Podbit · The a16z Show
Larry Page reportedly said he'd rather go bankrupt than lose the AI race. That mindset — treating AI as existential rather than strategic — explains the unprecedented CapEx commitments at Google and Meta. Winning is the only acceptable outcome, regardless of cost.
Where this was said
At 7:29 · chapter starts 2:50
David George opens with staggering numbers: $1 trillion of data centers in the US, plans to add $3–4 trillion over five years, a data center buildout that already dwarfs the entire US interstate highway system in inflation-adjusted dollars, and a 150x increase in Google's token processing over just 17 months [2] — Gavin Baker "10-point ROIC increase for big CapEx spenders: The biggest public GPU spenders have seen approximately a 10-point increase in return on inv…" 04:45 . The scale sounds terrifying, but Baker is unfazed [1] — Gavin Baker "In the 2000 bubble, 97% of all fiber laid in America was dark — unused and worthless. Today, every GPU is running hot. Baker's ROIC test co…" 03:38 . He walks through two key comparisons: Cisco peaked at 150–180x trailing earnings in 2000 versus NVIDIA's current ~40x multiple, and unlike dark fiber — where 97% of installed cable sat unused at the bubble's peak — every GPU is running at capacity. The most rigorous test, Baker argues, is ROIC: the biggest public GPU spenders have seen roughly a 10-point increase in their returns since ramping up CapEx [2] — Gavin Baker "10-point ROIC increase for big CapEx spenders: The biggest public GPU spenders have seen approximately a 10-point increase in return on inv…" 04:45 . The ROI on AI infrastructure spending, he concludes, has been unambiguously positive — even if future Blackwell spending will be a fair test.
Google reported a 150x increase in the volume of tokens processed in just the past 17 months, demonstrating explosive real-world AI usage.
Cisco peaked at 150–180x trailing earnings during the 2000 bubble; NVIDIA trades at roughly 40x — a much more moderate valuation.
In the 2000 bubble, 97% of all fiber laid in America was dark — unused and worthless. Today, every GPU is running hot. Baker's ROIC test confirms it: the biggest AI CapEx spenders have seen roughly a 10-point increase in return on invested capital since the buildout began. The numbers say this is not a bubble.
At the peak of the 2000 internet/telecom bubble, 97% of fiber laid in America was dark — unused and unlit — the defining sign of overbuilding.
The biggest public GPU spenders have seen approximately a 10-point increase in return on invested capital since ramping up AI CapEx, signaling strong returns so far.
The major hyperscalers collectively generate around $300 billion of free cash flow per year, with $500 billion of cash on their balance sheets, underwriting AI CapEx.
SiteGPT attracted over 1 million visitors and $500K in total revenue without spending a cent on paid marketing. The secret: engineering as marketing — building free tools that rank on Google.
Bhanu quit his first job after just 8 months, moved back to his parents' house to cut costs, and started building. One product sold for $250K; the next hit $10K MRR in its first month.
90% of SiteGPT's Google search traffic comes not from the main product but from ~50 free tools Bhanu built. Each tool targets a low-competition keyword and funnels users back to the paid product.
50,000 monthly visitors become 200 leads, 60 trials, and roughly 15–24 new customers per month at ~$100 average revenue each. Add a $1,700–$1,800 LTV and you have a very healthy SaaS.
Start with a blank Ahrefs search, layer in keyword filters (include term, KD < 10, volume > 1,000), list candidates in Notion, design a CTA linking to your main product, then score by volume, difficulty, build effort, and product relevance. That's the whole playbook.
Marketing feels painful for most builders. Engineering as marketing flips the script: instead of writing cold emails or blog posts, you build things — and those things rank on Google forever.
Don't spend months perfecting before launch. Ship the core feature, get real users, and let their feedback dictate the product roadmap. Premature polish is a trap.
SiteGPT launched and hit $10,000 MRR within its first month. That momentum was so overwhelming that Bhanu sold his existing SaaS, Feather, for $250,000 to free up all his time.
PropGPT launched with 20 downloads a day and strong influencer marketing but hit a ceiling at $1,000–$2,000 MRR. High download numbers masked a critical flaw: almost nobody stuck around after the free trial ended.
We use essential and analytics cookies to run Vuci. To understand how the site is used: Privacy Policy.
Install Vuci on your phone
Add it to your home screen for a faster, app-like experience.
Install Vuci on your phone
Tap the Share button, then “Add to Home Screen”.
A new version is available
Reload to get the latest Vuci.