For information only. Not investment advice, a recommendation, an offer or a solicitation, and without regard to any recipient’s objectives or circumstances. Figures come from third-party sources believed reliable but not guaranteed, and market data may be delayed. Past performance is not a reliable indicator of future results. |
AI InfrastructureOverview|> Scope note (2026-08-10): This page absorbed the former "Artificial Intelligence" theme, which tracked the same hyperscaler capex cycle and largely the same beneficiaries from the model/monetization side. Rather than maintain two near-duplicate pages, AI Infrastructure now covers the full stack — capex/hardware buildout through model deployment and monetization. DefinitionAI Infrastructure refers to the physical and software layers required to train and deploy large-scale artificial intelligence models — from compute (GPUs, ASICs), networking (InfiniBand, Ethernet), and storage (HBM, NAND) through the power systems (grids, data centers) and specialized electrical EPC engineering that support them, up to the models and applications running on top of that stack. Investment thesisWe are in the midst of an unprecedented capital expenditure expansion cycle. The global AI infrastructure buildout is backed by a massive forward contractual ledger. Post-Q2 2026 reporting, combined cloud/AI infrastructure backlogs across the hyperscalers and neocloud players have crossed $2.5T (tripling from ~$750B in Q3 2025). Rather than purely aspirational guidance, this represents take-or-pay long-term agreements, prepayments, and multi-year enterprise commitments that provide extraordinary visibility across the hardware and physical supply chains. Key drivers:
Key beneficiaries
Key risks
Recent developments
Related themesExposed companies (68) TD SynnexSNX US Micron Technology, Inc.MU US Amphenol CorporationAPH US SalesforceCRM US Apple Inc.AAPL US Vinci SADG FP Dell Technologies Inc.DELL US ASML Holding N.V.ASML NA American Electric Power Company Inc.AEP US Microsoft Corp.MSFT US Palantir Technologies IncPLTR US Analog Devices Inc.ADI US NVIDIA CorporationNVDA US Howmet Aerospace Inc.HWM US Amazon.com Inc.AMZN US Sandisk CorporationSNDK US Linde plcLIN US Trane Technologies plcTT US Western Digital CorporationWDC US Advanced Micro Devices Inc.AMD US Constellation Energy CorporationCEG US Broadcom Inc.AVGO US Vistra Corp.VST US Alphabet Inc.GOOGL US Arista Networks Inc.ANET US Cisco Systems Inc.CSCO US Flowserve CorporationFLS US Seagate Technology Holdings plcSTX US Interactive Brokers Group Inc.IBKR US Meta Platforms Inc.META US VAT Group AGVACN SW Quanta Services, Inc.PWR US Sanmina CorporationSANM US Talen Energy CorporationTLN US Lam ResearchLRCX US Semtech CorporationSMTC US GE Vernova Inc.GEV US Applied MaterialsAMAT US Vertiv Holdings CoVRT US KLA CorporationKLAC US Morgan StanleyMS US Schneider Electric SESU FP Corning IncorporatedGLW US JPMorgan Chase & Co.JPM US International Business Machines CorporationIBM US Bank of America CorporationBAC US Eaton Corporation plcETN US Marvell Technology Inc.MRVL US NextEra Energy, Inc.NEE US Oracle Corp.ORCL US ServiceNow Inc.NOW US Coherent Corp.COHR US Lumentum Holdings Inc.LITE US Qualcomm Inc.QCOM US Tesla Inc.TSLA US Palo Alto Networks Inc.PANW US CrowdStrike Holdings Inc.CRWD US Comet Holding AGCOTN SW Intel CorporationINTC US Siemens AGSIE GY SpaceXSPCX US BYD Company Limited1211 HK BYD Electronic (International) Company Limited0285 HK Kioxia Holdings Corporation285A JP SK Hynix Inc.000660 KS Samsung Electronics Co. Ltd.005930 KS Taiwan Semiconductor Manufacturing Co.TSM US |
AI Infrastructure
Overview
|> Scope note (2026-08-10): This page absorbed the former "Artificial Intelligence" theme, which tracked the same hyperscaler capex cycle and largely the same beneficiaries from the model/monetization side. Rather than maintain two near-duplicate pages, AI Infrastructure now covers the full stack — capex/hardware buildout through model deployment and monetization.
Definition
AI Infrastructure refers to the physical and software layers required to train and deploy large-scale artificial intelligence models — from compute (GPUs, ASICs), networking (InfiniBand, Ethernet), and storage (HBM, NAND) through the power systems (grids, data centers) and specialized electrical EPC engineering that support them, up to the models and applications running on top of that stack.
Investment thesis
We are in the midst of an unprecedented capital expenditure expansion cycle. The global AI infrastructure buildout is backed by a massive forward contractual ledger. Post-Q2 2026 reporting, combined cloud/AI infrastructure backlogs across the hyperscalers and neocloud players have crossed $2.5T (tripling from ~$750B in Q3 2025). Rather than purely aspirational guidance, this represents take-or-pay long-term agreements, prepayments, and multi-year enterprise commitments that provide extraordinary visibility across the hardware and physical supply chains.
Key drivers:
- Contractual Purchase Commitments ($2T+): Big tech non-cancelable contractual obligations have surged (e.g. Alphabet's obligations up ~9x Y/Y to $811B), locking in multi-year GPU/accelerator, networking, power, and memory capacity.
- Hyperscaler Capex Step-Up: Annual data center capex spending is projected by CFRA to reach $1.3T in CY 2026 and $2.0T by CY 2028. NVIDIA's purchase order backlog across Blackwell and Vera Rubin platforms has crossed $1.0T through CY 2027.
- Next-Gen Accelerator Inflection (Vera Rubin & Custom XPUs): Inflection confirmed for 2H 2026 / Oct-Q as NVIDIA ramps the Vera Rubin platform (carrying a ~$3.5M–$4.0M price tag per NVL72 rack, +25% vs Blackwell; CFRA puts Vera Rubin content at ~$42B per gigawatt vs ~$25B for Grace Blackwell) and Broadcom scales custom hyperscaler XPU revenues to $56B in FY26 and >$100B in FY27. NVIDIA's FQ3 guide of ~$108B ±2% and preliminary ~70% FY2028 growth indication put a number on the ramp.
- Standalone AI CPU & Agentic Orchestration: Entry of purpose-built agentic AI CPUs (NVIDIA Vera targeting $20B in FY27 revenue within a $200B TAM) to handle token orchestration and tool use, shifting compute economics toward token-throughput efficiency.
- Ecosystem Financing Platforms: Introduction of non-balance-sheet AI infrastructure financing vehicles (NVIDIA's $500B MOU with 6 major financial institutions; Broadcom/Apollo/Blackstone AI XPU platform) de-risking balance sheets and accelerating adoption among neoclouds, enterprise, and sovereign AI projects.
- High-Bandwidth Memory (HBM) Deficit — now the sector's binding constraint: Extreme supply tightness in HBM3e/HBM4 and rising component costs are compressing hyperscaler lease margins, providing immense pricing power to Samsung, SK Hynix, and Micron Technology through 2027. NVIDIA's FQ2 FY2027 call made this explicit: memory supply, not customer demand, is what caps its FY2028 growth at ~70% rather than ~100%. The cycle's ceiling is now set upstream of the accelerator vendors.
- Physical, Electrical & Optical Bottlenecks: High-speed clusters require advanced liquid cooling architectures, complex high-voltage data center substation engineering (Vinci SA), and high-density Ethernet networking silicon (Broadcom Tomahawk 6 @ 100 Tbps).
Key beneficiaries
- Networking: Arista Networks Inc. (Ethernet back-end; AI networking ~$3.5B 2026E -> ~$7B 2027E; AI back-end switching passed front-end in Q2 2026 per Dell'Oro), Cisco Systems Inc..
- Cloud Platforms: Microsoft Corp., Amazon.com Inc., Alphabet Inc., Oracle Corp., Meta Platforms Inc., CoreWeave Inc. (neocloud leasing).
- Memory: Samsung Electronics Co. Ltd. (HBM4/HBM4E leadership, DS margin ~70%), SK Hynix Inc., Micron Technology, Inc..
- Power, Cooling & Data Center EPC: Eaton Corporation plc, Schneider Electric SE, GE Vernova Inc., NextEra Energy, Inc., Vertiv Holdings Co, Trane Technologies plc (hyperscale central chillers and direct liquid cooling co-developed with NVIDIA), Howmet Aerospace Inc. (airfoils and hot-section castings for AI on-site gas turbine peakers), Vinci SA (high-density electrical & cooling engineering).
- Models & Applications: Microsoft Corp. (Copilot, Azure AI, Agent 365), Salesforce (Agentforce), ServiceNow Inc., Alphabet Inc., Meta Platforms Inc..
- AI Servers & Enterprise AI Systems: Dell Technologies Inc. (leading AI server OEM, first to ship NVIDIA Vera Rubin racks; AI backlog and orders running ahead of in-quarter sales, demand spilling into traditional servers and storage).
- Interconnect & Power Semiconductors: Amphenol Corporation (high-speed copper, power and fiber interconnect, widened by the CommScope CCS acquisition), Analog Devices Inc. (system-level power delivery, optical circuit switching control and the Empower Semiconductor acquisition).
- Sovereign / Enterprise AI Software: Palantir Technologies Inc (AIP deployment platform; US commercial revenue now above US government).
- AI Systems Manufacturing & Connectivity: Sanmina Corporation (US rack- and cluster-scale AI system integration for a leading accelerator vendor), Semtech Corporation (FiberEdge/CopperEdge connectivity silicon and photonics for 1.6T/3.2T optical modules).
Key risks
- Power Constraints: Grid capacity and data center power availability are becoming the primary bottlenecks for new deployments; since the 11/04/2026 US midterms, a possible state-level data-center moratorium (Texas) adds a permitting and policy risk.
- Capex Fatigue / ROI Scrutiny: Risk that hyperscalers may slow spending if AI monetization does not meet expectations — CFRA notes a ~40% probability of an AI capex bubble pop within three years if monetization lags supply catch-up.
- Technological & Execution Risk: Complex rack-scale manufacturing transitions (e.g. Rubin Ultra 78-layer PCB midplane / Kyber NVL144 execution risk) and co-packaged optics scaling.
- Model Gap / Regulation: Meta AI acknowledged behind Gemini/GPT, though reaccelerating; AI safety and DMA/antitrust scrutiny remain overhangs on the applications layer.
Recent developments
- 2026-11-16: CFRA 2027 Outlook: data center capex +90%+ to $1.2T in 2026 and +50% to $1.9T in 2027; binding constraints shift to power, cost of capital and local opposition; agentic workloads push CPU:GPU toward 1:1; memory tight through 2027. See 2027 Market Outlook Raising the Bar. Pre-IPO Anthropic ETF routes: Accessing Anthropic Exposure via ETFs and Mutual Funds.
- 2026-11-04: After a Democratic sweep of the US midterms, CFRA (note of 11/04/2026) frames possible data-center build-out restrictions (a moratorium in Texas, which has the largest pipeline) as a near-term risk for the chain but a structural positive for pricing power at the large hyperscalers. It expects capex to shift toward revenue-generating AI servers (about 60%-65% of spend today versus 35%-40% for buildings, land and power), flags that semiconductor revenue may slow more than consensus if a moratorium lasts months, and notes its semiconductor universe trades at about 13x 2028 P/E. Big Tech rallied (AAPL about +6%, MSFT and AMZN +6%-8%). This adds a policy leg to the Power Constraints risk below.
- 2026-10-16: CFRA's note on tighter financing conditions (CoreWeave's unsecured bonds yielding ~15%; CFRA cut its CRWV target to USD 70 from USD 140) argues the credit squeeze hits neoclouds and Oracle hardest and widens the advantage of MSFT, AMZN, GOOGL and META, which plan over USD 700B of combined capex this year.
- 2026-11-02: Anthropic's S-1 discloses >$500B of compute commitments (Broadcom >$160B lease, Google Cloud $114.5B, AWS $110B, SpaceXAI up to $120.5B, Azure $45B) and revenue-per-compute up 5.2x to $2.08. CFRA keeps 2027 capex growth at 40-50%, slowing to mid-teens in 2028-2030; Meta raised ~$100B equity ahead of >$200B 2027 capex; neocloud financing stays the pressure point. See Anthropic S-1 Compute Commitments & AI Capex Outlook.
- 2026-08-26: NVIDIA's FQ2 FY2027 print reset the capex debate. Revenue of $96.22B (+105.8% Y/Y) beat by $3.84B, Data Center hit $89.02B (+116.6% Y/Y), and the ACIE sub-segment — AI clouds, industrial and enterprise, i.e. the non-hyperscaler half — grew +138% Y/Y and +25% Q/Q. The material disclosure for this theme is the preliminary FY2028 revenue growth indication of ~70% against a ~45% consensus, given as explicitly supply-constrained: management stated customer demand implies closer to +100%, with memory supply the binding bottleneck. Supply and capacity commitments were raised to $279B. NVIDIA also disclosed guarantees to SB Energy Corp. covering a land, power and shell buildout of ~4.25 GW of IT load — the vendor-financing perimeter widening from GPUs into the physical estate. See NVIDIA Corporation.
- 2026-08-24: Broadcom detailed custom XPU roadmap with Meta (1 GW in 1H CY27 to 3 GW by CY28), Alphabet (TPUs through CY31), Anthropic (1 GW to +5 GW in CY27), and OpenAI (1.3 GW in CY27). Broadcom AI chip revenue projected at $56B in FY26 (+175%) and >$100B in FY27.
- 2026-08-20: Oracle Corp. reported Q4/FY26 RPO surging to $638B (+363% Y/Y) driven by massive AI infrastructure contracts (including the OpenAI Stargate partnership), with OCI cloud revenue up 39% to $34.0B.
- 2026-08-18: Ingested Q2 2026 updates highlighting the cloud backlog inflection to $2.5T and Alphabet's contractual purchase obligations reaching $811B.
Exposed Companies

TD Synnex
SNX US
Micron Technology, Inc.
MU US
Amphenol Corporation
APH US
Salesforce
CRM US
Apple Inc.
AAPL US
Vinci SA
DG FP
Dell Technologies Inc.
DELL US
ASML Holding N.V.
ASML NA
American Electric Power Company Inc.
AEP US
Microsoft Corp.
MSFT US
Palantir Technologies Inc
PLTR US
Analog Devices Inc.
ADI US
NVIDIA Corporation
NVDA US
Howmet Aerospace Inc.
HWM US
Amazon.com Inc.
AMZN US
Sandisk Corporation
SNDK US
Linde plc
LIN US
Trane Technologies plc
TT US
Western Digital Corporation
WDC US
Advanced Micro Devices Inc.
AMD US
Constellation Energy Corporation
CEG US
Broadcom Inc.
AVGO US
Vistra Corp.
VST US
Alphabet Inc.
GOOGL US
Arista Networks Inc.
ANET US
Cisco Systems Inc.
CSCO US
Flowserve Corporation
FLS US
Seagate Technology Holdings plc
STX US
Interactive Brokers Group Inc.
IBKR US
Meta Platforms Inc.
META US
VAT Group AG
VACN SW
Quanta Services, Inc.
PWR US
Sanmina Corporation
SANM US
Talen Energy Corporation
TLN US
Lam Research
LRCX USINFICON Holding AG
IFCN SW
Semtech Corporation
SMTC US
GE Vernova Inc.
GEV US
Applied Materials
AMAT US
Vertiv Holdings Co
VRT US
KLA Corporation
KLAC US
Morgan Stanley
MS US
Schneider Electric SE
SU FP
Corning Incorporated
GLW US
JPMorgan Chase & Co.
JPM US
International Business Machines Corporation
IBM US
Bank of America Corporation
BAC US
Eaton Corporation plc
ETN US
Marvell Technology Inc.
MRVL US
NextEra Energy, Inc.
NEE US
Oracle Corp.
ORCL US
ServiceNow Inc.
NOW US
Coherent Corp.
COHR US
Lumentum Holdings Inc.
LITE US
Qualcomm Inc.
QCOM US
Tesla Inc.
TSLA US
Palo Alto Networks Inc.
PANW US
CrowdStrike Holdings Inc.
CRWD US
Comet Holding AG
COTN SW
Intel Corporation
INTC US
Siemens AG
SIE GY
SpaceX
SPCX US
BYD Company Limited
1211 HK
BYD Electronic (International) Company Limited
0285 HK
Kioxia Holdings Corporation
285A JP
SK Hynix Inc.
000660 KS
Samsung Electronics Co. Ltd.
005930 KS