Henry Singleton?
No.
Henry Singleton was essentially the proto-Buffett. He ran this very successful conglomerate called Teledyne in the 1960s. I found him because when I miss somebody, I read everything ever written about them and everything they’ve written. They all tell you the people they studied. Buffett and Munger kept talking about this guy named Henry Singleton. Munger said Henry Singleton was the single smartest person he ever met. Buffett said it was a crime that American business school students didn’t study him. His returns were utterly ridiculous. A lot of the ideas I thought Buffett used for Berkshire were actually things he learned from Singleton at Teledyne.
He grew by acquisition, just like you did. At one point, I think he made— I should know this; I just did another episode on him. Let’s say 135 acquisitions. I think at one point he owned 130 companies. 129 of them were profitable, so he bought really well. But then he changed his strategy and stopped acquiring. From there, he never acquired another company for about 15 years. I’m thinking about him while you’re talking. Is there a chance that Palo Alto Networks stops acquiring?
Possibly. I said it’s not an existential strategy for us. It’s a way we can maximize the value of the business. It’s a way we can maximize the outcome for our customers. At some point, if you believe we’ve reached a point where the marginal value to accrue to the customer comes from us making sure all of our stuff works at a customer, maybe we focus on that. There’s no hard-and-fast rule that we have to keep acquiring.
Is anything going on in AI changing how you think about acquisitions?
Yeah.
Can you say more about this?
This is the most transformative technology we’ve ever seen in our lifetimes, and possibly will be. If you look at the impact it’s already having, it’s fundamentally changing how everybody works. Let me put it differently: When I was at Google, we democratized knowledge or information. This is democratizing intelligence. We all have something in our back pocket that gets more and more intelligent over time. So, we basically have access to intelligence we never had.
I’m pretty sure that in 2 years, AI will get to a point of reliability where you’ll trust it as much as an expert human being on the topic. You’ll talk to your nutritionist, talk to AI, and say, “You know what? My nutritionist isn’t smarter than my phone giving me nutritional advice.” The same goes for my physiotherapist, my physical trainer, or my trainer. You’ll suddenly find that AI has so much data and so much tribal knowledge that it’ll be able to make decisions. You’ll be able to say, “Listen, give me advice that’s at the frontier and carries a little risk, or give me mainstream advice.” You’ll be able to ask, “How much intelligence do you want? How much differentiation do you want in the intelligence you get?”
That’s going to be true. It’s going to fundamentally change the way we work. It has to, right? A lot of organizations are structured so that work is broken down into silos. Even product development is split across 8 different teams. You have documentation people, testing people, coding people, product people, and UI people. So, you have 8 different sets of people who combine to build a product, which makes no sense. We did it because we created specialization: Nobody could be great and amazing at everything.
But if you believe you can harness all this intelligence so that one person could master all 8 crafts, because they have very able assistants that will do 90% of the work for them, it fundamentally changes the way we get shit done, the way we build things, the way we do anything in our lives, and the way companies are created.
So, that’s going to be true. The shape of a company, the shape of the organization that we run today, is going to be fundamentally different 5 years from now.
Does AI make cybersecurity companies more valuable?
Yeah, I think even more so. More technology and more interconnectedness make cybersecurity companies more valuable. Look, cybersecurity didn’t exist 25 years ago because there was no connectivity. You had a mainframe, you logged in, it was secure. There was no problem. The problem is that now David can log in from here to his bank account, and that becomes an attack vector.
The more people who can log into something from everywhere just expands the attack surface. I saw this morning that Facebook has 3.65 billion daily users. There are 3.65 billion people getting into their infrastructure every day. Every door into their infrastructure is a potential cyberattack surface.
Across the world, with 8 billion people, I don’t know—we must interact with hundreds of consumer applications. There must be trillions of interactions every day online that allow you to interact with some enterprise infrastructure. The more you do that, the more likely you need security across all those vectors. The more technology changes, the more capability you need around security.
For example, the products that will secure AI don’t fully exist yet because AI is so new. So, yeah, it’s good for our business. The more you connect things—
You’re more of a stoic kind of person. You seem like you don’t get stressed. You said your business doesn’t stress you, which is very unusual. You said life shouldn’t stress you. You should wake up every morning, do the best you can, and then remember that if you grew up in India with Hinduism, there’s a strong sense of fatalism. All you can do is your best.
Okay, I don’t know about this. Explain.
The rest is—
Explain how you grew up with this fatalism thing.
Remember, you’ve got 1.8 billion people in the country. You’ve got to make sure they all understand that they can do their best, and eventually there’s a hand of God that dictates what happens. Or it’s the notion that there’s uncertainty. You can’t create full certainty. Things will happen to you that you can’t control. The only thing you can control is what you do.
If you do what you do really well, then you should sleep well, because you can’t worry about everything that can happen to you from an uncertainty perspective. That is the sense of fatalism. That is what keeps society somewhat orderly.
But you do understand it’s easy to say that and very hard for a lot of humans to apply.
Well, look, that doesn’t stop anyone from trying, right? We try to control every outcome around us. We try to minimize risk around us. We all do that. But at some point in time, you’re going to say, “I’ve done my best.”
Did you have this ability when you were younger and earlier in your career?
No.
This is like—
This is recently found, in the last 5 years. Yes.
Did you say this is just in the last 5 years? So, you’re working at Google. High stress? More anxious person?
No. Google’s the best business. The search business keeps going.
Still, that was the greatest. That seems to me—maybe I’m wrong—you had a lot of success before that, but you’re working at a once-in-a-100-year company at that point. You were young. You were in your 30s, right?
Yeah. It can get to your head.
Yeah. It feels like the pressure of this being a huge opportunity, maybe once in a lifetime. This opportunity will change my life.
It didn’t feel like that when you were there?
No. It just felt, “Wow, this is a great rocket ship that’s working.” Remember, you’re 34. You’ve never actually worked in any other company of a similar ilk. You’re now at the center of attention in technology, and everybody’s searching around you. Every morning you wake up, there are more searches. Every morning, more people want to advertise. So, it’s an execution problem. It’s like, “How can I get there?”
When I joined, people were spending money at Google in Australia with an American credit card. That was the only way you could buy advertising in Europe. Can you imagine? “I’m so sorry, I can’t collect your money. You’ve got to go get yourself an American credit card in U.S. dollars and then pay me.”
I’m looking and saying, “We’re making $10 million in Austria. We don’t accept local currency. How the hell is this happening?”
That’s a different business problem, right? You’re just basically saying, “Let me see if we can get out of the way.” So, you’re trying to eliminate friction and create scale, which is a great learning for me. I said, “Great, that I can do. I can eliminate friction and create scale. I don’t have to worry about demand.”
I used to joke—I used to go to customers and say, “What are you here for? What do you do?” They’d say, “I’m the head of sales for Google Europe.” “What does that mean?” “I sell advertising.” “Great. So, how much is it?” I’m like, “I actually don’t know. It depends on what the last guy bid. You’ve just got to bid more than that.”
“So, you can’t tell me what it’s going to cost?” I’m like, “Not really. It depends on how much you’re willing to bid.” “Great. How much can you take a day?” “I don’t know. It depends on the number of people who search, because we can only serve ads when somebody searches.” “So, you don’t know volume? You don’t know price?”
Yeah.
“What kind of sales guy are you?”
I’m selling ads. I don’t have control of volume, and I don’t have control of price. Price is controlled by bidders, and volume is controlled by people who search.
Mhm.
I could tell you when it rained in France, we used to get a lot more revenue because people sat at home and searched.
So, you’ve still got to go to customers and say, “Listen, the way it works is you work out an ROI. Every search is worth something to you. If I send you a search, look at your conversion rate. If you convert that into an insurance sale or an e-commerce transaction, figure out how much you’re willing to pay for a lead. When a lead becomes a conversion…”
We used to sit and do ROI math with our customers so they could understand what their maximum bid should be. Then we’d say, “Now go give us all your money, and as long as it hits below your bid, take as much search as you can.” That was my job.
It’s pretty straightforward. I’m friends with Adam Foroughi, the founder of AppLovin, and he’s very simple: “I sell revenue. If they’re giving me a dollar, I have to give them back more than that.” It’s very easy to sell as long as I can give them back more than that.
We’ll go back to Google in one second. I want to get to this, though. From looking in, it seems like because of the proliferation of AI—there’s, I think you were just on my friend’s show. I’m friends with John and Jordy from TBPN, and I caught a clip of that today. There was some kind of—
Yeah, there was a discovery of a vulnerability called PHP, which is used in 22.5 million places and is being used right now.
Yeah, and my point is, you’ve got to be waking up every day hearing more crazy stuff. It’s only going to get crazier in the future. You seem not to allow yourself to be stressed by it, which, from the outside, seems like this would be the most intense, stressful part of your entire career. Why am I wrong about that?
But I can’t do anything about it. We all have to go figure this out. We’re going to fix it. But stressing is not going to change the outcome. The outcome is going to change by us getting on and doing something about it, right?
You all discovered this morning. I found out about it yesterday morning. Our teams went and looked at everything we had to make sure we didn’t have the vulnerability. We found it. We blocked it. We found the patches. We patched ourselves. We said, “How do you protect our customers?” The team worked overnight and built a patch.
When it came out this morning, they announced their patch. We announced our patch. We put it in the public domain. We’re getting on the phone to any customer who has a problem. We solve it. That’s what we can do.
So how do you describe this philosophy?
I’m just saying: wake up every morning, do the best you can to the best of your ability, leave no stone unturned, and then let fate do the rest.
Is this real?
Yeah.
Well, okay. When were you the most stressed or anxious in your life?
When was I the most anxious in my life? When I had to give a speech at my daughter’s wedding.
Is this the daughter who just started the fund?
Yeah, this is the daughter who got married in February in India.
What’s her name?
Aisha.
Congratulations.
She used to work in marketing at Brex.
Wait, she used to work in marketing at Brex?
Yes. Kids don’t take advice from their parents for the most part. But I’m blessed that once in a while she’ll listen to something and process it differently. I said to her, “You know, what excites you?” She said, “Dad, you know, it’s exciting when marketing doesn’t seem to be the most important function in a company. So, working in marketing, what should I be thinking about?”
I said, “If you look around here in Silicon Valley, the most exciting jobs are for people who get to talk to founders, understand what they do, and get excited about business. To be excited about business, you’ve got to think about venture capital.” So she thought about it. She did her own research and figured out that she was going to go work in venture capital. She worked with Bill Maris at Section 32, and then she was getting married this year and said, “I think I should start my own fund.”
I said, “That sounds like a great idea. You should.” So she’s been through the motions. She’s done a lot of research and figured it out. She started going around thinking about who would invest in her. Then she said, “Will you be my advisor?” This is the best thing a father can get. Now she calls me once or twice a day. We talk about companies and what she’s thinking about. It’s a blessing. It’s great.
9. Coming to America with $200
That’s beautiful, man. When you think about your own childhood, was this innate drive you had to achieve largely internal, I would imagine? Did you get a lot of pressure from your parents?
No. My parents didn’t tell me what to do. I used to come home from school and do my homework. My mom would be happy that I did it. She’d be chasing my sister around, trying to find out what homework she had. I got my shit done.
My dad didn’t know how I got into university. He didn’t know I got into business school in the U.S. I just went to him and said, “I need $2,000.” He said, “I don’t have that.” I said, “Can you give me $200?” He said, “Yeah, I’ll take $200.” I took $200 and showed up here. Then he sent me $2,000 afterward to pay part of my tuition, and that was it.
You don’t put any pressure on your kids?
No.
How different are you as a parent from your parents?
That was a different era. At that point in time, parenting was different. Parenting is different now. Parenting in India is different. Parenting in the U.S. is different. You obviously have all those differences that come into play. I have 2 younger kids who are 11 and 10, so I’m more involved in their lives than I was with my 29-year-old.
Why?
More time. I consciously try to take out time, and I’m sure my kids will say it’s never enough, and my wife will say it’s not enough, but I try to take out time as much as I can. So it’s different. No, I don’t believe in putting pressure on my kids. The kids have to be able to express themselves and do what they believe is best.
But I have seen something bizarre. I don’t know how many people have been born since humanity started. It must be in the tens of billions, right?
No, I think we read about this a long time ago. It was over 100 billion.
Okay, 100 billion people have been born. It’s a well-known problem. Every kid has a parent by definition. Even now, we struggle with parenting as a problem because there are probably so many variants of parenting that most people get it wrong.
I just want my kids to be happy in their lives. If they’re happy, my mission is accomplished, because I’ve seen so many scenarios where kids are not happy, and it’s not clear where things go wrong. It’s not clear if it’s a parenting issue, or if it’s in the DNA or in the circumstances. I just want to make sure my kids are healthy, happy, and balanced. If I can achieve that, they can do whatever they want.
I love the framing you just put on there. Everybody’s had a parent. Parenting is an old thing. It’s been around, and we still don’t have it figured out.
10. Raising kids who didn't grow up poor
No. Show me the ultimate guide to parenting.
I’m particularly interested in this because most biographies that I read are about people who usually start out poor. There are some people who are born into wealthy families and wind up achieving a lot. Obviously, most of humanity isn’t wealthy, so most of the people you’re going to read about weren’t. There’s something they almost universally struggle with: “I was the son of a poor man, and now my kids are the sons of a rich man. What do I do about this?”
Nothing. You created a distinction by yourself. When you were the son of a poor man, there was somebody way poorer than you.
Mhm.
From their vantage point, you were the son of a rich man. I think we all get born into a circumstance. It’s very dangerous to take that circumstance and keep reminding your children, “Remember, I was poor. You should be more grateful.” That’s what happens, right? Especially for people who were poor before they got to where they are today, they feel this notion that they have to keep reminding their kids, “Well, it wasn’t like this when we were born. We used to do this, and we used to do that.” Like, dude, get over it.
They’re born in this circumstance. Figure out a way that they can enjoy it and make sure they learn how to yearn for more, how to have drive, and how to achieve. You’ve got to be careful. I think you’re blessed if you can find a way for your kids to want to achieve, to build, to make things happen on their own, and not take their circumstances for granted. The hardest part of parenting is making sure your kids have drive and ambition, and want to achieve.
My son is obsessed with basketball. Great. Go for it. Let him follow his passion. My daughter wants to be a gymnast. She’s got a tremendous work ethic. She’s great. He’s extremely friendly. He likes people. He likes to hang around people. He knows a lot about what he wants to know. That’s great. If my kids can have passion, chase their passion, and show ambition and drive, I don’t care if he becomes a basketball player or a basketball commentator, as long as he wakes up every morning happy and feels that he gave it his best.
I completely agree with this perspective. Of all the examples I’ve read about, I love what Sam Walton said in his autobiography, which he wrote when he knew he was dying, with cancer all over his body. He said, “I know I’m a fairly overactive fellow,” which is the understatement of the century. “I don’t put pressure on my kids. Let them do what they want. Why would I want them to be like me?”
It’s kind of interesting. If you do it the right way, your kids will watch your behavior. They watch more behavior than you think. They know more about how we are and who we are than we can ever understand. Sometimes they’ll want to do certain things like you. But if you try to force them to do certain things like you, you’re never going to get there.
I can see my kids would like to do certain things like me, and that’s great. I’m very happy that they want to. I’m sure sometimes they look at me and say, “I never want to do that thing my dad does.” That’s fine, too. As long as they can figure out what parts of what I do they like and what parts they don’t like, they’re fine. I’m okay with that.
11. Why AI hacking has been good for Palo Alto
I’m thinking back to all the other people I’ve had on the podcast, and I feel like AI is changing every single business. Tobi Lütke was on this show earlier this year. I’m going to see him again in about 2 weeks to record. His whole point was that we’re going to look back on 2026 and say that, because of AI, every business was up for grabs. Every business is being transformed somewhat, right? But I feel like yours might be the most transformed of anyone I’ve had on the show.
No, no, no. Let’s break it down. Cybersecurity is somewhat straightforward because I don’t come from that background. I told you that, 8 years ago, those were 2 different words in my mind. Cybersecurity is the idea that a bad actor is trying to get into your infrastructure and exfiltrate something important. Let’s take that as a premise. If I know someone is bad, I can stop them at the perimeter, right? A bad guy shows up at the door. We know he looks bad. He’s got a hoodie, he’s got a gun. This is bad. We know how to stop them.
Half of cybersecurity is stopping bad guys and bad things at the perimeter. Ten to 20% is analytics: how to get the hygiene right once they’re inside your house, to make sure things are solid. The remaining 20 to 30% is what to do if somebody gets in. Things will get through your perimeter and into your infrastructure.
How quickly can I find the bad guy? And how quickly can I eject them from my infrastructure before they can do harm? That’s cybersecurity in a nutshell.
How does AI impact this? Let’s start with your comment about vulnerabilities. AI says, “I can find holes in your perimeter faster than anybody else in the world. It used to take days to find them; today it takes forever. I can find holes to enter your perimeter.” Okay, great. My perimeter just became more porous. Now, AI isn’t going to patch the perimeter. I still have to look through the infrastructure, find where the porous parts are, and patch them.
I could be a software company or an infrastructure provider, because everybody’s IT stack is made up of tens of thousands of vendors. Vulnerabilities exist in our software, and our customers also build software themselves. That’s step one. That’s what we’re seeing right now. Models like OpenAI’s new models and GLM-5.3 have immense capability for finding vulnerabilities. Bad actors can rent them cheaply, throw them at you, and find everything you have vulnerabilities in. So what’s the call of the hour? It’s to find a way to patch these things swiftly so the bad guys can’t get in.
Guess who’s building the service? We’re building a service called virtual patching. We write signatures and put a bandage around the entire perimeter. We know where the holes are; these people can’t get in. Over time, our customers will patch the underlying systems, and they’ll be fine. That’s one problem. The good news is our business has gone up since Mythos came in.
That’s what I mean. That’s what I’m trying to get to.
Holy shit. Can you help me patch it?
Well, guess what? Yes, I can help you patch it. I inspect everyone’s perimeter. You’ve got a laptop somewhere. I live on a laptop. I charge you $25 a year, and I watch 150 megabytes a day, give or take, on your laptop to make sure nothing bad is coming in.
You told me you’re hacking my phone calls.
I was just kidding.
I wouldn’t know how to do it. Anyway, remember, cybersecurity is two words. The marginal cost of inspecting traffic on a laptop is zero for me. I don’t charge you more for the traffic that goes through. AI is set up to charge you every time you run traffic through it, in terms of tokens. If AI tried to do my job at the endpoint, it would cost you $5,000 a day. Or pick a number: $500 a day. It’ll become $50 a day. They can’t do what I do at $25 an endpoint.
If you go back inside, and the bad actor gets in, what happens? Yes, AI can help you analyze all the anomalous behavior quickly, but you still need harnesses, security expertise, data ingestion, and infrastructure. So, in a nutshell, our business is not at risk.
Yeah, from the outside, that’s what I was assuming was the case. That was what I was trying to get to.
Assume that CEOs are trying to maximize their businesses from the outside. Generally, that’s a good rule.
Yeah, but we talked about how no one wants to lose. A lot of people lose every day, so this is what I was trying to get to. Who do you go to when you want to learn about the frontier of AI? Who do I go to when I want to learn more about AI?
There are different people. I spend at least four or five hours a week with AI founders of some sort. These kids are young. They possibly worked at OpenAI; they probably worked at Google somewhere. They’ve got these companies. They’re very kind; they spend time with me and try to tell me what they do.
I’ve learned about open-weight models, training, training data, and guardian agents. I go around and talk to tons of people. I talk to probably five or ten cybersecurity startups a month, maybe 15. I try to absorb how they’re thinking about it because, in a bizarre way, this is a new and developing science, for lack of a better word. There are no real experts. People know how to do their work. Everybody else is trying to imagine the implications for everything around them, and they’re trying to build a mental model so they can build a product and win in this category.
There was no concept of data labeling before people started having to label data for self-driving cars. Now that’s turned into reinforcement learning, simulation models, and neural networks that are trying to simulate data so we can train AI. I literally have to stay in the conversation with tons of people to understand and absorb what people are talking about. You just have to spend a lot of time on Twitter, talking to people, and reading stuff. You kind of halfway understand how this gets through the other side. I can’t absorb everything. That’s how I learn.
You brought up another thing I wanted to ask you about. What’s behind your new presence on X?
Well, I think I said that I was giving one of your fellow podcasters a hard time.
Harry?
Yeah. I was giving Harry a hard time, saying, “This is great for you. You keep bringing all of us in, you build your own brand, and all of us are effectively props for your brand.”
He’s like, “Well, that’s your fault, not mine.” I’m like, “What do you mean?” He’s like, “Well, you get the right to build your own brand if you want to. You just don’t.” I’m like, “How would I do that?” He’s like, “Well, go tell people on X what you think. You have a lot to say in a podcast. Why do you wait for me to ask the question?”
That’s a good point.
Then he told me to tweet every day. I’m like, “Okay.”
I’m enjoying your tweets. I watch you almost every day.
No, it’s a lower frequency now. I listen. You can only put your foot in your mouth if you say something too often, because you almost have to not feel the pressure to do that. I like that advice: don’t feel the pressure to say something all the time. You’ll end up making it a job. I don’t want to make it a job. I want to do it when I have something I want to say. I think I say stuff when I want to say stuff. People ask me how much time I spend and who writes it. Actually, I do. I should write it on my guitar.
It’s obvious that you write it. Come on, people.
Once I had it checked by Claude or something, and it made some grammatical edits. People gave me so much grief.
Oh, I saw that one.
Yeah, you’re right. My idea is that I edit and check it, and I’m never going to do that again.
No, you’re really good at this. In fact, there’s a guy that’s always on this camera who gives me a heads-up when we hit a certain time. He just gave me the heads-up, and I was shocked at how fast this time went. So that’s usually a good sign.
It’s a good sign?
It’s a good sign, man. I really appreciate you doing this. This was awesome. We’ve got to do this every few months.
Sounds good. Thanks for your time.