Baby product imports: strollers are 97% China (HS 8715)

US baby-product imports split into two completely different sourcing worlds, and only one of them is a tariff problem. If you run a baby brand and you've been reading the same "diversify away from China" headlines as everyone else, the first useful question isn't whether to move, it's which half of your catalog the warning even applies to. The Census import data answers that cleanly. Hard goods are a China monoculture. Soft goods already did the work.

This page reads the full-year 2024 US Census imports-for-consumption data across the core baby-product HS (Harmonized System) codes, lines up the 2026 tariff stack against each one, and turns it into the planning view a finance team actually needs: exposure by HS line, not by vibes. The numbers below are import values by country of origin, which is the cleanest public proxy for where a category is made.

For the full cross-category view, see our DTC import origins by category breakdown.

The two-world split: gear versus clothing

Start with the headline. In 2024, 96.9% of US stroller imports came from China: $379.0M of a $391.3M total under HS 8715.00 (baby carriages, including strollers and parts). The next origin down the list, Vietnam, was $5.2M, about 1.3%. That is not a concentrated category, that is a monoculture. The car-seat-class heading tells nearly the same story: 81.2% of HS 9401.80 ("seats other than metal or wooden frames") came from China, $1.097B of $1.350B.

Now flip to apparel. China is only 22.2% of babies' knitted garments (HS 6111, $374.5M of $1.686B) and 17.9% of woven (HS 6209, $57.4M of $321.5M). For knitted, India ($328M), Vietnam ($279M), Cambodia ($218M) and Bangladesh ($152M) each rival or exceed China. For woven, Bangladesh ($100M) actually beats it outright.

That gap is the entire thesis of this post. The de-risking that the headlines keep demanding is mostly finished in clothing and barely started in gear. When I talk to founders running a baby brand this size, the moment that lands hardest is realizing the "China problem" they've been losing sleep over is really a problem about two or three SKUs (the stroller, the travel system, the car seat) and not their whole line. The blended number hides it: across the four core baby-specific lines, China is 50.9% of the $3.75B imported, a figure that's true and almost useless, because it averages a near-monopoly against a minority share.

Strollers and car seats: where the exposure actually lives

The reason gear stayed in China while clothing left comes down to what it takes to build the thing. A sewing line can be stood up in Bangladesh or India in months. A stroller needs tooling, molds, and assembly capacity. A car seat adds federal safety certification (FMVSS 213, the US standard for child restraint systems) on top of that, so a new factory isn't just a manufacturing-ramp problem, it's a re-certification problem. That's why the non-China bench is so thin: $5.2M of Vietnamese strollers against $379M of Chinese ones isn't a second source, it's a rounding error.

One honest caveat on the car-seat line. HS 9401.80 is "seats other than of metal or wooden frames," which is broader than child car seats: it also captures other plastic and upholstered seating. So read the 81.2% as the share for the heading that contains car seats, not as pure car-seat share. The direction is right (gear is overwhelmingly Chinese), the precision isn't, and a finance team should know which it's working with.

Baby clothes already did the work

Apparel is the part of the catalog the headlines are loudest about and the part that needs the least action. Look at where babies' knitted garments actually come from.

No single country owns it. China is nominally on top at $374.5M, but India, Vietnam, Cambodia and Bangladesh sit in the same band right behind, and the woven category (HS 6209) flips outright: Bangladesh ($100.4M) beats China ($57.4M). For a baby-apparel brand, that means the duty math, while it can produce a scary headline rate on the China-made portion, applies to a minority of units that are usually the easiest to re-source. The pattern we see again and again is that apparel founders have already qualified two or three overseas factories and can shift a style's production with a season's lead time, not a multi-year capital plan. That optionality is the whole point, and it's why a baby-clothing brand is structurally less exposed than a gear brand even when its per-unit China rate looks worse.

It's worth saying plainly that diapers are a different animal. HS 9619 ($1.795B) is dominated by Canada ($803M, 44.8%) and Mexico ($377M, 21.0%), which is P&G and Kimberly-Clark cross-border production, with China only 19.0%. It's also not a baby-specific code (it includes feminine sanitary products), so treat it as a North America story sitting off to the side of the China thesis, not part of it.

The data shows brands re-stocked, they didn't exit

Here's the part that keeps operators honest. There was a real dip in Chinese stroller imports, and a lot of "we're diversifying" press got written around it. Then the imports came back.

China stroller imports (YTD December value) ran $294.0M in 2018, climbed to a $407.5M peak in 2022, dropped 29% to $289.0M in 2023, then rebounded 31% to $379.0M in 2024. The 2023 trough lines up with post-2022 de-stocking and the first real tariff pressure. The 2024 rebound shows that brands worked down inventory and then went right back to the same supply base, because for hard goods there was nowhere else to go. The dip wasn't an exit, it was an inventory cycle dressed up as a strategy shift. When we've watched founders try to read their own category's diversification from a single year of softer imports, this is the trap: one year of destocking looks like structural change until the restock prints.

What the 2026 tariff stack does to your landed cost

The duty math is additive, and that's the thing to internalize. The 2026 China stack on a given line is roughly: base MFN duty + a Section 301 layer (commonly around 7.5% for List 4A consumer goods, higher on some lines) + a 2026 reciprocal/IEEPA layer (around 10% as a default). All three are assessed on the same customs value, so they add.

That gives you two very different stories. A stroller carried 0% base MFN historically, so the stack lands it around 17.5% on a product that used to clear duty-free. A China-made onesie already carries a double-digit MFN (roughly 8 to 20% depending on fiber), so it can stack to ~33 to 37%, a uglier number, except most baby apparel isn't China-made in the first place, so that rate applies to a sliver of units you can move.

A worked example for a gear brand: a stroller with a $120 FOB unit cost, sourced 100% from China, at a ~17.5% stack carries roughly $21 of duty per unit. At 50,000 units a year, that's about $1.05M in annual duty on a single SKU line, and because strollers were duty-free before, almost all of it is new cost relative to a few years ago. There's no second source to arbitrage it against, so the only levers are price, absorb, or negotiate cost-share with the factory.

One critical honesty flag: the exact 2026 reciprocal/IEEPA rate is genuinely unsettled. Some mid-2026 guidance suggests parts of the IEEPA layer were decollected and certain reciprocal tariffs on China suspended into November, which would push the blended effective rate around; other reads keep the ~10% layer live. So treat every stack percentage here as directional and as-of-mid-2026, and confirm the live rate at the 10-digit HTS line at entry before you commit a budget to it. The import-concentration data (the spine of this whole analysis) doesn't move with that uncertainty; the tariff percentages do.

The macro backdrop says tariffs really are the swing factor. The BLS Import Price Index for all commodities ran 127.1 in 2018, peaked at 144.5 in 2022, and sat at 141.0 in 2024 and 141.3 in 2025. Underlying import prices are flat. The duty stack is the new variable in your landed cost, not the goods themselves.

The "diversify away from China" headline is right for baby clothes and nearly impossible for baby gear. Strollers are 97% China with a $5M second-source bench; the rebound from the 2023 dip proves brands re-stocked rather than relocated. If you sell gear, the work isn't moving production this year, it's modeling a ~17.5% stack into price and margin now and qualifying a second source for the years it'll actually take to ramp.

The operator playbook: model exposure by HS line, not by vibes

Here's how a finance-minded operator should actually run this, and it's the same advice a fractional CFO would give across the desk.

First, build the stacked-duty matrix per SKU, not per brand. Pull the exact 10-digit HTS code for each product, look up the live base MFN, and layer the Section 301 and reciprocal components on top as separate, dated lines so you can update them when the rates move (and they will). The blended "China is half my imports" number is useless for decisions; the per-line stack is what hits gross margin.

Second, split your catalog into "can move" and "can't move soon." Soft goods are can-move: qualify a second factory in India, Vietnam, or Bangladesh, because the ramp is short and the duty savings clear the switching cost quickly. Gear is can't-move-soon: a stroller or car-seat re-source is a multi-year, capital-and-certification project, so for those lines the near-term levers are pricing, margin absorption, and negotiating a cost-share with your existing China factory rather than pretending you'll relocate by next season.

Third, decide price-versus-absorb deliberately and model the gross-margin point impact at your actual price. A ~17.5% stack on a $120 FOB stroller is real money, and "we'll just eat it" is a decision that should be made on a spreadsheet, not a vibe. When I talk to founders carrying gear-heavy catalogs, the ones who handle this well are the ones who already know, to the basis point, what each tariff scenario does to their contribution margin before the PO goes out, so the tariff is a number they planned for rather than a surprise on the customs entry.

If you want a hand building that matrix and pressure-testing the price-versus-absorb call against your real margins, that's exactly the kind of thing a fractional CFO is for. It's also worth reading the DTC tariff exposure index 2026 for the cross-category view of where this is hitting hardest.

Sources and methodology

The import data comes from the US Census Bureau's international-trade series, imports for consumption, via the timeseries endpoint at api.census.gov/data/timeseries/intltrade/imports/hs. The core variable is CON_VAL_YR (year-to-date import value) pulled at period 2024-12, which is the full calendar-year 2024 cumulative figure. Each HS code was pulled with the country code requested to force a country-level breakout, because omitting it collapses the result to a single "total for all countries" row.

The HS codes used are 8715.00 (baby carriages and strollers), 9401.80 (seats other than metal or wooden frames, which contains child car seats), 6111 (babies' garments, knitted or crocheted), 6209 (babies' garments, not knitted), and 9619 (sanitary and diaper articles). China is country code 5700. All shares are China's CON_VAL_YR divided by the all-countries CON_VAL_YR for the same code and period.

The stroller time series uses the same endpoint and HS 8715.00 for China across 2018 to 2024, taking each year's December year-to-date value as the full-year figure. That's what produces the $407.5M 2022 peak, the $289.0M 2023 trough, and the $379.0M 2024 rebound.

The tariff structure comes from regulatory research into the 2026 HTSUS, returning an additive stack: column-1 general MFN duty, plus Section 301 (commonly around 7.5% for List 4A consumer goods, higher on some lines and subject to exclusions), plus a 2026 reciprocal/IEEPA layer (around 10% as a default). These are directional. The exact reciprocal/IEEPA status was conflicting across sources in mid-2026 (one read had parts of the IEEPA layer decollected and certain China reciprocal tariffs suspended into November; another kept the ~10% layer live), so every stack figure in this post is framed as a planning estimate to be confirmed at the 10-digit line at entry.

Macro import-cost context is the FRED series IR, the BLS Import Price Index (End Use) for all commodities, index 2000=100, annual averages 2018 to 2025. Limitations worth restating: HS 9401.80 is broader than child car seats, HS 9619 is not baby-specific, and 2024 is the latest clean full calendar year, so a 2025 full-year refresh would sharpen the "did they actually exit" trend once that data is final.

Frequently asked questions