Ads Got Free to Launch. They Got Harder to Win.

The barrier didn’t move. It disappeared.

Until late 2025, launching a real ad campaign took a creative team, a media buyer who knew the platform, and a production budget you had to justify to someone. That was the moat. It kept the auction smaller than the number of people who wanted in.

That moat is gone. Meta’s AI creative tools now build a full ad set — images, video, copy, every placement crop — from a product URL, and advertiser adoption of these tools went from 4 million to over 8 million in about four months, the fastest adoption curve Meta’s ad business has ever recorded. Google isn’t far behind: Performance Max and AI Max now carry over 30% of search ad spend for advertisers who’ve adopted them, and starting September 2026, Google will auto-upgrade DSA and broad-match campaigns to AI Max whether you opt in or not.

Anyone with a Shopify store and a credit card can now have a fully AI-built, fully AI-targeted campaign live in under an hour. That’s not a hypothetical — it’s the median new advertiser on both platforms right now.

What actually changed in the auction

Here’s the mechanism, and it matters because it tells you where to spend your next hour, not just your next rupee.

When production cost collapses toward zero, the number of advertisers bidding in your auction doesn’t grow gradually — it explodes, because the thing that used to filter out casual entrants (cost, skill, time) stopped filtering. And because most of those new entrants are running the same generative tools, trained on the same platform data, their creative starts to look and sound the same. IAB’s 2025 Digital Video Ad Spend Report projects genAI-built-or-enhanced video creative rising from 22% of all video ad creative in 2024 to 39% in 2026 — and smaller advertisers are projecting even higher, around 45% reliance on genAI for their video creative.

Layer on top of that: Meta’s newer ranking systems read creative fatigue with finer-grained signal than they used to, which means a concept that used to have real runway burns out faster now — directionally, not to a precise day count, but consistently reported across the industry as a shortening cycle. Put a flood of near-identical AI variants into an account that’s being watched by a ranking system built to spot “spent” creative fast, and the system doesn’t see thirty fresh shots on goal. It sees one tired concept wearing thirty outfits.

The cost, in the numbers you actually track

You don’t need a rupee figure invented for this piece to feel the squeeze — you’re already watching it in Ads Manager. But here’s what the wider market data around you says about where the pressure is coming from and where it’s going.

Two things sit underneath that table. First, Indian D2C industry commentary consistently describes a 2026 shift away from growth-at-any-cost toward unit-economics discipline — contribution margin, payback period, LTV:CAC by cohort — as the framing investors and founders now expect from performance teams. Second, the clearest sourced baseline for what “slow” looks like in Indian D2C: RedSeer’s analysis found fashion and beauty D2C brands averaging an 18–24 month CAC payback period, against repeat-purchase rates rarely exceeding 30% at 180 days — a pre-AI-surge figure, but the honest floor to measure your account against, not a number to assume has improved on its own.

A note on what we’re not claiming here: a widely circulated figure puts Indian D2C Meta CAC rising from ₹380 to ₹502 in the last year. That number traces to a single agency’s blog post with no external source or methodology — it isn’t used anywhere in this piece as fact, and you shouldn’t cite it as an industry statistic either.

Why more AI volume doesn’t fix it

The instinct, when CPA creeps up, is to make more. If Advantage+ rewards fresh creative, the logic goes, thirty AI-generated variants should outperform five hand-built ones. The evidence points the other way. Those +3%+ and +6%+ conversion lifts Meta reports came from well-targeted use of the tools, not from flooding the account.

A brand that pairs five deliberately brand-anchored, AI-assisted concepts — same visual system, same voice, same product story told five ways — is testing a genuinely different hypothesis than a brand that generates twenty-five auto-variants off one template and hopes volume compensates for sameness. The first approach mirrors the conversion lifts Meta itself reports. The second is the trap the auction crowding creates.

The compliance layer nobody priced in

There’s a second cost most performance teams are treating as a production detail instead of a regulatory one. On May 8, 2026, ASCI published draft guidelines requiring visible disclosure — “Created using AI” or “Enhanced using AI” — for any ad content where AI use could materially influence a consumer’s decision. Comments on the draft are open until June 13, 2026, so as of this piece’s publish date the rules are proposed, not in force — but a synthetic voice-over on a UGC-style ad, or an AI-generated “customer” testimonial, is exactly the medium-risk category the draft targets.

The self-audit: run this before you scale anything further

Score three or more on the checklist below and the fix isn’t a new prompt in your AI ad tool. It’s a brief.

Three or more checked boxes means going back to the product story, the founder’s actual reason customers repeat-purchase, the thing a competitor’s AI can’t generate because it isn’t in anyone’s training data — and building the next concept from there, using AI for speed of production, not as the source of the idea.

That’s the same discipline a visual brand audit is built to surface before you spend another rupee scaling a creative concept that was never differentiated to begin with.

Conclusion and next step

The auction got crowded, not stupid. AI didn’t break ad performance — Meta’s own data shows real, if modest, gains from using these tools well. What broke is the assumption that speed of production is the same thing as quality of idea. Every brand got the same free launch. The ones still winning are the ones who used the time they saved to think harder about what makes their creative worth watching, not to make more of the same thing faster.