The episode opens with a montage of clips setting up Nikesh Arora as a major figure in cybersecurity, followed by Jason welcoming him to the pod. Nikesh opens with a memorable framing: Google Search democratized information; AI is democratizing intelligence. He reflects on money as a scoreboard rather than an end goal, immediately signaling the philosophical depth beneath the business conversation. The hosts flag that Palo Alto has grown from a $17 billion to a $238 billion company under his 8-year tenure, and tease that the conversation will touch on Mythos, the rise and fall of SaaS, and more.

This is the episode's most electric segment. Nikesh Arora drops a bombshell: Palo Alto, one of the most security-conscious companies on earth, ran Anthropic's Mythos model against its own codebase for six weeks and found vulnerabilities that would have taken five to seven years to uncover through traditional methods — all for a few million dollars in compute costs. ^{[1] — Nikesh Arora "In 6 weeks, we found vulnerabilities which would have normally taken us 5 to 7 years to find." 02:54 Play} He explains that Mythos' 'ultra mode' can chain vulnerabilities together to map entirely new attack paths, a capability that's great for offense but terrifying at scale. Even more alarming: Nikesh estimates that Mythos-level capabilities will be available in open-source or Chinese models within three months. ^{[2] — Nikesh Arora "3 months to open-source Mythos capability: Nikesh Arora estimates Mythos-level AI vulnerability-finding capabilities will be available in o…" 04:30 Play} He notes that models like Llama 4.8 and 5.5 already have similar capabilities, and you don't need to crack the hardest targets — an old industrial OT system is a much easier mark.

With the Mythos revelations still fresh, David Friedberg presses Nikesh on who's actually winning the AI cyber arms race. The answer is uncomfortable: defenders are falling behind. Every software vendor is now showing up at enterprise CIOs' doors asking them to patch newly discovered vulnerabilities — at exactly the same moment those CIOs are scrambling to audit their own codebases. The situation is compounded by open source, which nobody quite knows how to defend at scale, though IBM's $5 billion project signals the market is waking up. Nikesh frames this bluntly: 'Not as well as we should be doing — which is great for our business, but that's a different story.' ^{[1] — Nikesh Arora "AI is supercharging attackers faster than defenders can respond. Every vendor is now showing up at CIOs' doors asking them to patch newly d…" 05:15 Play}

This chapter contains the episode's most sweeping strategic insight. Nikesh carves the SaaS world into three buckets with surgical precision. First: analytical SaaS — any company whose core value is collecting and analyzing your data — is finished. LLMs can now query raw data directly, making the entire middleware analytics market redundant. ^{[1] — Nikesh Arora "Any SaaS company whose core value proposition is collecting and analyzing data for you is finished. You can now run an LLM directly against…" 07:03 Play} Jason illustrates this perfectly with a real example: his firm eliminated 17 of 20 SaaS seats, connected the data source to Claude via Slack, and cut the bill by 90%. Second: infrastructure software — databases, data platforms like Databricks, Snowflake, MongoDB, Oracle — is massively undervalued, because enterprises will need to store 10 times their current data over the next three years to train and defend AI systems. ^{[2] — Nikesh Arora "While analytical SaaS dies, infrastructure software is quietly undervalued. Enterprises will need to store 10 times their current data over…" 08:50 Play} Third: systems of work (CRMs, ERPs, sales tracking) need to be reinvented for an agentic world where agents, not humans, do the data entry — eliminating the need for the UIs that trillions of dollars have been spent building.

The conversation turns to model economics and where value will ultimately accrue in the AI stack. Nikesh offers a vision of AI models becoming pure utilities — you buy 120 IQ for routine tasks, 250 IQ for complex ones, paying fractions of cents versus dollars. The implication: model providers won't own the profit pools. That's why OpenAI is racing to build Codex and Anthropic is building Claude Code — they're attacking the application layer, which is where enterprise software spending already lives. ^{[1] — Nikesh Arora "AI models will commoditize into a utility layer where you buy intelligence on demand at different price-performance tiers. The profit pools…" 14:06 Play} But here's Nikesh's key observation: the AI-native application companies that will replace legacy SaaS are not yet fully formed. Fifty thousand enterprises need the same HR system, the same sales platform — it's wildly inefficient for each to build their own on top of raw model APIs. The opportunity waiting to be captured is building that next generation of enterprise application companies, complete with harnesses, memory, and data pipelines.

The tone lightens as David Sacks runs an armchair CEO segment with Nikesh. He declines to rate Uber (he's on the board). On Waymo: the cars work and they should be in more cities faster. On Google — his most memorable take — he calls it 'underrated' and predicts it will be the first $10 trillion company in our lifetime. ^{[1] — Nikesh Arora "Google is underrated. It has the models, the data, and critically — the largest enterprise sales force of any hyperscaler. Most people forg…" 21:35 Play} His reasoning is structurally compelling: everyone can build a model, but converting model capability into enterprise revenue requires a massive, disciplined sales force. The three hyperscalers — Google, Microsoft, Amazon — have the biggest enterprise sales organizations in the world, and that advantage is systematically undervalued by the market. On OpenAI, he notes they should be selling faster, citing Anthropic's apparently stronger ARR growth driven by an all-in focus on enterprise and coding. A brief, comedic interlude erupts when Chamath tries to walk back an earlier comment about founder vs. hired-hand CEOs — earning Nikesh's theatrical forgiveness.

The episode closes with Nikesh mapping Palo Alto's acquisition strategy. Phase one was elegant: buy product companies, plug them into a high-performing go-to-market engine, and increase revenue per customer at negligible incremental cost. That playbook ran the market cap from $17 billion to north of $150 billion. ^{[1] — Nikesh Arora "For years, Palo Alto bought product companies and plugged them into its sales engine — doubling revenue per customer at negligible incremen…" 28:20 Play} Phase two, signaled by the $25 billion identity security acquisition, is different in kind: it's about using AI to run the most operationally efficient enterprise in the sector. If Palo Alto can drive its operating margins far above industry norms, it becomes a universal acquisition machine — any asset bought at a lower margin can be upgraded to Palo Alto's margin profile, making the deal immediately accretive regardless of the target's adjacency to core cybersecurity. Nikesh closes with a contrarian workforce prediction: despite all the talk of AI-driven headcount reduction, Palo Alto has more technical employees today than it would have without AI, because AI-driven transformation creates demand for engineers faster than it eliminates them.