Watchlist

~15–19% Anthropic stake (10×/year revenue growth) plus AWS as Anthropic's primary compute partner. Anthropic's $9–10B 2025 run rate is not priced into consensus Amazon models.

Major Anthropic investor and Gemini frontier-lab operator. Dario cites cloud economics (3–4 differentiated players) as the profit model; Google is one of the named players. TPU supply chain adds compute leverage.

The central pillar of AI infrastructure. CUDA ecosystem moat, annual architecture cadence (Blackwell → Vera Rubin → Feynman), and ~$100B+ in supply chain commitments. Industry compute growing 3× per year to an estimated 300 GW by 2029.

Sole leading-edge logic manufacturer; 30-year Nvidia partnership. Every GW of AI compute requires TSMC silicon. Jensen is scaling CoWoS packaging at the same rate as logic nodes — TSMC benefits at every layer.

Sole EUV lithography tool supplier. 3.5 tools = 1 GW of AI compute; 700 total tools by 2030 = 200 GW ceiling. Chronically underprices product vs. value enabled. Named by Dylan Patel as the ultimate AI compute choke-point.

World's #1 HBM supplier; Nvidia's preferred memory partner. HBM uses 4× wafer area per bit vs. DRAM. Memory is 30% of Big Tech CapEx in 2026. Smartphone volume collapse to 500-600M units frees no HBM capacity. Structural scarcity through 2028.

Optimus humanoid robot (target 1M units/yr) × AI5/AI6 chips × domestic lithium/nickel refineries = three compounding recursive loops. America's only lithium and cathode refinery operator. FSD fleet of 10M cars is an unchallengeable physical AI training moat.

$62B record backlog; management guides power generation equipment sales to 3× by 2030 on AI data center demand. White House invoked Section 303 Defense Production Act for grid infrastructure. Trades at 36× forward PE — correctly pricing the AI buildout cycle through 2030.

Visser argues Bitcoin is the last surviving moat AI cannot destroy — physically constrained (energy + chips), socially chosen, and scarce. 100% of Bitcoin historical returns have occurred when CPI exceeds 3-month T-bill yields with the Fed on hold; that regime is approaching. "AI is to every software-based store of value as quantum is to Bitcoin."

Nuclear power for AI data centers. Jensen: energy is the one bottleneck that cannot be resolved in 2–3 years. Dario's 300 GW industry build by 2029 requires power generation that only nuclear can supply at the needed density.

Silicon photonics for AI cluster networking. Jensen invested directly in the Coherent supply chain and explicitly named it as having 'reshaped' AI cluster networking. Scales with NeoCloud and hyperscaler buildouts.

NeoCloud leader; Nvidia backstopped ~$6.3B and made a direct $2B investment. Named as OpenAI's third infrastructure track alongside Azure and OCI. Converts Nvidia CapEx supply into OpEx rental demand.

Power management and electrical distribution infrastructure for data centers. Structural beneficiary of the multi-trillion annual CapEx build — every gigawatt of AI compute needs power delivery hardware.

Silicon photonics supply chain partner; Jensen made direct investments in Lumentum alongside Coherent. AI cluster networking at scale requires optical interconnects — demand tracks directly with GPU cluster deployments.

HBM memory is now mainstream, not a specialty product. Jensen called out Micron specifically for being an early HBM believer. HBM demand tracks directly with Nvidia GPU shipments.

Azure committed to hundreds of billions in Hopper/Blackwell deployments — 'several years of work' per Jensen. OpenAI equity stake and Azure as primary hyperscaler for OpenAI Stargate legs one and two.

OCI contracted 5–7 GW for OpenAI Stargate buildout. Jensen separately flagged an imminent GPU-native data processing initiative — Oracle's SQL/database business is the largest CPU workload in enterprise, making it a direct migration target.

Data center power and thermal management. Dario's 3× per year compute build puts industry at ~100 GW by 2028 and ~300 GW by 2029 — at $10–15B/GW, cooling and power delivery infrastructure is structurally undersupplied.

CVD and etch equipment; second-tier fab tooling bottleneck downstream from ASML. Every new TSMC N3/N2 wafer start requires their deposition and etch tools. Expanding capacity as TSMC ramps AI chip production.

Etch chamber systems; direct fab production bottleneck alongside Applied Materials. Both logic and memory fabs require Lam equipment, giving dual exposure to the AI chip and HBM scarcity thesis.

Gas turbine manufacturer for data center behind-the-meter power. Sold out through 2030. Elon Musk: only 3 casting companies make turbine vanes/blades worldwide — massively backlogged. Oligopoly pricing power confirmed by order backlog.

Largest US lithium producer; upstream feedstock supplier to Tesla's Corpus Christi lithium refinery. Tesla is building America's only domestic lithium and cathode refinery — Albemarle supplies the ore. US government critical minerals reshoring benefits both.

Only US rare earth mine and processing facility (Mountain Pass, CA). China controls 98% of gallium refining and 95% of rare earth processing. Elon explicitly flagged US ore going to China for processing. Bipartisan reshoring policy tailwind.

Dominant copper producer. Optimus at 1M units/yr × ~10 kg copper/robot = 10,000+ metric tons/yr in new demand from robotics alone — before EV and data center wiring growth. Copper is the under-owned AI infrastructure material.

Optical fiber infrastructure; Meta committed $6B in a long-term take-or-pay fiber supply contract. Jensen Huang directly invested in the Corning supply chain. Early-cycle AI buildout lock-in play with multi-year contracted demand.

Specialty chemicals and materials for advanced semiconductor packaging (CoWoS, HBM stacking). Named by Visser (22V Research) as the anchor of the 'chemical layer' of AI — a mid-cycle theme absent from every AI portfolio and analyst model.

Power semiconductor play; lead times for AI server power management chips are 'extremely stretched.' TXN broke out to a 5-year high as part of a 6-name power semiconductor basket. Fits both inside the AI edge and the data center power layer.

Visser's "most under-owned scarcity play" — silver is required in every semiconductor, solar panel, drone, and humanoid robot. Supply is finite; the metal base-built from $60, is consolidating around $78, with a prior high of ~$120 as the next target. A secular bull driven by the physical AI buildout.

Layer 1 of Jensen's five-layer cake; Huang names Siemens as doing well as AI factories drive once-in-a-generation grid reinvestment.

Layer 1 energy beneficiary; Huang names Mitsubishi as doing fantastically in the AI-driven power generation supercycle.

Ives' most mispriced mega-cap alongside MSFT; ~20% of world accesses AI via iOS; WWDC Gemini launch is the consumer AI inflection.

NVLink Fusion partnership with Nvidia opens mobile and edge architectures to the Nvidia AI stack. Jensen framed ARM as a distribution channel into compute markets Nvidia cannot reach alone.

Jensen's contrarian agentic-tool thesis: AI agents will run thousands of Cadence instances in parallel for chip design, removing today's human headcount ceiling. Named explicitly alongside Synopsys as a seat-expansion play.

NVLink Fusion reframes Intel as a Nvidia distribution channel into the enterprise x86 installed base. Jensen called Intel a 'large customer' and said it's a 'great win for both.' Not in any analyst's current AI model.

AI-native drug discovery platform. Dario identified pharma as AI's highest per-token value application — molecule redesign tokens worth 'tens of millions of dollars.' FDA pipeline reform advocacy adds a regulatory tailwind.

Computational chemistry + AI for drug discovery. Dario flagged the FDA pipeline bottleneck as the key constraint; companies combining AI discovery with regulatory-path expertise capture disproportionate value.

Jensen's highest-conviction contrarian call: 'The number of instances of Synopsys Design Compiler is going to skyrocket.' Agents as tool-users will multiply today's human-capped seat count. Market is selling SNPS on AI commoditization fears.

Jensen named Snowflake directly in the CPU→GPU data processing migration. Almost entirely CPU-based today. If Nvidia's 'very big initiative in accelerated data processing' lands, Snowflake becomes a massive GPU customer.

Aerospace-grade turbine blade and vane casting. One of only 3 companies worldwide that produce turbine vanes/blades. Massively backlogged per Elon Musk. Data center power buildout requires turbines; turbines require castings from a 3-player oligopoly.

Fuel cells for behind-the-meter data center power. Named specifically by Dylan Patel (SemiAnalysis) as 'very positive for 1.5 years' with fastest production ramp and quick payback. No dependency on the turbine blade backlog that constrains GE Vernova.

Metrology and inspection equipment; every fab needs KLA tools after each deposition and etch step. Indirect but essential fab tooling play — scales with TSMC capacity expansion for AI chips and HBM memory.

Leading domestic US solar manufacturer. Tesla and SpaceX both have 100 GW/year solar cell mandates; US tariffs block Chinese cells at several hundred percent. First Solar is the primary existing US production beneficiary of any domestic solar ramp.

Thacker Pass (Nevada) — largest undeveloped domestic US lithium resource. Long-dated option on US lithium mining independence as Tesla's refinery needs grow with Optimus and EV production. Government critical minerals designation adds political support.

World's largest nickel producer; feeds Tesla's Austin cathode refinery. Nickel is the key cathode precursor for NMC batteries (Optimus + EV). Tesla as a committed long-term nickel customer provides demand floor.

Bitcoin proxy and inflation hedge. Visser purchased calls for year-end; trigger is CPI exceeding 3-month yields (Fed constrained, rates can't rise, inflation running hot). Up 33% from $60K lows; positioned for breakout if CPI/Fed dynamic plays out.

Flash storage directly in Visser's personal scarcity portfolio alongside Micron and Marvell — oriented to the agentic AI buildout wave requiring ultra-low latency storage for inference workloads.

Only true internet giant Baker calls AI-first internally. Muse near Pareto frontier; Zuckerberg paid up for talent. Rates of change matter over 3-year horizons.

Miscategorized in copper-loser baskets; biggest product is a switch connecting accelerators via copper and optics. Baker flags as mispriced factor-trade victim.

Orbital racks-in-space thesis; operates 98% of satellites in orbit. Starlink V3 scaling 20 kW toward 100–120 kW Blackwell-rack power levels.

GLP-1 cash flows finance LillyPod 1,000-GPU campus and the human-software applications layer; Visser names Lilly as the capital sink when infra capex digests.

Prior Visser third-wave winner alongside Micron; agentic networking and custom ASIC beneficiary as the AI trade broadens beyond GPUs.

Grid storage partner in Nvidia-Siemens solid-state battery work; helps solve US peak-capacity constraint blocking data-center power buildout.

10 GW on OpenAI paper build plan; custom TPU/ASIC supplier competing with Nvidia — Baker skeptical Nvidia lands at 30% deployed share.

Cybersecurity contrarian vs Mythos/Anthropic fear; agentic AI expands attack surface; stock recovered sharply post-RSA per Ives customer checks.

Same Ives cybersecurity contrarian as CRWD; RSA narrative that Anthropic eats security is wrong per enterprise customer conversations.

Memory/components puzzle beneficiary alongside Micron; Meta fills Nvidia chip shortfall from Sandisk and other suppliers in Ives' supply-chain map.

SaaS apocalypse selloff disconnected from outcome per Ives; Benioff wake-up call but models commoditize → people/product differentiation wins.

SaaS fear trade overdone per Ives; enterprise stack replacement harder than narrative; monetization use cases dominate over cost-cutting.

Named in Ives' US AI software leadership stack alongside hyperscalers; benefits from US chip/software/model advantage vs China.

Second neo-cloud position alongside CoreWeave per Fiorillo; paired bet on compute-constrained AI buildout when hyperscalers cannot meet demand internally.

Edge AI devices thesis — down 20% YTD, Visser positions this as a 6–9 month ahead trade before the market prices on-device AI inference (handsets, autos, humanoids). Physical AI needs to run at the edge; Qualcomm is the dominant chip supplier there.

Atreides venture holding; wafer-scale compute that is different and hard. Prefill/decode disaggregation extends legacy GPU fleet life when paired with Hopper/Ampere.

6 GW on OpenAI paper allocation with warrants; Baker calls 30% Nvidia share of 27 GW plan extremely unlikely in watt-constrained world.

Space-sector derivative off AI spatial revolution; Ives names Planet Labs alongside Rocket Lab and Voyager as new sector starters beyond SpaceX IPO.

Space-sector derivative play on orbital compute buildout; Ives cites Rocket Lab as among names defining the new spatial-AI sector alongside SpaceX.