Nikesh Arora says Palo Alto Networks will not sign a definitive term sheet until an acquired founder signs a joint product roadmap, and admits it once backed out of two deals.
Arora says acquisitions are a tool for executing strategy, not a strategy itself, and says current startup valuations make deals more expensive than they were three years ago.
Arora predicts AI will become as reliable as a human expert within a few years, and argues connected technology makes cybersecurity more valuable.
Nikesh Arora turned Palo Alto Networks from a firewall company into a platform of more than 70 products, largely by buying companies . Enterprise software's biggest cost is winning customers, so selling many products to the same buyers pays off, and building good products takes years . Acquisitions frequently fail on execution, so his edge lies in how he picks and integrates them .
Takeaways
Cross-selling is the economic engine. Customer acquisition is the biggest expense in enterprise software, Arora argues, so a company that sells many products to the same customers can run a much higher operating margin . Palo Alto grew from one firewall business to more than 70 products .
Buy because building good products takes years. A strong product takes four to seven years to build, so buying is often necessary. Acquisitions frequently fail through bad execution, so Arora studied data on which deals succeed and fail . He likens his portfolio to pharma companies trying many drugs and cutting losses early .
Acquisitions serve strategy and may someday stop. Arora calls M&A a tool for executing strategy and says Palo Alto may stop acquiring eventually. Senra compares this to Henry Singleton, whom Charlie Munger called the smartest person he ever met, who made over 100 acquisitions at Teledyne before pausing .
Timing matters, and target prices have soared. Palo Alto bought Evident.io too early compared with RedLock, which covered all three major clouds. A company Arora could have bought three years ago now costs three to four times more . He seeks targets likely worth $5 to $10bn within a decade .
AI will match human experts within years. Arora calls AI the most transformative technology of his lifetime and predicts it will be as reliable as a human expert in two to three years, changing how companies are organised . He spends at least four or five hours a week with AI founders .
More connectivity makes security more valuable. Arora says Palo Alto's business is not at risk from AI . More connected technology raises the stakes, he argues, citing Facebook's 3.65bn daily users and a PHP vulnerability used in about 22.5m places, which Palo Alto addressed with virtual patching .
How Arora runs an acquisition
Target the number one company in a category, judged on criteria such as ARR .
Negotiate personally without bankers, and refuse to sign a definitive term sheet until the founder signs a joint product roadmap.
Give the founder thirty days after agreement to settle product direction before closing. The rule came from an early mistake, when a founder spent six months debating direction after a deal .
Put acquired founders in charge of their own product areas instead of rewarding existing teams with the prized asset , and keep them three years, ideally four .
Treat acquired teams as future colleagues, in contrast with the more extreme approach Senra attributes to Hock Tan. Founders must still adapt to a company with 6,000 people in the field .
