Podbit · The Prof G Pod with Scott Galloway
Anthropic claims 80%+ gross margins, but strips out model training costs and revenue share with partners. The Financial Times estimates the real number is 40–50%. It's WeWork math — profits before all the actual costs.
Where this was said
At 4:38 · chapter starts 1:51
The episode's meatiest segment opens with Fred's question about whether frontier AI labs can survive the rise of powerful, free local models. Scott methodically layers the data: a UK AI Security Institute test found the gap between open-weight and frontier models narrowing, with Deepseek v4 Pro costing $0.30 per task against $12.50 for Claude Opus 4.5 — a 40x price difference. [1] — Scott Galloway "The price gap between open-weight and frontier AI models is staggering. Deepseek v4 Pro runs 30 cents per task. Claude Opus 4.5 runs $12.50…" 02:25 Only 3% of US households pay for AI, and OpenAI converts just 5% of its billion users to paid plans, signaling that enterprise — not consumer — is the real battleground. Anthropic's claimed 80%+ gross margins are demolished with a WeWork comparison: the Financial Times estimates real margins at 40-50% once training costs and partner revenue shares are included. Scott then delivers his sharpest analogy: open-weight models are doing to frontier AI what Old Navy did to Gap and Southwest did to Delta — 80% of the product at a fraction of the price. [2] — Scott Galloway "The best business model in history is offering 80% of the premium product at 50% of the price — Old Navy did it to Gap, Southwest did it to…" 06:35 The segment closes with a geopolitical grenade: China's incentive to subsidize cheap AI amounts to an AI-dumping strategy that could trigger a US market crash, given that 40% of the S&P 500 is tied to AI valuations. His verdict — frontier models cannot justify their current valuations — lands as a confident, data-backed conclusion.
The price gap between open-weight and frontier AI models is staggering. Deepseek v4 Pro runs 30 cents per task. Claude Opus 4.5 runs $12.50. That's a 40x difference — and the UK's AI Security Institute found the performance gap is narrowing.
On tasks both models completed reliably, the open-weight GLM 5.2 costs $6 per task versus $15 for Claude Opus 4.6, highlighting the price gap between open and frontier models.
Deepseek version 4 Pro costs about 30 cents per task versus $12.50 for Anthropic's Opus 4.5, a price difference of over 40x.
Bank of America found that only about 3% of US households paid for an AI subscription as of February, with adoption highest among households earning more than $125,000.
OpenAI reported 50 million paid subscribers against roughly 1 billion total users in June, meaning only about 5% were paying customers.
Anthropic reportedly tells investors its total addressable market is more than $30 trillion, a figure Scott Galloway dismissed as meaningless without context.
Anthropic has cited gross margins above 80%, but this figure excludes major expenses including model training costs and revenue share with partners, making it analogous to WeWork's adjusted EBITDA.
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.
Data from Scott Galloway's firm L2 showed that top performers shared three traits no one will say out loud: they were disproportionately female, attended elite schools, and had competed in sports. One example he cited: a gymnast from Yale was an automatic hire.
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.