Harley-Davidson:Outside
In
Sixty-six years of buying what it could build
- What they saidMay 2026, the new plan: “leveraging existing powertrain, existing platforms” and “executing better with the platforms we already have rather than introducing entirely new ones.” The first new bike is the Sprint, on a 440cc single Hero MotoCorp builds in Rajasthan.
- What it isThe sixty-sixth year of the same move. Every brand Harley ever bought was sold or shut, with the same sentence. MV Agusta: $105.1M in, $268.4M lost, sold back for €1. Alta: a stake too small to disclose, gone in six months. StaCyc: $14.9M, and it outsells LiveWire 33 to 1.
- What decided itHarley builds when the product carries the full brand at the full price. Everything below the big twin (a small bike, a balance bike, a battery, a loan book) gets bought, licensed or partnered. It stays when it feeds the dealer. It goes when it competes with the badge.
Harley doesn’t want electric. It wants floor traffic.
Faster Faster, Inc. Forms D (2016–2018) · LVWR 8-K (30 Sep 2022; 22 May 2026) and 10-K (FY2022) · LiveWire FY2025 and Q2 2026 results · Hero MotoCorp X440 release (3 Jul 2023) · H-D call transcripts Q1 2019, Q3 2021, Q2 2025, Q1 2026, Q2 2026
- Part IThe recordSixteen moves in one table. Brands go. Channels and components stay.
- Part IISame sentenceBuell and MV Agusta. $125M and $268M. “Focus on the Harley-Davidson brand.”
- Part IIIToo small to sayAlta, never disclosed. StaCyc, $14.9M, and it outsells LiveWire 33 to 1.
- Part IVWhat it buildsTwo engines kept. One small bike lost money and left. The rule.
- Part VThe Sprint rhyme1960: an Aermacchi. 2026: a Hero. Same hole, somebody else’s engine.
- Part VIThe final wordSeven CEOs, one cycle. LiveWire is on the Buell script. What it was for.
The finding, and the trail that led to it
Harley-Davidson has spent sixty-six years buying, licensing and partnering its way into every segment below the big twin, and building from within only when the product carries the full brand at the full price. Here is how three old filings led me to that sentence.
Start with what they said. May 5, 2026, the Q1 call, the new CEO explaining the plan: “By using and leveraging existing powertrain, existing platforms, we can have a much broader assortment of motorcycles to present.” The first new motorcycle under that plan is the Sprint. Its engine is a 440cc single that Hero MotoCorp co-developed with Harley and builds in Neemrana for the X440, a ₹2.29 lakh motorcycle. On the same call: “we’re finalizing the specific production plans.” A company that says it will lean on the platforms it already has is opening the plan with one it does not own.
That sent me back through the record. Not the press record; the filed one. What did Harley pay each time it reached outside the building, what did it get, and how did it end?
- Open the FY2009 and FY2010 10-Ks and MV Agusta is not the “~$163M” everyone quotes. It is $268.4M of net loss from discontinued operations in twenty-four months, on $105.1M of consideration, $20.1M of which was written off on the day it closed.
- Open the 8-K of August 9, 2010 and the sale agreement is attached. The shares went for €1. The U.S. company for $1. A €103.8M receivable for another euro. Harley put €20.0M into escrow on the way out.
- Search every 10-Q and 10-K Harley filed in 2018 and 2019 for “Alta.” One hit, one sentence, in a press release. No amount. Alta’s own Form D chain bounds it at $5.65M.
- Open the Q1 2019 10-Q and StaCyc cost $14.9M. LiveWire’s carve-out statements show the earn-out was paid in full. In 2025 it sold 21,633 units to LiveWire’s 653.
- Line them up by what each one was for and the pattern writes itself: the finance company and the balance bikes were bought to feed the dealer, and they stayed. The brands were bought to be a second Harley, and they went.
None of this is secret. All of it is in the exhibits, and most of it has not been read in a decade.
The trail, in five documents
The order I read them in. Each one sent me to the next.
- “Existing platforms.” Starrs on the plan, and on the Sprint: “we’re finalizing the specific production plans.” The engine is Hero’s. Why does a company with the Milwaukee-Eight not build a 440?
- $115.4M. The impairment, fourteen months after closing. Consideration €68.3M, goodwill $85.8M, IPR&D of $20.1M written off at once. The Buell decision, taken the same week, is in the 10-Q beside it.
- €1. §2.1.1(a): the MV Agusta shares “for a consideration of Euro 1 (one).” §7.1.1: a capital increase of €20,000,000. §7.2.2(a): the earn-out “finally and irrevocably waived.” The 10-K says “nominal consideration.” This is what nominal means.
- $20,843,998. Alta’s round, closed by amendment four weeks before Harley’s announcement. Six investors and $5.65M added since July. Harley’s own filings never give a number; this is the ceiling.
- $14.9 million. StaCyc: $7.0M of cash, $9.5M of goodwill, and a stated reason: “building the next generation of riders.” Then LiveWire’s FY2025 release: 21,633 of them, against 653 motorcycles.
Four instruments
The paperwork that decided each ending.
- Why. Back to the Bricks is scored on HDMC alone and opens with a licensee’s motorcycle. Whether the Sprint and the 883 get built in Harley’s own plants, at a price where Harley makes money, is the whole test of the plan. The record says what usually happens.
- Who for. Anyone deciding whether the Sprint is a product or a channel program; anyone holding HOG into a 2027 that contains the 883, a Sprint, and a LiveWire decision; and the people at Juneau Avenue who already know this and cannot say it.
- What I did. Read the acquisition, impairment and discontinued-operations notes for every outside move since 1993, the sale agreement for MV, Alta’s Form D chain, and LiveWire’s carve-out statements for StaCyc. Derived figures are arithmetic on filed numbers, shown in the captions. Secondary sources are used only where flagged.
- Who I am. I ran product at Harley-Davidson: Touring, CVO, Trike. I have sat in the meeting where a small-bike program gets its cost target. §13.
The record
Every time Harley-Davidson reached outside the building, 1960 to 2026, from the filings.
Sixteen moves, one table
Every figure is from a filing or a company release unless marked ~, which means a secondary source and an open item. “n/d” means the company never disclosed it. Read the last column first.
| Year | Move | In / out | Cost, as filed | How it ended |
|---|---|---|---|---|
| 1960 | 50% of Aermacchi’s motorcycle division, Varese | Out | ~$250K | The original Harley-Davidson Sprint, 250/350 singles. Full ownership early 1970s. Sold to the Castiglionis (Cagiva) 1978 |
| 1986 | Holiday Rambler, recreational vehicles | Out | ~$155M | Segment sold 1996, ~$105M, gain $22.6M, “in order to concentrate on its core motorcycle business” |
| 1993 | 49% of Eaglemark Financial Services; the rest Nov 1995 | Out | $10M + ~$45M | Became HDFS. $248M of operating income in 2024. 9.8% sold to KKR/PIMCO at ~1.75× book, 2025 |
| 1993 | 49% of Buell; substantially all of the rest Feb 1998 | Out | ~$500K; 1998 n/d | Shut Oct 2009. Announced exit cost ~$125M. “Focus both our effort and our investment on the Harley-Davidson brand” |
| 2008 | MV Agusta + Cagiva | Out | €68.3M ($105.1M) | Sold back to Castiglioni 6 Aug 2010 for €1 + $1 + €1, with €20.0M of H-D cash contributed. Net loss from discontinued ops $268.4M in 24 months |
| 2013 | Street 500/750, Revolution X, Bawal (India) | In | n/d | “First all-new platform in 13 years.” India manufacturing exited Sept 2020, ~$75M restructuring. “Unprofitable,” per the CFO, Q3 2021 |
| 2016 | Milwaukee-Eight | In | n/d | Still the Big Twin |
| 2017 | Rayong, Thailand plant | In, offshore | n/d | Rev Max models for the U.S. moved there 2024; returning to York and Menomonee Falls before 2027 |
| 2018 | Equity in Alta Motors (Faster Faster, Inc.) + co-development | Out | never stated | H-D out by August; Alta closed 17 Oct 2018; BRP bought the IP. H-D’s own Silicon Valley EV R&D facility announced 5 Sep 2018 |
| 2019 | StaCyc, Inc. | Out | $14.9M | LiveWire’s only profitable unit. 21,633 units in 2025 |
| 2019 | Qianjiang (Geely), 338cc for China | Out, licence | n/d | X350/X500 in Asia. The X350RA is the U.S. Riding Academy bike, not for sale |
| 2020 | Hero MotoCorp, India distribution and brand licence | Out, licence | follows ~$75M exit | X440 launched Jul 2023, Hero-built, ₹2.29 lakh. The Sprint’s engine |
| 2021 | Revolution Max / Pan America | In | n/d | Pan America, Sportster S, Nightster. Thailand 2024–26; coming home |
| 2021 | LiveWire SPAC; KYMCO $100M | Out, capital | H-D $100M; $1.77B EV | 923 units TTM; $422M of consolidated losses (No. 02) |
| 2025 | HDFS: 4.9% each to KKR and PIMCO; >$5B of receivables; two-thirds forward flow | Out, capital | ~$1.25B in | HDFS operating income guided from $490M to $55–70M (No. 02 §2) |
| 2026 | Dust Motorcycles, Inc., via LiveWire | Out | $375K cash + stock | Pre-revenue; product “2H 2026”; earn-outs to $11.25M |
| 2026 | S4 Honcho, produced by KYMCO | Out, contract mfg | $4,999 / $5,499 | KYMCO exclusive for five years on the maxi-scooter and “any future products on which the parties may agree” |
| 2026–27 | Sprint (Hero 440 single); Sportster 883 | In? | Sprint production “being finalized”; sub-$6,000 (Jul 2025) now “less than $10,000” (trade). The 883: first air-cooled twin program since the Evolution |
Two patterns fall out before any analysis. Every outside purchase that carried its own brand was sold or shut: Holiday Rambler, Buell, MV Agusta, Aermacchi. Three of the four went within two years of a CEO change, with the same sentence. Every outside move that carried no badge into a Harley segment has been kept, quietly: Eaglemark’s balance sheet, StaCyc’s balance bikes, Hero’s engine, Qianjiang’s twin, KYMCO’s assembly line, KKR’s capital. They feed the dealer without competing with the Harley-Davidson brand.
Same sentence
Buell and MV Agusta: what the two biggest purchases cost, and the words used both times.
10 more sections behind one form.
Name and email, once, and this browser is through for a year: the rest of the brief, every figure linked to its filing, the source index, and the whole thing as a PDF. The open items and the corrections log stay public; that is the method.
02 · Buell · 03 · MV Agusta · 04 · Alta Motors · 05 · StaCyc · 06 · Built in-house · 07 · Aermacchi to Hero · 08 · The 883 · 09 · The cycle · 10 · Analysis, not reporting · 11 · Sources
Who wrote this, and how
Left: the Low Rider on the freeway. Middle: leading a group through the canyons. Right: the day job on weekends.
I hold no position in Harley-Davidson, LiveWire Group, Hero MotoCorp, KYMCO, or any company named here: long, short, or derivative. I have no engagement, adverse or friendly, with any of them. I know Marc Fenigstein, Alta’s co-founder, and have worked with him since Alta closed; his account quoted in §04 is public record and is identified as his.
Where a conclusion rests on my time in the industry rather than a document, I say so. Where an earlier version of this was wrong, the fix is in the log at §13, not made quietly. If something here is wrong, tell me.
Primary documents first. Every dollar figure on MV Agusta, Buell, StaCyc, Eaglemark, India, Dust and LiveWire is from a filing, linked in §11. Alta’s bound is from Alta’s own Form D chain, and is marked as an inference.
Derived figures are labelled. Loss-to-consideration multiples, loss per month, units per unit, revenue per unit and the Form D increment are arithmetic on filed numbers; the script is in the repo at tools/gt03/calc.py.
Fact and opinion are separated. §§01–09 are sourced. §10 is judgment and says so. Where a reading rests on my time in the industry (the York cost base, the planning-office calculation) it is tagged Industry context.
No inside sources. Nothing here rests on anything I was told. The public record carries every claim.
Major findings carry one of five tags, so a reader can see at a glance what kind of claim is being made and how to attack it.
What is not yet nailed down, and what Rev. 0 got wrong
Ten open items. Anything in this list is tagged Unverified where it appears above. The corrections log is public policy: nothing is fixed silently.
| # | Item | Status |
|---|---|---|
| 1 | Alta investment amount | Not in any H-D filing (full-text search, 2018–2019). The bound in §04 is inferred from Faster Faster, Inc.’s Form D chain. The FY2018 10-K MD&A paragraph on investment income was not retrieved. It is the one place an unnamed write-off might be described. |
| 2 | “Alta Motors” in LiveWire’s 10-Ks | LiveWire’s FY2024 and FY2025 10-Ks contain the string (EDGAR full-text hit); location not retrieved, probably an officer biography. |
| 3 | Aermacchi | 1960 price (~$250K), 1978 sale price, Sprint volumes and the year of full ownership (1972–74) are secondary only. |
| 4 | Holiday Rambler purchase price | $155M is trade press (Family RVing, 2003); the 1986 annual report is not on EDGAR. |
| 5 | Buell 1993 price and cumulative units | ~$500K and 135,000+ / 136,923 are secondary. |
| 6 | Hero X440 volumes; Sprint build source | 8,974 (Apr–Dec 2024) is Autocar Professional’s. Neither Harley nor Hero has confirmed Hero builds the Sprint. Check the dealer-meeting materials and the Q3 call. |
| 7 | Sprint price; 883 price and plant | “Less than $10,000” and “~$10,000 / York” are press paraphrases of the 5 May 2026 investor deck. Fetch the deck. |
| 8 | Milwaukee-Eight and Rev Max program costs | Not in either release; check capex commentary 2014–2020. |
| 9 | LiveWire TTM units | 923 to June 2026 is from No. 01; the Q1 2026 release was not re-pulled. |
| 10 | Realized India exit charge | $75M is the estimate for actions approved September 2020; FY2020 restructuring was $130.0M for all 2020 actions and does not break India out. |
Corrections log
Rev. 0 (8 September 2026) was an outline built from secondary sources. Rev. 1 (9 September 2026) corrects it on reading the filings.
| What was claimed | What the document said |
|---|---|
| MV Agusta: “~$163M of losses in 24 months.” | That is the 8-K’s “prior write-downs of $162.6 million, net of tax,” which is impairments only. The filed net loss from discontinued operations is $268.4M (2008–10), $217.4M after the Q4 2011 tax reversal. |
| MV sale: “€1 with €20M of H-D cash inside.” | Right but incomplete: three nominal payments (€1 shares, $1 U.S. LLC, €1 for a €103.8M receivable), €20.0M into escrow, earn-out waived. |
| StaCyc 2025 units: 21,141. | 21,633 (LiveWire FY2025 release). Ratio to motorcycles 32 → 33. |
| HDFS: “9.8% to KKR/PIMCO, ~$6B book sold.” | 4.9% each to KKR and PIMCO (9.8% collectively, FY2025 10-K) and “over $5 billion” of receivables (30 Jul 2025 release). |
| Thailand plant, 2018. | Announced 25 May 2017; producing 2018. |
| “Zeitz, 2025: unprofitable” on Street. | The on-the-record statement is CFO Gina Goetter’s, Q3 2021 call. Zeitz said “prune unprofitable motorcycles” without naming Street. |
| Buell exit cost $125M, stated as fact. | It is the October 2009 estimate (Q3 2009 10-Q, Note 19). No filing reports a realized Buell-only total. |
| Holiday Rambler sold 1996 for ~$50M. | The whole Transportation Vehicles segment went for ~$105M, gain $22.6M (FY1996 10-K); ~$50M was the RV division per wire reports. |
| Alta: H-D “would instead build its own R&D center in Silicon Valley.” | Paraphrase. The facility was announced in a separate release, 5 Sep 2018. |
| Dust: cost undisclosed. | Terms are in LiveWire’s 22 May 2026 8-K. |
| S4 Honcho $4,999. | $4,999 Trail; $5,499 Street. |
| Aermacchi full ownership 1974. | Sources split 1972/1973/1974; stated as “early 1970s.” |