Scaling tech responsibly: Sridhar Ramaswamy on search, ads, and monopoly power

Executive overview

Google's ad model created a structural conflict: the search engine optimises for advertisers, not users. Sridhar Ramaswamy, who ran Google's $100B+ ad business, left to build Neva — a subscription search engine with no ads and no personal data extraction.

The subscription model changes the incentive entirely. At 1% market share, Neva can be profitable. The goal is not to unseat Google but to prove a different model produces a better product.

The core insight: when a company's revenue depends on advertisers, users are the product — not the customer.

Why Ramaswamy left Google

• Google shifted from search-first to ad-load maximisation over 10+ years
• Organic results in flights, local, and shopping were progressively displaced by Google's own products
• Google developed what he calls "manifest destiny" over all information — wanting to be the ultimate arbiter of truth
• 90%+ market share made this concentration of power alarming, not just competitive

How Neva's subscription model changes search

• No ads means the top result is always organic
• Users can customise: prioritise paid newspaper subscriptions, exclude big-box retailers, filter medical results by source type (government sites, etc.)
• Revenue aligns with user satisfaction, not advertiser spend
• Subscription companies are valued on top-line revenue; a small market share can still produce a viable, high-value business (Costco model)

Competing against platform giants

• Don't compete head-on in their core lane — differentiate on principles they structurally cannot copy (privacy, no ads)
• Big companies announce projects but abandon them quickly if growth is slow; size of investment is invisible from outside
• Identify the segment that cares about what you uniquely offer, not the mass market
• Defaults and distribution are the real moat — getting default search placement on browsers is Neva's hardest challenge

Responsible scaling and the Facebook problem

• Early-stage startups can ignore platform-scale harms; they don't apply yet
• As companies scale, societal responsibility rises — optimising a single metric (engagement, revenue) is no longer acceptable
• Facebook's internal research showed long-term harms; leadership repeatedly chose engagement over mitigation
• The structural problem: scientists cannot guarantee they "can't do better," so fixes never launch without leadership forcing the trade-off
• Ramaswamy's example: Google walked away from nine-figure revenue by banning porn ads — a qualitative call, not a proven ROI

Tech monopoly and government response

• Companies at GDP-rival scale are autocratic by nature — they serve shareholders, not democracy
• Big tech acts as a toll collector on every customer-merchant relationship; combined ad revenues amount to hundreds of dollars per person per year
• Antitrust law exists for this reason; the Sherman Act reflects society's rejection of unchecked scale
• Congress is the wrong body to regulate tech — specialist agencies (modelled on the FDA, FTC, FAA) are needed

Running a startup vs. running a division of Google

• Attention, resources, and platform leverage disappear instantly
• The biggest trap for ex-big-company founders: overplanning
• Startups must iterate fast to find product-market fit; no one cares until they do
• The emotional reality: single moments of external validation carry a week of energy; low points are frequent
• Walking the line between optimism and delusion is a permanent founder condition