Stijn Elebaut (@StijnEl) header
Stijn Elebaut (@StijnEl) profile picture

Multi asset Portfolio Manager | Crazy about the stock market | Bird lover | Opinions are my own

Joined June 2011
Photo by Stijn Elebaut (@StijnEl)Photo by Stijn Elebaut (@StijnEl)Photo by Stijn Elebaut (@StijnEl)Photo by Stijn Elebaut (@StijnEl)Photo by Stijn Elebaut (@StijnEl)Photo by Stijn Elebaut (@StijnEl)Photo by Stijn Elebaut (@StijnEl)Photo by Stijn Elebaut (@StijnEl)Photo by Stijn Elebaut (@StijnEl)Photo by Stijn Elebaut (@StijnEl)
Stijn Elebaut retweeted
X-FAB has been written about all year. There is nothing new to learn about X-FAB that is not already reflected in the price. Yet the stock moves up as much as 70% today because of this post. So either the market has been completely wrong until now, or the claims being made are wrong. Spoiler, it is the claims that are wrong. X-FAB is being pitched as “photonics + power semis” with four supposed AI hooks. The reality is that these segments together account for only about 23% of revenue, and almost none of that has meaningful AI exposure. Microsystems and photonics are mainly tied to automotive sensors, medical microfluidics, and industrial sensing, while integrated photonics contributes only about $5–6M in NRE and remains in the R&D phase. Wide-bandgap exposure is still overwhelmingly automotive and renewables, with datacenter only described on the Q1 call as a recent add-on market. Real AI datacenter revenue is likely well below 2%. The remaining 77% is classic analog and mixed-signal CMOS, heavily concentrated in automotive at around 62%, and that core business is currently shrinking with automotive down 10% YoY. X-FAB is today a European specialty foundry for third-party designs running at roughly 60% utilization, FY2025 revenue around $770M, with only marginally positive EBIT, $291M net debt, and negative owner FCF. This is a high fixed-cost contract manufacturer sitting at the bottom of a cyclical downturn. The claimed 800V power semi exposure through Navitas and Power Integrations is simply wrong. There is no pathway from NVIDIA’s 800V push into X-FAB. Navitas runs through Powerchip and, from late 2025 onward, GlobalFoundries for U.S. capacity. TSMC exits GaN entirely by mid-2027. Power Integrations manufactures PowiGaN internally. Infineon, Innoscience, MPS, ROHM, STM, and TI are vertically integrated. X-FAB’s GaN remains developmental with no commercial revenue, and its production-capable dMode platform is not aligned with the eMode datacenter mainstream. There is no NVIDIA, Navitas, or Power Integrations design win. The silicon photonics argument is half true but massively overstretched. NVIDIA is indeed part of X-FAB’s photonixFAB consortium, but only as an application development partner evaluating technology. That does not imply high-volume manufacturing. NVIDIA’s actual CPO supply chain runs through TSMC and Tower. X-FAB’s photonics business is still pre-revenue, with product commercialization likely not before 2027–2028, and LIGENTEC controls the direct customer relationship. Consortium membership is not a production contract. The MEMS claim is technically correct but irrelevant for AI. X-FAB is a meaningful pure-play MEMS foundry for automotive sensing, medical microfluidics, and industrial applications. But the MEMS segments that matter for AI datacenters, optical circuit switch micro-mirrors and MEMS timing oscillators, are controlled elsewhere. The total addressable pure-play MEMS foundry market is only around $1B globally. X-FAB is a top-five player in a niche that is simply too small to move the valuation narrative. The “only high-volume SiC foundry in the U.S.” statement is also true but misleading. The global SiC market is dominated by integrated manufacturers like Infineon, STM, Wolfspeed, onsemi, and ROHM, which control over 90% of industry revenue. X-FAB’s share is around 1% globally and still operates on 150mm wafers while the industry transitions to 200mm. The headline 152% YoY growth is coming off a tiny base and is constrained by consignment substrate supply, limiting margin expansion. Government support is real, but it is not an earnings backstop. The funding is tied to growth capex for future capacity, with commercial ramp from Fab4Micro only expected around 2029. The U.S. support remains politically uncertain. These subsidies imply years of additional capital outflow before any return, not reduced near-term risk. second post ➡️
36
23
13
333
168,319
Stijn Elebaut retweeted
Het is vandaag exact 1 maand geleden sinds de laatste koersvorming bij #ABO. 1 maand ! Zonder nieuws, zonder update, zonder iets. Dat kan je toch niet serieus menen @FSMA @euronext ?
6
5
40
5,865
Er is vandaag veel verdriet in Beleggersland want...onze traditionele "aandelengids" is niet meer. Een fantastisch naslagwerk dat menig kleine beleggers en beleggingsclubs bestudeerden. Aandelengids, we gaan je heel hard missen.😭 – at Herzele, België
7
1
15
3,861
Stijn Elebaut retweeted
Belang van sociale media? Even voor de duidelijkheid: zonder de uitgesproken reacties via de sociale media en vooral via deze @X zou deze ´leugen´ nooit een item geworden zijn in onze #msm. Nooit. Integendeel. Ze gingen er zelfs heel graag in mee. #colruyt & @pvdabelgie
16
67
4
367
9,292
Stijn Elebaut retweeted
Volgens @RaoulHedebouw betaalt Colruyt 0,27% vennootschapsbelasting. Hierbij uittreksel uit de geconsolideerde jaarrekening per 31.03.2023. Waarom zo liegen? @De7deDag #de7dedag
77
112
24
506
192,982
Stijn Elebaut retweeted
Een president die als strategie heeft om steden plat te bombarderen om zijn doelen te bereiken, moet volledig uitgesloten worden door de westerse wereld. Stop gas- en olieaankopen tot Poetin inbindt. Stop de waanzin in Oekraïne.
280
101
16
1,239
Tel jij ook vogels dit weekend? ⁦@MijnNatuurpunt⁩ #vtmnieuws
5
3
16
Stijn Elebaut retweeted
Topspeech. Over landen die meer uitgeven dan ze binnenkrijgen (via belastingen), en die dat dan nog eens compleet inefficiënt doen. 1ste stap = sito presto begrotingstekorten wegwerken (door minder uitgaven). Alles onder 0% is immorele belasting op de toekomst. RT voor impact
One of the best - if not the best - speech I’ve ever seen by a politician. Please WATCH. It’s just two minutes.
15
138
10
324
The Dow “Joans” is down 1031 points today @realDonaldTrump & it is hot where you’re going so there’s that.
7
20
3
51
Stijn Elebaut retweeted
Bloomberg geeft cijfers #melexis vervroegd (?) vrij. Omzet zou in 2019 dalen volgens bericht bloomberg. Maar nog geen officieel persbericht van melexis zelf.
1
3
Stijn Elebaut retweeted
Dat klopt, een beleggingsclub bekijkt elk dossier afzonderlijk en luistert naar zijn clubgenoten. VFB steunt de regering met een gratis lidmaatschap. Peeters: 'Regering is geen beleggingsclub' tijd.be/r/t/1/id/10044922 via @tijd
1
4
20
VIX ontwaakt ...
2
2
4,613
159,301
12,334
274,439
Stijn Elebaut retweeted
VFB stelt voor speculatietaks te beperken tot bedrijven met marktkapitalisaties van minstens 2,5 miljard euro
4
1
Het voorstel van JVO op 25 juli. Welke drukkingsgroep heeft hier de rode pen gehanteerd? #dtv #speculatietaks
6
59
30
Stijn Elebaut retweeted
Het standpunt van De Belegger over de #speculatietaks. Waarom deze maatregel oerdom is. http://t.co/O61E5hNfht
5
1
Het is ver gekomen dat we de SP.A moeten gelijk geven... teapaw.org/cderijcke/status/65422…
SP.A die het moet opnemen voor de kleine belegger.
4
1
Stijn Elebaut retweeted
[VIDEO] - Hevige brand in houtbedrijf in Oudegem ift.tt/1NpE9Lw
Stijn Elebaut retweeted
"12 Europese overnameprooien", volgens SocGen een overzicht: bolero.be/nl/analyse-en-inzi…
1
1
Stijn Elebaut retweeted
Staar u niet blind op de lage rente http://t.co/8TQH8BtlxK
1
2