The Silicon Gold Rush: Unearthing Investments via AI's Picks & Shovels

Written by:
Ram Ahluwalia
Devesh Aggarwal

Introduction: The Dawn of the AI Era

Today Artificial Intelligence (AI) is not just a buzzword; it's a revolution reshaping our future. As we stand at the brink of this technological transformation, investors are grappling with a crucial question: Have they missed the boat on AI investments, or is the best yet to come? This blog post takes a step back to showcase the AI value chain revealing why the real treasure may lie not in the flashy AI apps, but in the robust silicon infrastructure that powers them – the picks and shovels of this modern gold rush.

Source: Wall Street Journal

The AI Investment Dilemma: Is It Too Late?

AI's potential economic impact is staggering. Experts predict that generative AI could boost the global GDP by a massive 7%. By 2025, AI-related investments are expected to surge to around $200 billion globally. Yet, the true shift in global labor productivity, a key indicator of AI's impact, is still on the horizon. This scenario presents a unique dilemma for investors: Is the AI investment wave still building, or has it already crested?

Source: Goldman Sachs
From previous trends labor productivity growth takes 5-10 years to materialize

The Enduring Value of Silicon

While AI applications come and go, the enduring value lies in the silicon layer, the robust foundation upon which AI stands. It's this durable infrastructure, often overlooked, that should be the focus of savvy investors. This is the technological bedrock which will power all future AI transformations. Unlike the fleeting nature of specific AI applications, the silicon infrastructure is enduring. For investors seeking longevity and stability, this foundational layer holds significant allure.

AI Represents a “Platform Shift”

Every decade or so, a new platform revolutionizes technology, and AI is the latest. A fierce “jump ball” is underway among tech firms racing to define the new user experience (UX) form factor. And this race isn't just about the applications; it's about reinventing our interaction with technology itself.

Where to Invest: Capex Payers vs. CapEx Receivers

When considering where to invest in the AI landscape, it's crucial to understand the concept of Capital Expenditure (CapEx). CapEx refers to the funds used by companies to acquire, upgrade, and maintain physical assets such as property, industrial buildings, or equipment.

Large tech firms (Google, MSFT, Amazon.) and startups such as OpenAI are spending billions in research to develop the infrastructure because the stakes are high, and it’s too early to bet on the winners in the Application Layer.

In the AI world, the application layer – consisting of AI applications and services – is a CapEx payer. This means they invest heavily in acquiring and upgrading technology infrastructure like GPUs and cloud AI services.
The Application Layer is expensive with intense competition.

OpenAI is now valued at $86 Bn with $1 Bn+ in revenue. Meanwhile, OpenAI is spending more than $1 Bn on the Silicon Layer

And this is where the Capex receivers come in. CapEx receivers, like the semiconductor companies, OEMs, chip designers etc provide the essential infrastructure. The Silicon Layer (Nvidia, TSMC, Amkor etc) is growing revenues quickly, and often has better valuations than the well-known brands comprising the Application Layer. It represents a more stable and potentially lucrative investment.

Beyond the Ordinary: The Silicon Layer

While AI applications grab headlines, the silicon layer quietly powers these innovations. This layer includes the hardware and infrastructure essential for AI operations. Investing in this layer means investing in the tools and machinery – the 'picks and shovels' – that make AI possible.

The Silicon Layer is not just important; it's indispensable. It facilitates efficient compute delivery and equips Large Language Models with essential inference and training capabilities. Today, it's at the forefront, addressing the massive strain on current compute capabilities caused by growing datasets and model sizes.

Navigating the Semiconductor Value Chain

The semiconductor value chain is a complex ecosystem involving:

  1. Chip Designers: Firms like Nvidia and AMD, designing the brains of our digital devices.
  2. IP Providers: Firms like Cadence and Rambus, providing the blueprints and crucial licensed technologies for chip designers.
  3. EDA Software Companies: Such as Cadence and Synopsys, these firms develop software to help design complex chips.
  4. Foundries: TSMC and GlobalFoundries, chip manufacturers, these are the factories where chip designs come to life.
  5. Lithography: The art of imprinting designs onto silicon, mastered by companies like ASML.
  6. Packaging and Testing: Specialists like Amkor Technology and ASE Group ensure that the chips are ready for real-world use.
  7. OEMs and Cloud/AI Providers: The end-users who integrate these chips into products and services.

Soure: McKinsey Research

Conclusion: A New Investment Frontier

The AI landscape is vast, filled with noise, and it's easy to be swayed by the allure of overvalued application companies. However, the real investment gems lie in the semiconductor firms, the unsung enablers of AI innovation. Follow the smart money that is betting on the less glamorous, yet crucial semiconductor sector.

In this Silicon Gold Rush, it's the providers of the picks and shovels – the silicon infrastructure firms – that could offer the most enduring and lucrative investment opportunities. For investors looking to capitalize on the AI revolution, the silicon layer is an opportunity to invest in the very backbone of this transformative technology.

If you are interested about investment opportunities in this sector join our waitlist, to talk to our relationship managers.

If you are looking for more content check out this interview on AI Chip Wars: LPUs, TPUs & GPUs with Jonathan Ross, Founder Groq.

Groq is the innovator of the novel Tensor Streaming Processor compute architecture, accelerating workloads in AI, ML, and HPC through their product portfolio.

Jonathan was one of the builders of the first Google TPU.

Text is copied
This is some text inside of a div block.