History
The Story
The current chapter of Nextronics Engineering (Nextron) is the story of an old, twice-burned connector OEM that re-founded itself as a high-margin niche supplier — first to medical, then aerospace, and finally to the AI compute supply chain. The pivot was named in FY2020 ("MATIC"), funded with two convertible bonds and a Taiwan factory purchase in 2020–2021, and validated by what management told shareholders would happen: communication/cloud connectors riding a generative-AI wave. Five years on, AI accounts for roughly 30% of revenue, gross margin has climbed from 32.7% (FY2019) to ~40% (FY2025), and the May 2026 board decision to abandon its newly-authorised convertible bond suggests cash flow has outrun the original capital plan. Credibility has improved — concrete operational promises (Taiwan plant, ESG certifications, 400G shipping, AI thesis) have largely landed; only product-rollout timing has slipped.
1. The Narrative Arc
The strategic chapter that matters today began in 2009, not 2020. Management has been candid that the 2008–2009 financial crisis — when gold-price spikes turned six-month-old Huawei/ZTE backplane orders into losses and "nobody dared raise prices" — was the trauma that drove the firm out of red-ocean OEM and toward specialty connectors. What the annual reports document is the visible phase of that pivot: the MATIC framework was formalized in FY2020, Taiwan capacity was bought in late 2020, and the AI optionality (seeded by a 2017–2018 Mellanox win, before Nvidia acquired Mellanox in 2020) compounded into 2025's record print.
FY2025 is annualized from 9M actuals (+30% YoY revenue, NI +73%) and from the March 2026 Commonwealth Magazine report citing December 2025 revenue +80% YoY and 40% gross margin. FY2024 figures are external estimates; the company has not posted its 113-nian annual report as of May 2026.
The non-obvious inflection is the gross-margin line, not the revenue line. Revenue dipped 10.8% in FY2023 from cyclical destocking — but gross margin kept climbing, from 37.3% (FY22) to 38.9% (FY23) to ~40% (FY25). That gap is the proof that the MATIC pivot is structural, not a cycle.
Anchors for downstream tabs. Current chapter began in FY2020 (MATIC formalized in the shareholder letter). CEO Yen-Chen "Kevin" Chen (陳言成) was already in operational leadership during the 2008–2009 crisis — i.e., this is not an inherited high-quality business. Chairman Hsu Chi-Lin (徐季麟) is the founder (1986).
2. What Management Emphasized — and Then Stopped Emphasizing
The shareholder-letter heatmap below scores each theme on a 0–3 emphasis scale across four annual reports (FY2020–FY2023, the four documented years).
Topic emphasis (0=absent, 3=heavy) — FY2020 through FY2023 shareholder letters.
| Topic | FY2020 | FY2021 | FY2022 | FY2023 |
|---|---|---|---|---|
| MATIC framework | 3 | 3 | 3 | 3 |
| Cloud / Communication | 3 | 3 | 2 | 3 |
| Medical | 2 | 2 | 3 | 2 |
| AI / GenAI / high-speed | 1 | 1 | 3 | 3 |
| Transportation / EV | 1 | 1 | 3 | 2 |
| Aerospace | 2 | 1 | 1 | 2 |
| Industrial | 2 | 2 | 1 | 2 |
| Thermal / cooling | 0 | 1 | 2 | 3 |
| ESG / sustainability | 1 | 2 | 3 | 3 |
| Smart factory / automation | 1 | 2 | 2 | 3 |
| Capacity expansion / capex | 3 | 2 | 1 | 2 |
| Low-margin OEM / contract bid | 2 | 1 | 1 | 1 |
What rose: AI/high-speed (1→3), thermal/cooling (0→3), smart factory/automation (1→3), ESG (1→3). The thermal management story is the most telling — absent in FY2020, by FY2023 it has its own business unit and is one of the only product lines explicitly tied to the AI bull case.
What faded: Capacity expansion language (3→2, peaked in FY2020 with the Xizhi purchase and two convertible bonds), low-margin OEM/contract-bid discussion (2→1, because the share is no longer big enough to talk about — 12% in FY2020 → 7% in FY2023 guidance). The conspicuously thin line is aerospace, which never got past 2/3 emphasis and stalled at ~2% of revenue.
What stayed: MATIC. Four years, four shareholder letters, four uses of the same word. That is rare — most pivots get renamed within two years. Treat the consistency as a credibility positive.
3. Risk Evolution
The risk-factor section of the AR is boilerplate; the shareholder letter is where management actually telegraphs what is keeping them up at night. Below is the same 0–3 scale applied to the macro/operational risks management chose to discuss.
Risk emphasis in shareholder letters (0=not mentioned, 3=heavy).
| Risk | FY2020 | FY2021 | FY2022 | FY2023 |
|---|---|---|---|---|
| COVID / pandemic | 3 | 2 | 2 | 1 |
| US–China trade / tariff | 3 | 1 | 2 | 2 |
| Russia–Ukraine | 0 | 0 | 3 | 1 |
| Middle East / Red Sea | 0 | 0 | 0 | 3 |
| China cost / labor / supply chain rivalry | 2 | 2 | 2 | 1 |
| FX / TWD | 2 | 2 | 3 | 2 |
| Inflation / raw materials | 1 | 3 | 2 | 1 |
| Supply chain disruption | 2 | 3 | 2 | 1 |
| Chip shortage (customer-side) | 0 | 2 | 1 | 0 |
| ESG / carbon compliance | 1 | 2 | 3 | 3 |
The risk register churns more than the strategy register. COVID dominated 2020; supply-chain/inflation took over 2021–2022; Russia–Ukraine appeared in 2022 and then quickly receded; Israel–Hamas and Red Sea shipping arrived in 2023. What is missing — quietly — is any line on customer concentration. Top-10 customers were 45.7% (FY21), 41.3% (FY22), 43.5% (FY23), and the AI-driven inflection in FY2025 was attributed by external reporting to a single descendant of one customer (Mellanox→Nvidia). That risk lens is not in the letters.
4. How They Handled Bad News
There are two bad-news episodes in the documented window, and the company handled both honestly.
FY2021 — Gross margin slipped 0.8pt on input-cost inflation. Rather than hide it inside an operating-leverage story, the FY2021 letter named the cause directly:
"Overall gross margin declined approximately 0.8%, principally due to rising raw materials, manufacturing expenses, and freight costs."
What this earns: management chose to call out a sub-1pt margin drag in the same paragraph as a record revenue print. That is the opposite of obfuscation.
FY2023 — Revenue fell 10.8%. The letter explained it as customer inventory destocking (medical -4.3%) with no attempt to dress it as strategy. And then management let the gross-margin print do the rebuttal: 38.9%, the highest in the company's recorded history.
No misses were rewritten in subsequent letters. Across four annual reports, every multi-year promise that recurred (MATIC mix, AI/cloud direction, Taiwan production base, ESG certs) was carried forward with consistent framing. There is no detectable narrative laundering.
5. Guidance Track Record
Taiwan-listed companies are not required to publish numeric guidance and Nextron has not. What they have published, repeatedly, are operational commitments — capacity, certifications, product roadmaps, strategic mix targets — and on those, the track record is good.
Credibility score (1–10)
Why 8/10. What pulls it up: every concrete operational deliverable named in a shareholder letter — Taiwan plant, certifications, solar capacity, product roadmaps, the AI thesis itself — arrived. The May 2026 decision to not issue an authorised convertible bond is the single most credibility-positive event in the file; very few Taiwan small-caps walk away from approved dilution. What holds it short of 10: the "multi-fold" autonomous-vehicle ramp was rhetorically overshot in FY2022, the FY2024 annual report was still missing as of May 2026 (5+ months past the typical filing window), and customer-concentration risk has been quietly absent from the letters even as the AI cluster grew.
6. What the Story Is Now
The current chapter — "high-margin niche specialist in AI compute, medical, transportation, and aerospace, made possible by exiting consumer/telecom red oceans after 2009" — is now five years into formal articulation and seventeen years into actual execution. Four points to leave with:
De-risked. The structural margin (38.9% FY23 → ~40% FY25) is durable across both a +18% growth year and a -10% growth year. The MATIC framework has been carried through four annual letters without renaming. AI exposure is real (~30% of revenue per March 2026 reporting), not aspirational.
Still stretched. FY2025's 30% growth is anchored on a single supply-chain pedigree (the 2018 Mellanox engagement that flowed into Nvidia). The FY2024 annual report has not been filed as of May 2026 — a real governance friction even if the 9M financials look strong. Customer concentration is unaddressed in the letters.
What the reader should believe: the gross-margin trajectory, the operational delivery record, and the AI revenue mix as a structural — not cyclical — change. The pivot from contract OEM to design-led niche supplier appears genuine; the lesson the CEO has cited from 2008–2009 ("nobody dared raise prices") has translated into actual business-model discipline.
What the reader should discount: extrapolating the FY2025 +30% rate forward without acknowledging the lumpiness of AI-cluster purchase cycles, and management's silence on top-10 customer concentration. Also: the May 2026 CB withdrawal is good now but raises the question of why the dilution authority was sought in mid-2025 — either a precautionary capital plan that proved unnecessary, or a misread of FY2025's organic momentum.
Net: a company whose narrative has converged with its results. That is the rarer kind.