In Nacho We Trust

GAN is the way forward … lets discuss some alpha in the Photonics Supercycle: Signals in AIXA earnings call $AIXA had earnings call and popped up another +11%, bringing YTD upside to +130% and 1 yr upside to +288%. You guys know that am already invested in Aixtron ($AIXA) and have written about this German champion in the past, framing it as a quintessential "hidden monopoly" within the semiconductor equipment space. I also think that this is early innings of a massive Photonics Supercycle. As AI clusters migrate from 800G to 1.6T and eventually 3.2T optical modules, the bottleneck for AI scaling has shifted from the H100/B200 compute engine to the interconnect fabric itself. You cannot run a frontier model if the data can’t move between the GPU and the memory stack fast enough. Aixtron sits at the very mouth of this funnel. I would not regurgitate the numbers here, rather analyse and unpack the things I consider important with impact on overall AI and semisector. The Great Divergence: GaN vs. SiC One of the most tactical takeaways from the call was the stark contrast between Gallium Nitride (GaN) and Silicon Carbide (SiC). 🐊 SiC Overcapacity: Management was candid about Silicon Carbide being in a state of "substantial market overcapacity," with SiC expected to be only 10% of their revenue this year. This is a warning shot for pure-play SiC substrate and device makers who are still digesting inventory. 🐊 The GaN Opportunity: Conversely, GaN is entering a golden era. Management expects AI data centers to become the largest single application for GaN power semiconductors. Because GaN is significantly more efficient than traditional silicon at high frequencies, it is becoming the standard for the power supply units (PSUs) that feed power-hungry AI servers. For those looking to play the GaN trend beyond the equipment layer, keep an eye on these players: • Navitas Semiconductor ($NVTS) – A US-based pure-play leader in GaN power ICs. • Infineon Technologies ($IFX.DE / $IFNNY) – A European giant with a massive, growing footprint in GaN following their acquisition of GaN Systems. • STMicroelectronics ($STM) – Another European heavyweight with strong GaN-on-Silicon capabilities. • BluGlass ($BLG.AX) – An Australian-listed speculative play involved in specialized 🐊 Datacom Doubling: Management is forecasting that demand for datacom lasers will more than double YoY in 2026. This is a massive green flag for the laser makers themselves, such as Lumentum ($LITE) and Coherent ($COHR), who are racing to expand Indium Phosphide (InP) capacity. (am invested in both) 🐊The "Tool of Record" for AI Interconnects The standout signal from the call was the performance of the G10-AsP system. Management confirmed it has become the "tool of record" for the next generation of photonic components 🐊 Order Intake Explosion: Order intake hit €171 million, a ~30% YoY increase that blew past analyst expectations. Crucially, 65% of these orders came from the optoelectronics segment 🐊Implications for the Memory (HBM) Ecosystem While we often view memory through the lens of SK Hynix or Micron $MU, the Aixtron update provides a critical "read-through" for how these chips will be connected in the future. 🐊 Structural Bottlenecks: For those holding SK Hynix, $DRAM or Micron ($MU), the shift toward optical interconnects is the solution to the "power wall" that threatens to limit HBM performance. Aixtron’s order backlog suggests the equipment for this transition is being installed now. 🐊 Near-Stack Integration: The acceleration of Co-Packaged Optics (CPO) means that optical engines will eventually sit directly on the package alongside High-Bandwidth Memory (HBM) stacks. Look the "Photonics Supercycle" isn't a theory anymore it’s appearing in the capex cycles of the world's largest chipmakers. And remember, none of this is financial advice. The crocodile can make errors. So do your own research.