NeutralAMZNAmazonAmazonCOHRCoherent Corp.Coherent Corp.GOOGLAlphabetAlphabetLITELumentum Holdings Inc.Lumentum Holdings Inc.METAMSFTMicrosoftMicrosoftNVDANVIDIANVIDIATSEMTower SemiconductorTower Semiconductor

Today on

- $GOOGL began developing OCS in the 2010s and deployed it in Jupiter data center networks and TPU infrastructure. - Google reported 30% lower networking CAPEX and 41% lower power consumption from its OCS- and SDN-enabled Jupiter architecture. - OCS reconfigures optical connections without converting optical signals into electrical signals at the switching layer. - $MSFT joined the OCP OCS initiative alongside Google, $NVDA, $LITE, and $COHR in 2025. - $META developed Disaggregated Scheduled Fabric to address the scaling limitations of traditional Clos networks and is exploring alternative switching technologies. > Google's decade-long OCS deployment establishes a proven foundation for broader adoption across AI data centers. > Growing AI clusters are exposing the power, capacity, and cost constraints of electrical packet switching. - $AMZN introduced its ShuffleBox-based flat network architecture in May 2026, but this passive optical interconnect approach is distinct from OCS. - $META continues to rely heavily on Ethernet packet switching in its Prometheus infrastructure, disclosed in October 2026. > A wholesale transition by $MSFT, $AMZN, and $META to all-optical switching is not yet established by public disclosures. > However, rising electrical packet switching costs strengthen the economic case for replacing selected switching layers with OCS. > The OCS market opportunity could expand dramatically if adoption extends beyond Google to multiple hyperscalers. > $LITE and $COHR offer direct OCS hardware exposure, while SiPh-based architectures could create additional manufacturing opportunities for $TSEM. > Production deployments, port shipments, and revenue conversion matter more than the number of companies participating in OCS initiatives.

NeutralCOHRCoherent Corp.Coherent Corp.GOOGLAlphabetAlphabetLITELumentum Holdings Inc.Lumentum Holdings Inc.OCSOculis HoldingOculis HoldingTSEMTower SemiconductorTower Semiconductor

Today on

- $LITE identified OCS and pump lasers as leading growth contributors in fiscal Q1 2027. - $LITE reported significantly improved OCS demand visibility. - Management sees OCS potentially becoming a multi-billion-dollar annual revenue business. > OCS may be transitioning from early deployment into a large commercial market. > $LITE stands out for directly linking OCS to its revenue growth outlook. - Salience Labs partnered with $TSEM to manufacture PIC-based OCS. - $TSEM's PH18DA integrates III-V lasers. - TPS45PH supports low-loss SiN waveguides. - The partnership has advanced into pre-production. > Cascaded insertion loss, crosstalk, and fiber-attach yield could determine SiPh OCS manufacturing competitiveness. > $TSEM offers foundry exposure to PIC-based architectures competing with MEMS OCS. - Voyant Photonics announced plans to adapt its FMCW LiDAR silicon photonics technology for data center OCS. - Its chip-scale OCS targets switching times below 5 microseconds. - Voyant plans to supply OCS engines rather than complete networking systems. - Key development priorities include non-blocking topology, optical loss, crosstalk, and packaging. > LiDAR photonic switching technology is expanding into AI data center infrastructure. > Fast device switching alone does not guarantee better network throughput. Scheduling and reconfiguration overhead remain critical. - OCP published a Fast-Switching OCS white paper in 2026. - Contributors include $AMD, $INTC, $CIEN, $LITE, iPronics, and Lumotive. - An April 2026 OCP paper compared MEMS and SiPh switching technologies. > OCS competition is expanding from relatively static optical paths toward dynamic AI fabrics. > End-to-end reconfiguration time and network software maturity matter more than isolated device switching speed. - $COHR offers digital liquid crystal-based OCS. - Its configurations range from 64x64 to 512x512 ports. - The company has demonstrated a 300x300-port system. > OCS competition extends beyond MEMS versus SiPh, with liquid crystal emerging as another high-radix option. > Port cost, insertion loss, reconfiguration time, and customer adoption will shape competition between $LITE and $COHR. > The central shift is OCS diversification beyond Google's MEMS-centered architecture. > MEMS offers attractive high-port-count and low-loss characteristics, while SiPh targets faster reconfiguration and semiconductor manufacturing scalability. > Investors should monitor $LITE's actual OCS revenue growth, $COHR's customer adoption, and $TSEM's transition into volume manufacturing.

NeutralAMDTSMTSMCTSMC

Today on

- $AMD announces multi-generation EPYC processor roadmap aligned with $TSM's advanced node progression - $TSM's 2nm process maturation provides manufacturing foundation for $AMD's server CPU development > $AMD's disclosure of its extended EPYC roadmap signals confidence in $TSM's ability to deliver advanced nodes on schedule, which is material for both companies' competitive positioning in data center processors. > The alignment between $AMD's product pipeline and $TSM's process maturation suggests stable supply chain execution. https://t.co/7J2mFsMSbY

TQQQBullish

Today on

Anyone remember $SPLG? It's now $SPYM. My YTD return was once approaching +150%. Today, it's fallen well below +100%. I thought I was doing pretty well investing in AI infrastructure stocks. The market has a way of teaching humility. Haha. I trimmed some individual pAnyone remember $SPLG? It's now $SPYM, with a slightly lower expense ratio than $VOO (0.02% vs. 0.03%). My YTD return was once approaching +150%. Today, it's fallen well below +100%. I thought I was doing pretty well investing in AI infrastructure stocks. The market has a way of teaching humility. Haha. I trimmed some individual positions today and bought 200 shares of $TQQQ around $55 in premarket trading. But I've been wondering. Would I be better off quietly accumulating something like $SPYM and doing nothing? Maybe beating the market isn't the hardest part. Staying in the game is.ositions today and bought 200 shares of $TQQQ around $55 in premarket trading. But I've been wondering. Would I be better off quietly accumulating something like $SPYM and doing nothing? Maybe beating the market isn't the hardest part. Staying in the game is.

Neutral000660SK hynixSK hynix005930Samsung ElectronicsSamsung ElectronicsASICAtegrity SpecialtyAtegrity SpecialtyAVGOBroadcomBroadcomGOOGLAlphabetAlphabetNVDANVIDIANVIDIAORCLOracleOracle

Today on

- $GOOGL has expanded its TPU customer base to nearly 100. - $ORCL plans to offer Google TPUs through OCI starting in 2027. - Counterpoint Research forecasts a 35-fold increase in HBM bit demand for AI ASICs from 2024 to 2028. - Average HBM capacity per AI ASIC is projected to increase approximately fivefold over the same period. - Samsung Electronics ($005930) and SK hynix ($000660) reportedly supply HBM for Google TPUs. - $AVGO maintains a long-term agreement to develop and supply Google's next-generation TPUs through 2031. - Reports suggest $GOOGL could expand direct HBM procurement, but the scale of the transition and a 15–20% cost reduction target remain unconfirmed. > TPU commercialization diversifies HBM demand beyond $NVDA, creating new growth opportunities for $005930 and $000660. > Direct HBM procurement could strengthen $GOOGL's purchasing leverage while putting pressure on memory suppliers' ASPs and margins. > The key distinction is between reducing $AVGO's procurement role and replacing its TPU design partnership. > Investors should monitor not only TPU shipments but also HBM capacity per TPU and changes in supplier market share.