$FABC Everyone is chasing the next NVIDIA. I’m more interested in the next bottleneck. AI isn’t slowing down because GPUs are getting weaker. It’s slowing down because thousands of GPUs need to constantly communicate with each other, and moving that much data efficiently is becoming one of the biggest challenges in computing. Think of an AI data center like a city. NVIDIA builds the skyscrapers. Broadcom and Marvell build the highways. Fabric AI $FABC is trying to replace those highways with a network that moves information using MicroLED optical technology—faster, cooler, and with far less power. If that sounds familiar, it’s because every major computing revolution eventually hits an infrastructure limit. The internet needed fiber optics. Cloud computing needed hyperscale data centers. AI may need optical interconnects. That’s why people are paying attention to FABC. The company is partnering with Kopin ($KOPN), has announced NDAs with two major chipmakers, and is developing a Neural I/O platform targeting data transfer speeds around 3.6 TB/s. If the technology works and reaches production, FABC could become an important supplier to the AI infrastructure ecosystem. But here’s the part many people ignore. This is still a pre-revenue micro-cap company. There are no guaranteed customers, no proven large-scale deployment, and dilution remains a real risk. A great technology doesn’t automatically become a great investment. The opportunity is enormous, but so is the uncertainty. That’s why I don’t see FABC as “the next NVIDIA.” I see it as a venture-capital style bet on one specific idea: That the next trillion-dollar AI race won’t be won by the company building the smartest chips—but by the company that allows every chip to communicate at the speed of light. Sometimes the biggest winners aren’t the ones creating intelligence. They’re the ones connecting it.

🇫🇷 Fiers d’être Français ! 🇨🇵 À Grenoble, l’entreprise française Aledia maîtrise une technologie que ni Apple ni Samsung n’ont encore réussi à industrialiser à grande échelle : les microLED. Des écrans plus lumineux, plus économes en énergie, plus durables et potentiellement bien moins chers grâce à une approche révolutionnaire : des nanofils en 3D cultivés directement sur du silicium. Cette technologie pourrait peser entre 25 et 60 milliards de dollars d’ici 2030 et devenir le nouveau standard des écrans (téléviseurs, montres, etc.). Mais Aledia ne s’arrête pas là : avec sa technologie µLink, elle se positionne aussi sur les interconnexions photoniques pour les centres de données et l’intelligence artificielle, un marché encore plus stratégique. C’est une magnifique démonstration du savoir-faire français dans les semi-conducteurs et les hautes technologies. C’est ça, la France qui gagne et qui peut devenir un champion mondial dans les technologies de demain ! 💪🔵⚪🔴 #Aledia #MicroLED #InnovationFrançaise #SemiConducteurs #Grenoble #FiertéFrançaise #HighTech

Bugatti just dropped a $500k MicroLED television. You won’t believe how it’s inspired by the Bugatti Tourbillon.

La #France n’a pas de pétrole, mais elle a des sacrés ingénieurs … 👉 #Aledia, c’est l’une d’elles. La start-up grenobloise fabrique des #MicroLED ✨, des écrans de demain, dans une usine de 5 000 mètres carrés à #Champagnier, au pied du #Vercors.

Cette entreprise française #Aledia 🇫🇷 maîtrise une technologie que ni #Apple 🇺🇸 ni #Samsung 🇰🇷 n’ont encore réussi à dompter et qui pourrait peser 50 milliards d’euros en 2030 : les #MicroLED ✨ media24.fr/2026/06/14/a-cham…

📺🇫🇷 𝗖𝗲𝘁𝘁𝗲 𝗲𝗻𝘁𝗿𝗲𝗽𝗿𝗶𝘀𝗲 𝗳𝗿𝗮𝗻𝗰̧𝗮𝗶𝘀𝗲 𝗺𝗮𝗶̂𝘁𝗿𝗶𝘀𝗲 𝘂𝗻𝗲 𝘁𝗲𝗰𝗵𝗻𝗼𝗹𝗼𝗴𝗶𝗲 𝗾𝘂𝗲 𝗻𝗶 𝗔𝗽𝗽𝗹𝗲 𝗻𝗶 𝗦𝗮𝗺𝘀𝘂𝗻𝗴 𝗻’𝗼𝗻𝘁 𝗲𝗻𝗰𝗼𝗿𝗲 𝗿𝗲́𝘂𝘀𝘀𝗶 𝗮̀ 𝗱𝗼𝗺𝗽𝘁𝗲𝗿 𝗲𝘁 𝗾𝘂𝗶 𝗽𝗼𝘂𝗿𝗿𝗮𝗶𝘁 𝗽𝗲𝘀𝗲𝗿 𝟱𝟬 𝗺𝗶𝗹𝗹𝗶𝗮𝗿𝗱𝘀 𝗱’𝗲𝘂𝗿𝗼𝘀 𝗲𝗻 𝟮𝟬𝟯𝟬 : 𝗹𝗲𝘀 𝗺𝗶𝗰𝗿𝗼𝗟𝗘𝗗 À Grenoble, une entreprise française tente de réussir là où les plus grands géants mondiaux de l’électronique peinent encore à passer à l’échelle industrielle. Cette entreprise s’appelle Aledia. Depuis plus d’une décennie, elle développe une technologie considérée comme l’un des futurs standards de l’affichage : les microLED. Des écrans capables d’être plus lumineux, plus économes en énergie et beaucoup plus durables que les OLED actuels. La promesse est immense. Chaque pixel devient une véritable micro-source lumineuse inorganique, offrant une qualité d’image exceptionnelle tout en réduisant fortement la consommation électrique. Apple, Samsung, Sony ou LG travaillent sur le sujet depuis des années, mais la production de masse reste extrêmement complexe. L’originalité d’Aledia réside dans son approche industrielle. Là où la concurrence fabrique ses microLED sur des substrats plats en saphir, les ingénieurs grenoblois font pousser des nanofils en trois dimensions directement sur du silicium, le matériau utilisé dans l’industrie mondiale des semi-conducteurs. Cette méthode pourrait réduire les coûts de fabrication d’un facteur cinq à dix. L’entreprise ne vise pas seulement les téléviseurs ou les montres connectées. Avec sa technologie µLink, elle se positionne également sur les futures interconnexions photoniques destinées aux centres de données et à l’intelligence artificielle, un marché qui pourrait devenir encore plus lucratif que celui des écrans. Dans notre article, on vous explique ce que sont réellement les microLED et pourquoi elles pourraient remplacer les OLED, comment Aledia a développé une approche unique basée sur des nanofils en 3D, pourquoi les géants de l’électronique peinent encore à industrialiser cette technologie, quel rôle joue Grenoble dans cette aventure industrielle française, et comment un marché estimé entre 25 et 60 milliards de dollars pourrait faire émerger un nouveau champion européen des semi-conducteurs. Notre article complet : 👉 media24.fr/2026/06/14/a-cham… #France #MicroLED #Aledia #Technologie #SemiConducteurs #Innovation #IndustrieFrancaise #HighTech

And one to watch, not bet on yet: microLED interconnects— the rumored next generation after this one. Too early to pick a winner, but the industry already planning CPO's replacement tells you how durable this theme is.

$KOPN update I did last month that shifted the math: Defense: Platform-agnostic MicroLED displays → potential to supply all 3 competing primes in the $1B US Army SBMC program (direct DoD funding). AI: Clear 2026 demo 2027 production timeline for Neural I/O interconnects with $FABC ( $KOPN owns 19.9%). $100B TAM potential. Worth the read and listen. 👇

Sooo adding onto @aleabitoreddit masterful deep dives for robotics, physical AI. I like U.S. companies i can invest in. Timken's Cone Drive, Timken Company, $TKR offers U.S.-made harmonic (strain wave) gearing as a strong domestic alternative to Japan's Harmonic Drive Systems (HDSI, 6324.T / group including Harmonic Drive LLC in the US). Both provide zero/low-backlash, high-reduction-ratio precision reducers ideal for humanoid robot joints, robotics, automation, aerospace, and medical applications. Core Technology/Performance: Both use the same fundamental strain wave gear principle (flexspline, wave generator, circular spline) for compact size, high torque density, and precision. Similarities: Zero or ultra-low backlash (<1 arc-min or low mRad typical) for excellent positional accuracy and repeatability. High single-stage ratios (typically 50:1 to 160:1, with some higher options). Compact, lightweight designs with hollow shaft options in certain configurations. Configurations: Cup, Hat, and Ring/Pancake styles for flexibility in integration (e.g., robotics joints). Torque density suited for actuators; both support customization. Key Differences: Torque Capacity & Sizes: Japanese Harmonic Drive (HDSI/HD LLC) often leads in peak torque for larger frames (up to thousands of Nm in big units) and broader catalog depth, including miniature and ultra-high-torque options. Cone Drive specs show continuous torque up to ~172 Nm and peak ~750 Nm in common sizes (11–32), with strong performance in cup/hat series. Stiffness & Durability: HDSI is renowned for high torsional stiffness, long-life materials (e.g., specialized flexspline treatments), and proven track record in high-cycle robotics (decades of dominance in industrial robots). Cone Drive emphasizes robustness, U.S. manufacturing quality, and testing; users note it as "decent" to competitive, with good integration options (e.g., with motors or cycloidal complements via Timken's Spinea). Efficiency & Heat: Both have good efficiency (~70-85% typical for harmonics), but specifics vary by design/load. Harmonics generally trade some efficiency for compactness vs. cycloidal alternatives. Cone Drive positions its harmonics for lighter-duty humanoid joints (hips, knees, wrists) alongside cycloidal for higher-load axes. Market Position, Availability & Cost HDSI / Harmonic Drive: The pioneer and market leader (especially in Japan/Asia robotics OEMs like Fanuc, Yaskawa). Premium pricing, global installed base, and tight integration with robot makers. Strong in high-end, high-volume production. U.S. subsidiary (Harmonic Drive LLC) provides local support. Cone Drive (Timken): Smaller but growing player Tier 3-ish in some rankings, focused on U.S./domestic supply chain advantages. Often more competitive on lead times (2-4 weeks standard), customization, and "Made in America" for reshoring/defense/humanoid projects. Price: Japanese originals typically command a premium; Cone Drive may offer better value for U.S. customers needing shorter supply chains. Timken (TKR): Direct U.S.-listed exposure. Industrial Motion segment (including Cone Drive, Spinea, Rollon, recent CGI) drives growth in automation/robotics/humanoids. Strong balance sheet, dividends, and diversification beyond harmonics (bearings, etc.). Robotics exposure has shown double-digit CAGR internally since Cone Drive acquisition (2018). HDSI (6324.T / OTC: HSYDF): Pure-play harmonic leader with robotics tailwinds but more Japan/Asia-centric and subject to currency/competition from Chinese players (Leaderdrive, etc.). Bottom Line: Choose HDSI for proven highest-end performance and robot OEM incumbency. Opt for Cone Drive/Timken for U.S. sourcing, solid precision, faster/custom support, and broader Timken ecosystem—especially appealing for domestic humanoid scaling or defense. #physicalAI #robotics #AI #memory #microLED

A microLED tv running tvOS.

Wall Street is mispricing the $FABC pivot because they are looking at the wrong math. The market sees a $19M micro-cap with a history of bizarre corporate shifts, completely ignoring the $30M pro forma cash pile and an exclusive global monopoly on MicroLED optical interconnects. If their 3.6 TB/s prototype succeeds in late 2026, it fundamentally breaks the thermal and speed bottlenecks of legacy silicon photonics in AI clusters. Here is the hidden truth behind the massive 400% dilution overhang and why this is the highest-beta binary bet in semiconductors today. ⚡🔬 #Semiconductors #AI auth.flash.stocksentinel.ai/…

MicroLED TVs are destined to fail - here's why bit.ly/4ekrsUb

Why do the newest AR glasses run on 30-year-old LCoS technology instead of MicroLED? 🕶️ It comes down to six layers of hidden bulk. A MicroLED chip itself is microscopic, but the engines currently require heavy heat sinks, memory boards, and complex glass prisms that make them too large for normal frames. Discover the exact engineering math behind the AR display war and how new factories from Foxconn and JBD are building the single-chip future for 2028. 🏭💡 $AAPL $META open.substack.com/pub/thinki…

One company. Multiple growth engines. Defense. $THEON Drones. $UMAC Autonomous Systems. $ONDS AI Infrastructure. $FABC MicroLED. @MiTknlg and Team @kopin are seeing beyond and connecting it all. The market is beginning to connect the dots. $KOPN

Do a microLED SCREEN instead.

Absolutely once I got OLED, LCD tech joined the likes of CRT and plasma obsolete, only microled and nanoled will be considered whenever they launch in the next 5 years or so.

🇯🇵2026年6月第2週（2026/6/7～14）の週間XRニュースヘッドライン。 #XRニュース #XRnews orecen.com/x-reality/xr-news… 【今週の注目トピック】 ✅韓国ETRI、XRデバイス向けマイクロLED製造技術を開発 ✅NTTドコモ、体験型MR技術の実証実験を開始 ✅RobloxでサッカーW杯連動イベント開始 🇺🇸Weekly XR News Headlines: June 6 - 14, 2026 【Weekly Highlights】 ✅ETRI unveils microLED manufacturing technology for XR devices ✅NTT DOCOMO begins testing immersive MR technology ✅Roblox launches FIFA World Cup-themed event

$ORCL - Scaling model-agnostic AI data center capacity $KOPN $FABC - Revolutionary MicroLED Optical Interconnect tech for AI Data Centers

This. $FABC working on the leading MicroLED interconnect while the AI companies about to spend a trillion dollars on AI infrastructure. $CRDO and $ALAB have $40B market caps with worse technology. Excited to be early.