IPG Photonics' defense lasers are a stickier, higher-margin business even after a hollow beat sent the stock cratering. Its CROSSBOW lasers, built on IPG's fiber laser technology, neutralize drones up to 25kg, with an 8kW version reaching 2,500 meters in a demo. A $10 million Lockheed Martin follow-on order in February 2026 is small against $1 billion of revenue but shows Lockheed standardizing on IPG's laser for mobile counter-drone use. Defense/medical targets 45%+ gross margin. x.com 

IQE's 43% first-half revenue growth overstates the recovery, because the group is still slightly smaller than two years ago. Photonics is a genuine breakout at £38.5m, 44% above H1 2024, but wireless at £26.0m remains a third below that level and is still recovering. A £4.0m working-capital outflow absorbed the margin gain, 61% of May's £81m raise went to lenders, and depreciation runs about nine times capex. Converted InP lines delivering H2 volume without more capex would prove efficiency. … Show more

IPG Photonics is a trade at best, not an investment, while Lincoln Electric is the preferred investment for cutting and welding. A 23% earnings drop would normally attract interest, but IPG is a retail photonics hype name whose stock is where it was 10 years ago, and the telecom division it sold to Lumentum years ago was small. Buying it would take hard facts of a proven business pivot, like Rambus's, showing the next 10 years won't look like the last 10. x.com  x.com  x.com 

Applied Optoelectronics is the best dip to buy after the photonics selloff took it from the mid-$200s to around $140. The market is misreading Meta's plan to sell compute as bearish for AI names. Memory names like Micron and SanDisk are near highs, where bad sentiment could hit harder, and Nebius takes the most direct hit from fears Meta becomes a competitor or cancels deals. All of these should bounce soon. x.com  x.com 

Photonics stocks look like a crowded rotation rather than a better bet than memory. Photonics has a smaller addressable market than memory and moves with it anyway. Buyers are driven by recency bias after Applied Optoelectronics' 90% recovery, traders liquidated in memory seeking higher-risk bets, and smaller companies with higher multiples. @therogueoracle adds that these stocks had massive runs and have not pulled back nearly enough, so they are more likely to chop sideways and down. x.com  … Show more

AI photonics is the next scarce layer after GPUs and memory, with demand exceeding supply until mid-2027 as copper hits its wall at 1.6T. Indium phosphide wafers rose from about $1,400 to $5,000 after China's export licensing, supply covers under 30% of demand, and new lines are 2-3 years away. Applied Optoelectronics makes its lasers in-house in Texas, Lumentum's five InP fabs trail demand by over 30%, Coherent grows its own InP, and Credo gets paid on both copper and optical links. x.com  x.com  … Show more

Photronics looks overextended, with risks far outweighing rewards after a 170%+ surge to near its all-time high. It trades at 23x earnings on only 4-5% growth, insiders have been aggressively selling, and the chart shows clear bearish divergence. Semiconductor and display markets are highly cyclical, so a slowdown in downstream demand could trigger a sharp pullback. x.com 

Optical component stocks such as Lumentum, Coherent, Applied Optoelectronics and AXT look early in an unwind, as a wave of new photonics paid newsletters signals a crowded narrative. Price pays; narratives do not. x.com 

Credo is one of the cheapest AI growth stocks, trading around 26x forward earnings after a hard correction while revenue grew 115% to $479M and full-year growth is guided above 85%. Margin worries mostly reflect DustPhotonics amortization and stock comp; non-GAAP gross margin held near 68%. DustPhotonics adds a second engine in optics, with over $600M expected this year, and Credo's low-laser designs can ship through the laser shortage. One customer at about a third of revenue is the main risk. … Show more

AXT remains a vital InP substrate bottleneck, and keeping some exposure makes sense because a ton of AI capex flows directly to it, though much is priced in. Transceivers moving from 800G to 1.6T and 3.2T need lasers built on InP substrates, which AXT and Sumitomo supply; demand exceeds supply even as AXT doubles capacity by end-2026. Management sees CPO inflecting late 2027 and beyond. Earnings should improve as backlog converts, though another parabolic move looks unlikely. x.com  x.com