Acquisition costs 12% more every year, through a front door half the size.
Golden Hour backs consumer companies whose economics survive that arithmetic, at seed and Series A.
Two numbers define building here.
Both are published by the companies causing them. Meta reported average price per ad up 12% year over year in the second quarter of 2026. Independently, Pew tracked 68,879 real searches and found that when an AI summary appears, click-through to a traditional result falls from 15% to 8%, and 26% of those sessions end without any click at all.
Informational and comparison queries, the exact top of a consumer funnel, are the queries that trigger summaries most often. So the paid floor rises about 12% a year while the organic front door narrows by nearly half. A consumer company now has to improve contribution margin every year just to stand still.
Meanwhile the funding data says consumer is back, and the funding data is lying. Consumer AI took $89B across 668 deals in 2025. Ten rounds took $71.5B of it. Everything else split $17.5B, and roughly 95% went to late and growth stage. Pre-seed consumer is down 73% from 2021 and Series A is hollowed out.
That is a vacated stage. We think it is the right one to stand in, and we think it vacated because the underwriting got harder, not because the returns disappeared.
Three things we believe.
We underwrite contribution-margin LTV, or we do not underwrite.
Almost every consumer deck shows LTV to CAC around 4:1. Recompute it on contribution margin instead of gross margin and the same business is usually near 2:1. Our bar is top quartile, not median: contribution margin above 28%, first-order payback under six months on a contribution basis, gross margin above 70%, and twelve-month repeat above 35%. We will show you our math on the first call so you can argue with it.
The sector is funding the inverse of its own returns.
Across the last decade, consumer hardware returned 8.3x and health, wellness and lifestyle returned 2.4x, the worst of any consumer segment. Health and wellness just produced the two largest consumer health rounds ever recorded. We would rather own the form factor than the category label. Of 31,410 consumer tech companies funded over that decade, 0.4% reached a billion dollar exit, and the top 11.7% of disclosed exits produced 87.2% of all exit value.
Merchants won round one of agentic commerce.
In March 2026, five months after launching Instant Checkout, OpenAI deprioritized it in favor of merchant-controlled apps. The reason is the reason it will keep happening: checkout is inventory, subscriptions, tax, merchandising, and the transaction data flywheel, and merchants will not hand that over. We invest in the merchant side of agentic commerce, not in the agent that was supposed to own the customer.
Five places we look first.
Two passes we make every time.
Any plan that requires paid social CAC to hold flat. That is not a risk. It is a disclosed, measured headwind with a published number attached to it.
Subscription-first DTC, and food delivery, where investment is down 92% since 2021.
Bring the cohort file, not the deck.
We will run the contribution-margin version with you on the first call.
Say hello