LESSON Lead Safe University · Free

Easy Wins · 09 · Baby food brands

Eighteen thousand lots. Ten winners, ten cautions.

California's AB 899 compelled 32 baby-food brands to publish every lot's heavy-metals testing. 18,124 records, 2,381 exceedances. Here are the ten brands with the cleanest aggregate profile, and the ten where you should rotate out of a monotonous diet. But read to the bottom, the brand is less important than the ingredient.

Lots tested: 18,124
Brands: 32
Exceedances: 2,381
Any-metal fail rate: 13.1%

Source: AB 899 reanalysis · data current to April 2026 · full table at /babyfood.

Brand matters. Ingredient matters more.

Top of the list (zero or near-zero exceedances): Pumpkin Tree, Kroger Simple Truth, Brainiac Foods, Cerebelly, Stonyfield, White Leaf, Piko, Lil' Gourmets, Kekoa, Nurture Life. Most are smaller, niche, or store-brand, and several have small sample sizes, which deserves its own caveat.

The worst: Aldi Little Journey (100% of lots tested above an action level), Else Nutrition, Puffworks, Amara, Ready Set Food, Babylife Organics, Nestum/Cerelac, Beech-Nut (3,248 lots, 25% exceedance rate), Happy Baby (1,900 lots, 25%), Sprout Organics. Most of those exceedances are arsenic, mostly tied to rice and rice-derived ingredients.

BUT: the brand is the wrong unit of analysis. The same brand can have clean apple-pear puree and failing rice puffs. Read the ingredient breakdown and our deep-dive at /babyfood.

The cleanest 10

Brands with the lowest aggregate exceedance rate.

Sorted ascending by “exceedances per 100 records”, the share of all tested lots that exceeded any AB 899 action level for any of the four metals (lead, inorganic arsenic, cadmium, mercury). Filtered to brands with at least 10 lots reported.

Brand	Lots	Fail %	Worst metal	Peak ppb
Pumpkin Tree	549	0.0%	lead	12.0 ppb (cadmium)
Kroger Simple Truth	260	0.0%	lead	9.8 ppb (mercury)
Brainiac Foods	115	0.0%	lead	9.4 ppb (lead)
Cerebelly	51	0.0%	lead	1.0 ppb (arsenic)
Stonyfield	28	0.0%	lead	0.0 ppb (lead)
White Leaf Provisions	25	0.0%	lead	18.0 ppb (cadmium)
Piko Provisions	24	0.0%	lead	0.5 ppb (lead)
Lil' Gourmets	15	0.0%	lead	13.5 ppb (cadmium)
Kekoa	14	0.0%	lead	12.0 ppb (cadmium)
Nurture Life	159	1.3%	arsenic	11.2 ppb (arsenic)

“Worst metal” = the metal with the most exceedances for that brand (or the most violations on the brand's max ppb if exceedances are tied at zero). Peak ppb = the single highest test value across the brand's lots, with the metal in parentheses. Several of these brands (Cerebelly, Stonyfield, White Leaf, Piko, Lil' Gourmets, Kekoa) have small sample sizes, their zero exceedance rate is real but lower-confidence than for high-volume brands.

The worst 10

Brands you should rotate out of a daily-staple diet.

Sorted descending by exceedance rate. None of these are “don't ever buy”, the AB 899 action levels are conservative and most exceedances are arsenic in rice-based products, which is the entire industry's structural problem. But if your kid is eating only from one of these brands as a daily staple, this is where rotation moves the needle.

Brand	Lots	Fail %	Worst metal	Peak ppb
Aldi Little Journey	76	100.0%	arsenic	98.7 ppb (arsenic)
Else Nutrition	55	67.3%	arsenic	33.8 ppb (cadmium)
Puffworks	66	66.7%	arsenic	35.1 ppb (arsenic)
Amara	63	54.0%	lead	33.5 ppb (cadmium)
Ready Set Food	36	50.0%	arsenic	69.9 ppb (arsenic)
Babylife Organics	78	34.6%	arsenic	84.0 ppb (lead)
Nestum/Cerelac	149	26.9%	arsenic	95.0 ppb (arsenic)
Beech-Nut	3,248	25.1%	arsenic	35.9 ppb (arsenic)
Happy Baby	1,900	24.5%	arsenic	137.0 ppb (arsenic)
Sprout Organics	279	21.9%	arsenic	79.0 ppb (cadmium)

Aldi Little Journey's 100% rate reflects that nearly every lot tested was a rice-based teether or rice rusk, the worst possible category for inorganic arsenic. The peak 137 ppb arsenic for Happy Baby is in their Sweet Potato & Banana Teethers (a rice-flour-based product). The peak 84 ppb lead for Babylife Organics is in their Raspberry & Fig Oat Bar, fig is unusually lead-prone (see ingredient breakdown).

The pattern in one sentence.

Brands that lean heavily on rice, fig, sweet potato, sunflower seed, and amaranth show high exceedance rates. Brands that lean on apple, pear, banana, oats, peas, and yogurt show low exceedance rates. The brand-level ranking is mostly a measure of which brands bought the cleanest rice and which ones built their formulations around lower-uptake crops in the first place.

This is why the ingredient breakdown at /babyfoodingredients is the right tool. A “cleanest” brand can still ship a failing fig bar; a “worst” brand still ships clean apple-banana pouches. Switch from one bad brand to another bad brand and you may not move the needle. Switch from rice cereal to oat cereal and you cut arsenic exposure 5–10× regardless of brand.

The three easy wins

What to actually do tonight.

Rotate brands weekly. No brand is 100% reliable across all lots.

Even the “cleanest 10” brands have lots in the dataset that came back at unusual peaks, White Leaf hit 18 ppb cadmium in a single lot, Lil' Gourmets hit 13.5 ppb cadmium, Pumpkin Tree hit 12 ppb cadmium. The aggregate stats are excellent, but variance per lot is the rule, not the exception. Rotation is the cheap, easy way to bound any single bad lot.

Practically: pick three brands you trust, alternate them across the week. Don't make any one of them the daily staple.

The “cleanest” brands still failed 2–4% of the time.

Translation: 96–98% of lots from the top brands are clean, but you don't know which 2–4% you've got without lot-level testing data. Most parents won't (and shouldn't) chase that level of granularity. Rotation does the same job statistically.

If you want lot-level confidence on a specific package, this is where in-home testing earns its place, one drop, painted-decoration glow, or no glow.

Read the ingredient breakdown, it's 5 crops, not 32 brands.

The story of the AB 899 dataset is that five ingredients drive the bulk of all exceedances: rice, sweet potato, amaranth, sunflower seed, and fig. Every brand that uses them inherits their soil-uptake problem. The brand rankings above are mostly a measure of which brands lean on those crops and how much.

So the real action: skip daily rice cereal, swap to oats / quinoa / barley; treat fig bars and sunflower-butter snacks as occasional rather than daily; rotate sweet potato with squash, carrot, parsnip. That single move outperforms switching brands.

Methodology & caveats

How to read this responsibly.

What “exceedance” means here: a lot whose reported value for any of lead, inorganic arsenic, cadmium, or mercury exceeded California's AB 899 / FDA Closer to Zero action level for that metal in that food category. This is a conservative threshold, not a regulatory hard limit, and many lots flagged as exceedances are still well below typical adult-food levels.

Detection-limit caveats: brands report values either as “measured ppb” or “below detection limit”, and detection limits vary across brands (typically 2–10 ppb). A brand that reports a high typical detection limit can hide low-level lead under the floor; a brand that reports a 0.1 ppb detection limit (Once Upon a Farm) is more transparent but its peak values can look higher because more lots have measurable values to begin with. Read the per-brand methodology at /babyfood.

Sample size matters: Stonyfield's perfect record is across 28 lots; Beech-Nut's 25% rate is across 3,248 lots. The big-sample brands are statistically tighter; small-sample brands could move materially as more lots are tested.

This is not a blood-test substitute. Per-lot heavy-metal testing tells you about food intake, not about your child's body burden. If you have a specific exposure concern, ask your pediatrician about a venous blood lead test, that's the actual measure of what got absorbed.

Next easy win