The compute economy

Growing application demand must first translate into a need for more computing. Investment then spreads across processors, manufacturing, memory, and networking. A computing system depends on all of these layers, so one chip trend cannot explain the entire infrastructure cycle.

NVDA and AMD cover computing products; TSM and MU cover manufacturing and memory; AVGO and ANET cover custom silicon and connectivity. Customer-designed systems, product substitution, and changing architectures can move value between these layers.

The building blocks of the idea

01

Computing products and ecosystems

42%

NVDA and AMD connect through accelerated computing, software, CPUs, and GPUs. Review how capabilities translate into customer adoption and repeat investment.

NVDAAMD

02

Manufacturing and memory

32%

TSM and MU help make computing designs manufacturable and available. Compare capacity commitments with actual demand.

TSMMU

03

Custom silicon and networking

26%

AVGO and ANET connect through dedicated processing and cluster connectivity. Customer architecture choices can create opportunities or substitution.

AVGOANET

Select a theme to filter the holdings. Its allocation reflects your current weights.

Weight	Stock / company	Role in the thesis	Onchain asset
28%	NVDANVIDIA	Accelerated computing chips and software support AI infrastructure.	0xd060…9EEC18 decimals · Active
20%	TSMTaiwan Semiconductor Manufacturing	Advanced foundry processes serve high-performance chip manufacturing.	0x58Ff…e7AA18 decimals · Active
17%	AVGOBroadcom	Custom accelerators and networking chips connect to data center investment.	0x156E…fbcF18 decimals · Active
14%	AMDAMD	CPU and GPU products participate in high-performance computing.	0x8692…3fdC18 decimals · Active
12%	MUMicron Technology	Memory products support servers and accelerated computing systems.	0xfF08…4afD18 decimals · Active
9%	ANETArista	Data center networking connects large computing clusters.	0x28bA…d1Bf18 decimals · Active
Filtered total 100%Portfolio total 100%

Asset addresses come from the Robinhood catalog snapshot. Links show each stock token’s onchain information, not this portfolio’s reserve contract.

1. 01

   Usage may outgrow spending

   Efficiency gains, higher utilization, or budget limits can slow hardware purchases even as applications grow.

2. 02

   Products and architectures can shift

   Tradeoffs between general processors, accelerators, and custom systems can redistribute value across chips, software, and networking.

3. 03

   Supply may arrive after demand weakens

   Manufacturing and memory capacity require advance investment. Lower demand can pressure inventory, utilization, and pricing.

4. 04

   The same customers drive the cycle

   Different suppliers can depend on the same large infrastructure projects and face spending reductions together.

1. Original allocation

   09/06/2026, 08:00

   A curated research portfolio. Review the assumptions and adjust the allocation to reflect your view.

   Current allocation · 6 holdings

This history is saved in your browser. Publish the portfolio to create a permanent link.

Research source

Curated theme

Model

Curated research

Asset catalog

Robinhood asset API · 194 assets

Catalog snapshot

09/07/2026, 00:28 UTC+8

Network

Robinhood Chain Mainnet · 4663

Strategy validation

No backtested performance

The initial holdings and reasoning come from curated theme research.Stock addresses are drawn from the official catalog. Before saving, the portfolio is checked for duplicates, holding limits, individual weights, and a 100% total. Review each company’s identity and business connection before making a decision.