Grid-scale storage pencils on cost per megawatt-hour, and Tesla has said as much — in narrower, more careful language than the slogan suggests. The FY2024 Form 10-K, surfaced through SEC filings, attributes the year's jump in energy gross margin to two cost-and-mix levers:

"Gross margin for energy generation and storage increased from 18.9% to 26.2% in the year ended December 31, 2024 as compared to the year ended December 31, 2023, primarily due to margin improvements for our energy storage products driven by cost reductions, including benefits from IRA manufacturing credits, and a higher proportion of our storage business, which operated at a higher gross margin, within the segment as compared to the prior periods."

Read that sentence closely, because the numbers inside it are the whole case. Energy generation and storage gross margin rose from 18.9% in 2023 to 26.2% in 2024 — a 7.3-point swing, and the second consecutive year of expansion after the segment printed just 7.4% in 2022. In dollar terms, segment gross profit climbed from $1.14 billion to $2.64 billion year over year, even as segment cost of revenue rose 52% to $7.45 billion on a 16.7 GWh increase in Megapack and Powerwall deployments. Revenue scaled, costs scaled, and margin still widened — that is the signature of a learning-curve business, not a pricing fluke.

The filing names three forces, not one. First, "cost reductions" on the storage products themselves — the manufacturing learning curve every standardized hardware product rides as the Lathrop and Shanghai Megafactories ramp. Second, "benefits from IRA manufacturing credits," the 45X production credits that lower effective cost of revenue for U.S.-built cells and packs. Third, "a higher proportion of our storage business, which operated at a higher gross margin, within the segment" — mix shift away from lower-margin solar toward higher-margin Megapack. Two of those three are durable operating levers; the third, the IRA credit, is a policy subsidy that a careful reader should hold separately, because it can change with the statute.

Does it pencil? The mechanism is sound. Megapack is a standardized, factory-built product, and standardized products ride a manufacturing learning curve — each doubling of cumulative volume tends to pull unit cost down. A higher Megapack mix compounds that, because the better-margin product carries more of the revenue. The 2022-to-2024 arc — 7.4% to 18.9% to 26.2% — is exactly the trajectory the theory predicts as volume compounds. That is real, durable margin, not a one-off.

The caveat the filing itself forces is the IRA line. Strip out the manufacturing credits and the underlying product cost-down is still positive — the 10-K lists it first — but you cannot read 26.2% as a clean measure of factory economics, because some of it is a federal subsidy flowing through cost of revenue. A markets desk should model the credit-adjusted margin separately from the headline, precisely because the subsidy is the part most exposed to events outside Tesla's control.

The other caveat is competitive: cost-down and mix have to keep outrunning price erosion in a procurement market that bids aggressively. The FY2024 filing celebrates a favorable period; it does not guarantee the relationship holds every quarter. A learning curve that flattens while bidding stays sharp would compress the very margin this disclosure highlights — which is why the durable question is the race between the cost curve and the price curve, not either alone.

It helps to see the segment's full multi-year shape. In 2022 the storage business was barely above breakeven at 7.4% gross margin on $288 million of segment gross profit; by 2024 it was the highest-margin major segment Tesla reports, ahead of the 18.4% total automotive gross margin that year. That inversion — storage out-earning the car business on margin percentage — is the single most important fact a does-it-pencil reader takes from this filing, because it confirms that at sufficient scale a containerized battery product can be a genuine profit engine rather than a strategic loss leader.

The deployment math underneath the margin is worth restating, because scale is what makes the cost-down credible. The FY2024 10-K ties the 52% increase in segment cost of revenue to a 16.7 GWh year-over-year increase in Megapack and Powerwall deployments — a volume step large enough that even a modest per-unit cost decline produces a meaningful absolute gross-profit gain. The filing notes the cost increase was "partially offset by increases in IRA manufacturing credits recognized as compared to the prior year," which is the mechanism by which the subsidy shows up as lower net cost of revenue rather than as a separate revenue line. That accounting detail is why the credit is easy to overlook: it does not appear as income, it appears as a quieter cost.

For grounding, it helps to convert the percentages into the dollars Tesla actually reported. Energy generation and storage segment cost of revenue was $7.45 billion in 2024 against $4.89 billion in 2023, and segment gross profit was $2.64 billion versus $1.14 billion — so the business more than doubled its gross profit dollars in a single year while expanding its margin rate. A segment that grows gross-profit dollars and gross-margin percentage simultaneously is doing something a pure price-cut story cannot explain; it is the fingerprint of cost-down and favorable mix working together, exactly as the disclosure claims.

We count, we don't advise. The takeaway: Tesla's own FY2024 10-K locates storage profitability in product cost reductions, IRA manufacturing credits, and a richer Megapack mix — three named drivers, one of them a subsidy worth tracking on its own. Watch the cost-and-mix story against average selling price, and watch the credit-adjusted margin against the headline. Language from Tesla's 10-K at sec.gov, indexed by SEC filings.