Build-A-Bear's real asset is not the bear. It is a lease book that makes most of the store base cancellable within about three years. The short sellers holding a quarter of the float are betting on falling discretionary revenue. The filings show a company that has spent a decade making its costs fall alongside that revenue. That changes what a consumer slowdown does to the equity, and it is the part of Michael Burry's Part II case that a headline reading of "mall retailer, weak consumer, heavy short interest" leaves out.
The number that best captures the gap appears in the company's own FY2025 10-K: $529.8 million of revenue and $64.6 million of operating income, a 12.2% operating margin. Burry puts 2007 revenue at $474 million. After eighteen years, the GFC, the smartphone and a pandemic that sent the stock to $1.80, the top line is only about 12% above its pre-crisis level in nominal terms. Growth is not the thesis. What the business earns on that revenue is.
Margin Without Growth
Burry describes what new management inherited in 2012. Roughly 20% of locations were unprofitable. The store base averaged less than a 10% contribution margin. 85% of stores sat in traditional malls, and e-commerce was 4% of net retail sales.
That 2012 figure and today's 12.2% measure different things, so the comparison needs care. Contribution margin is a store-level measure taken before corporate overhead. Operating margin is taken after it. The company now earns more at the consolidated operating line, after paying headquarters, marketing and depreciation, than its average store earned before overhead a little more than a decade ago. That improvement did not come from growth. It came from changing what kind of business sits behind roughly the same revenue.
Three changes drove it, and each shows up as a different line in the financials:
- Lease duration. By 2022, Burry notes, more than 70% of leases ran less than three years. Over 80% of visits are planned, so Build-A-Bear is a destination tenant that brings traffic to a mall. That gives it leverage in renewals that most mall apparel tenants never had.
- Partner-funded growth. Partners such as Carnival, Landry's and the Girl Scouts pay for buildout, real estate and labor, and they buy inventory at wholesale. Wholesale locations grew from 60 in early 2022 to 177 at June 30 of this year. Each one adds revenue that needs little or no capital from Build-A-Bear.
- Format and footprint. Mall exposure fell from 85% to 65% by 2022, with more tourist sites and small concourse stores. The refreshed Discovery format now covers 55% of corporate-owned stores.
None of this happened in a single quarter, which is why the market missed it. Burry's description is accurate: the model "changed dramatically, but over an extended period of time." Slow change rarely gets priced as change. It gets priced as the old business with a lucky run.
The Debt Line That Isn't Debt
This is where the filings and the summary narrative seem to disagree. Burry describes Build-A-Bear as having "no financial debt." The FY2025 balance sheet shows total debt of $127.3 million against total assets of $345.5 million. A quick screen reads that as a leveraged small-cap retailer. An investor checking Burry's claim against the headline balance sheet might conclude he was wrong.
The two statements fit together only if most of that $127.3 million is lease obligations rather than borrowings. Under current lease accounting, store leases sit on the balance sheet as liabilities, and many data summaries fold them into "total debt." This article infers that split. The condensed figures here do not break the line out. But the inference agrees with Burry's description, and it is the most important accounting point in the case.
If the inference holds, the debt line works in the opposite way from bank debt in a downturn. A term loan stays fixed when sales fall. A lease book that mostly rolls within three years gets smaller as stores close or renewals are renegotiated. The liability is not a fixed claim on the equity. Management can reprice it on a rolling basis. That is the specific reason an earnings decline at Build-A-Bear may fall less to the bottom line than the short thesis assumes.
Cash Conversion and the Coverage Gap
The cash-flow statement supports the earnings picture. FY2025 operating cash flow was $65.1 million, almost exactly equal to operating income. Earnings are therefore not being propped up by accruals or working-capital timing. Capex was $25.5 million, 4.8% of revenue and 39.3% of operating cash flow, leaving $39.5 million of free cash flow.
Burry estimates EBITDA "in the $80 million range" and a market capitalization under $300 million at a share price near $25. On those figures, the FY2025 free cash flow implies a yield in the low teens. That number depends on his market-cap estimate, which is not reproduced from a filing here, so treat the yield as approximate. The direction is clear, though. A company converting nearly all operating income to operating cash and keeping about 60% of it after capex should not trade like a business in decline.
The weak coverage helps explain why it does. Burry says most of Wall Street does not cover the name. Standard peer screens show the same problem in a different way. Put Build-A-Bear into a Consumer Cyclical / Specialty Retail screen and it gets compared with a homebuilder, a restaurant chain and a funeral operator:
- Cracker Barrel (FY2026): $3.3 billion of revenue at roughly breakeven operating margin, capex 55.9% of operating cash flow.
- Beazer Homes (FY2025): $2.4 billion of revenue at a 1.5% operating margin, capex consuming 89.1% of operating cash.
- Carriage Services (FY2025): a 23.5% operating margin, but capex at 131.2% of operating cash flow and negative free cash flow of $19.0 million.
Build-A-Bear has the second-highest margin in that group and the lowest capex burden relative to operating cash. These are not real comparables, and that is the point. When a stock has no natural peer group, it tends to be valued against the category it is filed under, here discretionary mall retail, rather than against its actual economics. The short interest Burry cites, 25% of the float and almost nine days to cover, is the market expressing that category view with conviction.
The Case the Shorts Are Making
The bear case deserves a fair hearing, and part of it is strong.
First, the stock was above $75 last year and is now around $25. A two-thirds decline does not happen without some damage to the earnings outlook, and the market may be correctly treating FY2025 as a peak. Burry himself says the consumer is weakening and trading down on discretionary purchases. A $30 stuffed animal with an accessory upsell is about as discretionary as retail gets, and he concedes it might seem so. If comparable-store traffic falls sharply, a 12.2% margin can compress quickly. Flexible leases soften that over three years. They do nothing for the next two quarters.
Second, the partner model has a cost the filings summarized here do not resolve. Wholesale revenue from 177 partner locations needs little capital, but it is probably lower-margin per unit than a corporate store selling the full experience at retail. If growth is moving toward that channel, the mix could cap margins even as the store count rises. This analysis does not have segment-level margins, so the size of that effect is unknown.
Third, the capex line is not optional for long. Discovery covers 55% of corporate stores, so 45% still have not been refreshed. Keeping capex at 39% of operating cash depends on that program continuing to pay back. A refresh cycle in a falling-traffic environment is the point where disciplined retailers sometimes over-spend.
Fourth, international expansion through seven master franchise agreements covering twelve countries is, by Burry's own account, still early in planning. It is an option on the future, not a source of current earnings, and the bull case should not depend on it.
What Would Settle It
The short thesis needs revenue to fall faster than costs. The long thesis, which Burry clearly favors, needs the reverse: a cost base that reprices downward quickly enough to protect free cash flow through a discretionary slowdown. Both sides agree revenue is at risk. They disagree about the speed of the cost side, which is a matter of contract terms.
Three observable items will resolve that disagreement more reliably than any commentary on the consumer:
- The lease liability relative to revenue. If the $127.3 million balance falls faster than sales in the next filings, the flexibility is working. If it holds steady while revenue drops, the lease book is less cancellable than the 2022 duration figures suggested.
- Free cash flow conversion. Operating cash flow tracking operating income, as it did at $65.1 million against $64.6 million in FY2025, is the cleanest sign that reported earnings are real. A widening gap would weaken the argument faster than a revenue miss.
- The wholesale location count. The June 30 figure of 177 is the baseline. Continued growth while corporate stores hold or shrink would confirm that expansion is being paid for with other companies' capital.
On the filings available, the market is pricing a revenue problem and ignoring a cost structure that has been rebuilt to handle one. The thesis breaks if the next downturn shows that the short leases were short on paper only. Until a filing shows that, the lease book is the better guide to how this equity behaves than the mall it sits in.
