LiveWire:5 Years In
and 1%
of Plan

The finding, and the trail that led to it

Five years in, one percent of plan. Here is how a percentage led me to the contracts.

Start with what they said. 13 December 2021, the investor call announcing the merger, filed with the SEC as a Form 425. Harley-Davidson’s CFO, on the record, for 2026: “approximately 100,000 units,” revenue “from $35 million today to almost $1.8 billion,” and “profitable on both an EBITDA and cash flow basis.” That plan carried a $2.31 billion valuation.

It is 2026. Trailing twelve months: 923 motorcycles. $31.3 million of revenue. A parent guiding a $70–80 million loss. Units at 0.9% of plan, revenue at 1.7%, and the top line below where it stood the day the plan was announced. Call it one percent. I have watched a lot of product plans miss. I have never watched one miss by the whole plan.

That is the finding. The rest of this is about why, and whether the structure under the company lets anyone fix it. The trail started somewhere much smaller.

I saw a number online: LiveWire, 386% unit growth in Q2. Accurate. Also 55 motorcycles to 267, and fifty-five motorcycles is not a base, it is a parking lot. Normally that is worth a shrug. What turned it into two weeks of reading was where it led:

• Open the 10-Q for the denominator and you find a write-down eating 87% of the gross loss.
• Ask why anyone writes down brand-new inventory and you land in a manufacturing agreement priced cost-plus.
• Read that agreement and there is a second one, with KYMCO, signed the same day.
• Read the services contract next to them and Harley can terminate on ninety days’ notice, no reason required.
• At the bottom of the pile: a loan with its conversion feature removed and a lien over basically everything put in its place. $76.8M, due December 2027.

None of that is in the earnings release. None of it has been reported. It lives in the exhibits, the attachments nobody opens, including most of the people paid to cover the stock.

The trail, in five documents

The order I read them in. Each one sent me to the next.

1. Q2 2026 earnings release · 8-K, 23 Jul 2026 →

   386%. 55 motorcycles to 267. I went looking for the retail number the company itself calls its share measure and found it grew 171%, not 386%. Same table, different line.
2. Form 10-Q · 5 Aug 2026 · segment note →

   The write-down. About 87% of the quarter’s gross loss was a net realisable value adjustment on brand-new S2s. You don’t write down new inventory unless it is carried above what it will fetch. The filing doesn’t say why. So: who set the price?
3. Contract Manufacturing Agreement · H-D · 26 Sep 2022 →

   Cost-plus. Harley builds, LiveWire buys into its own inventory at Harley’s cost plus a mark-up. Miss the volume and there is a provision for the parts Harley already bought: $6.1M of it, once, in 2023. And there is a second agreement, signed the same day, with KYMCO.
4. Master Services Agreement · H-D · 23 Dec 2024 · §8.3(c), §8.4 →

   Ninety days. Everyone quotes the “forty percent” clause. The paragraph under it lets either party terminate for convenience on ninety days’ notice. No trigger required.
5. Amended & Restated Term Loan · H-D · Nov 2025 · §1(c)(iii), §1(d) →

   The seat change. Conversion feature gone. Security interest over substantially all assets in. First $10.0M of any ATM proceeds goes to Harley. $76.8M due 15 December 2027, compounding to about $85M. This was signed before the quarter above was even reported.

The paperwork, on one screen

Four instruments. Read the cards and you can skip the rest; every section below is the arithmetic they imply.

• Why LiveWire. It is the only pure-play electric motorcycle company in the U.S. that has to file with the SEC. Everyone else hands you the numbers they feel like handing you. Here the press release can be checked against the arithmetic.
• Who this is for. Suppliers deciding whether to tool up. Investors sizing a position. Dealers weighing a franchise. Riders wondering if the warranty outlives the company.
• What I did. Read all 260 filings for both companies, the agreements and the exhibits nobody quotes. Every figure links to its source. Corrections are published, not made quietly.
• Who I am. Ex-Harley product manager (Touring, CVO, Trike), ex-Honda planning lead on the Grom, now an independent expert witness. No position in anything named here. Full bio and method in §14.

386% more motorcycles, and the loss did not move

Both numbers come from the same table, same segment, same two periods. No outside denominator, nothing to argue about. It is the cleanest comparison in the filing, which is probably why nobody used it.

The other number: 76% market share

The same post carried a second number: 76% U.S. market share. Of what? Three annual reports later, nobody has written it down.

The company’s own filing names the number that measures share, and it isn’t the one in the post. Q2 2026 10-Q:

Retail is the share measure, by LiveWire’s own definition. The 10-Q’s table:

The 386% is 94% wholesale: bikes shipped to dealers, an invoice, not a customer. The number the company itself calls the measure of demand grew 171%. Anyone who has worked a month-end close at an OEM knows the difference between those two numbers, and knows which one goes in the post.

To be fair on the point that matters: retail beat wholesale in the quarter (293 vs 267) and the half (412 vs 358). Dealer stock went down. This was not channel stuffing; the demand is real at its scale. But 386% and 171% are different claims about different things, and only one is about riders.

One day, two filings

The 386% quarter went out on 23 July 2026, 8-K, Item 2.02. That was the good news. The bad news arrived the same afternoon, by letter.

Same day, the New York Stock Exchange told LiveWire it was out of compliance with Section 802.01C: thirty straight trading days under $1.00. That letter was disclosed six days later, in a separate 8-K under Item 3.01.

Who wrote this, and how

Corrections log

Each of these was asserted in an earlier revision and corrected on reading the document. The log is published because a method is only worth what its error-handling is worth.