Ticker Theses (45)
Post-AI sector crash, explicit Strong Buy call on GOOGL/MU/SNDK as highest-conviction names. AI infrastructure/memory/semis oversold — contrarian bullish call.
NVDA CEO confirms everything across supply chain is in shortage — memory, wafers, advanced packaging, SiPho, cable connectors. Multi-year supply tightness supports NVDA pricing power.
Post-AI sector crash, explicit Strong Buy call on GOOGL/MU/SNDK as highest-conviction names. AI infrastructure/memory/semis oversold — contrarian bullish call.
Macro-driven dip in AI names viewed as opportunity; thesis on AI capex cycle intact; nonfarm payroll shock transient
LPK AGM pending; glass substrates superior to ABF for CPO; LPK is the major equipment play, very early innings
IQE £81M raise complete: 80% strategic (MTSI 152M + noteholders 115M), only 65M shares to market — dilution overhang removed
SaaS is alive. NOW +23% in a week. AI not replacing SaaS names - durable moats across PLTR, MSFT, NOW, SHOP, DDOG, CRM, SNOW, APP, MDB, ZETA. SaaS companies integrating AI become stronger.
QCOM up 6.84% AH after Jensen Huang says 'Buy their stock'. Edge AI thesis playing out.
SIVE $8.2M order from All Space for BFICs. Huge order for 2027 vs Q1 2026 rev of $6.7M. Embeds SIVE in US defense ecosystem via YSS.
AAPL passing GOOGL in market cap signals GOOGL is relatively undervalued; personal buy signal
SOFC companies positioned to solve data center power bottleneck; BE preferred in sector due to scale/positioning
GOOGL raising $80B for AI infra; total 2026 AI capex ~$185B (2x 2025). Broad AI supply chain beneficiaries.
GOOGL raising $80B for AI infra; total 2026 AI capex ~$185B (2x 2025). Broad AI supply chain beneficiaries.
OUST competes across four lidar pools: automotive OEM (fastest growing), Western automotive/robotaxi (main auto opportunity), automation/robotics/mapping (largest revenue), and industrial; lidar adoption curve accelerating
FY26 outlook reaffirmed with revenue+EPS at high end of guidance, memory/AI infrastructure demand growing, DDR5 DRAM tailwind
Bullish from first principles: chips need stress-testing before deployment, AEHR provides burn-in testing for NVDA/AMD/INTC chips
AXTI vital InP substrate bottleneck: AI capex -> GPUs -> optical interconnect -> 800G/1.6T/3.2T transceivers -> EML/DFB/CW lasers -> InP substrates
Complete basket of analog/power semis for GaN/SiC power conversion thesis
CRDO up 25% since earnings. High confidence in re-rating higher alongside CPO TAM expansion. Bringing SiPho in-house via DustPhotonics acquisition enables vertical integration.
AI supply chain beneficiaries cemented in multi-year theses; hyperscaler capex growing massively
Still early innings; demand exceeds capacity; 2027 inflection on 800G/1.6T ramp; capacity expansion underway now
SOFC companies positioned to solve data center power bottleneck; BE preferred in sector due to scale/positioning
MTSI positioned for parabolic CPO growth with SiPh drivers and CW lasers; management confirming CPO optimization
Actively researching/positioned in OUST, PENG, FCEL, ENSI — semiconductors/fuel cells names
Robotics supply chain play; Kawasaki Heavy partnering with NVDA/MSFT/ADI for physical AI, confirmed by Jensen's TAM expansion comments
Samsung's $1.5B Vietnam investment will compete directly with Hana Micron's captive Samsung memory packaging/test operations, long-term negative for Hana
Testing is critical bottleneck in CPO mass production - manual and non-standardized approach creates severe throughput bottleneck. FORM (probe cards), ONTO (inspection/metrology), CAMT exposure.
Testing is critical bottleneck in CPO mass production - manual and non-standardized approach creates severe throughput bottleneck. FORM (probe cards), ONTO (inspection/metrology), CAMT exposure.
Testing is critical bottleneck in CPO mass production - manual and non-standardized approach creates severe throughput bottleneck. FORM (probe cards), ONTO (inspection/metrology), CAMT exposure.
McKinsey humanoid supply chain mapping confirms actuators are key battleground for mass production. BofA: actuators = 51% of total BOM. Humanoid robotics supply chain investable.
McKinsey humanoid supply chain mapping confirms actuators are key battleground for mass production. BofA: actuators = 51% of total BOM. Humanoid robotics supply chain investable.
AI capex expected $7.6T from 2026-2031 per Goldman Sachs. Capital flows upstream into memory (MU, SNDK), photonics (SIVE, LITE), substrates (GLW, AXTI), foundries (INTC, TSEM).
AI capex expected $7.6T from 2026-2031 per Goldman Sachs. Capital flows upstream into memory (MU, SNDK), photonics (SIVE, LITE), substrates (GLW, AXTI), foundries (INTC, TSEM).
AI capex expected $7.6T from 2026-2031 per Goldman Sachs. Capital flows upstream into memory (MU, SNDK), photonics (SIVE, LITE), substrates (GLW, AXTI), foundries (INTC, TSEM).
AI capex expected $7.6T from 2026-2031 per Goldman Sachs. Capital flows upstream into memory (MU, SNDK), photonics (SIVE, LITE), substrates (GLW, AXTI), foundries (INTC, TSEM).
SMCI was too cheap heading into earnings. Baseline fundamentals intact despite negative headlines (China smuggling scandal, ORCL competition). Forecast clean beat and raise. Contrarian long.
Down 12.28% on Blue Origin selloff, no negative Rocket Lab news, waiting for dip after stellar earnings, sees as buying opportunity
Complete basket of analog/power semis for GaN/SiC power conversion thesis
Complete basket of analog/power semis for GaN/SiC power conversion thesis
Complete basket of analog/power semis for GaN/SiC power conversion thesis
GFS adds Sivers InP lasers to SiPh reference designs incl CPO platform, value chain alignment with AVGO/MRVL/NVDA widens funnel
MXL +70% since going long post-earnings. Holding through June for remaining catalysts. Expresses continued bullish conviction.
AI supply chain beneficiaries cemented in multi-year theses; hyperscaler capex growing massively
BRUN owns almost nothing; assets borrowed, GPUs leased, DCs rented colocation; '125 MW capacity' mostly owned by others; fundamentally weak setup
Raspberry Pi H1 EBITDA $38M (+96% YoY), well ahead of $43.7M FY forecast. Holding 2025 DRAM inventory purchased at low cost — direct beneficiary of memory inflation.