The rapid expansion of artificial intelligence is creating what Piper Sandler describes as an “insatiable demand” for computing power, opening significant opportunities for technology companies that provide the hardware and infrastructure required to run increasingly sophisticated AI systems.
As businesses accelerate their adoption of artificial intelligence, demand is moving beyond software and AI applications themselves. Behind every generative AI platform, autonomous system and advanced machine-learning model is an enormous amount of computing infrastructure, ranging from high-performance processors and memory to networking equipment, storage systems and data centers.
Piper Sandler believes this structural increase in computing requirements could create substantial opportunities for companies positioned at the center of the AI infrastructure ecosystem.
The investment firm has highlighted two technology stocks that it believes could be particularly well positioned to benefit from this trend, reflecting a broader Wall Street view that the artificial intelligence boom remains heavily dependent on continued spending on physical computing infrastructure.
The fundamental investment thesis is relatively straightforward: as AI models become more capable, they generally require more computational resources. Companies capable of providing that computing capacity could therefore benefit even as competition among individual AI software developers intensifies.
AI Is Creating Enormous Demand for Computing Power
Artificial intelligence has rapidly evolved from a specialized technology into one of the largest areas of corporate investment.
Companies across industries are integrating AI into customer service, software development, advertising, cybersecurity, financial analysis, healthcare, manufacturing and logistics.
Generative AI has accelerated that trend.
Modern AI models must process enormous amounts of information during both training and inference.
Training involves exposing models to massive datasets and adjusting billions — or potentially trillions — of parameters.
Inference occurs after training, when users interact with an AI model and request answers, images, videos, code or other outputs.
Both processes require computing resources.
As the number of AI users increases and models become more sophisticated, the amount of computing infrastructure required to support them can rise dramatically.
That dynamic is at the center of Piper Sandler’s investment thesis.
Rather than viewing the current AI infrastructure boom as a temporary increase in hardware spending, the firm sees the potential for sustained demand as artificial intelligence spreads throughout the global economy.
Bigger AI Models Require More Infrastructure
One of the defining characteristics of modern artificial intelligence is scale.
Developers have discovered that increasing the amount of training data, model complexity and available computing resources can significantly improve the capabilities of AI systems.
However, those improvements come with substantial infrastructure requirements.
Large AI models can require clusters containing thousands of specialized processors working together.
Those processors need high-speed networking systems capable of moving enormous amounts of data between servers.
They also require advanced memory technology, storage systems, cooling equipment and significant electrical power.
The result is an AI infrastructure ecosystem that extends far beyond any individual semiconductor company.
Processor manufacturers may receive much of the attention, but AI computing also depends on memory suppliers, networking companies, server manufacturers, data-center operators, power-management providers and semiconductor manufacturing equipment.
As AI workloads increase, investment can spread across this entire supply chain.
Data Centers Are Becoming AI Factories
Data centers are one of the most visible beneficiaries of the computing boom.
Traditional data centers were largely designed around conventional cloud computing workloads such as websites, databases, file storage and enterprise applications.
Artificial intelligence changes the equation.
AI workloads can require much greater computing density and significantly more power than traditional cloud applications.
This is forcing technology companies and data-center operators to reconsider how facilities are designed.
Instead of simply adding more conventional servers, operators increasingly need infrastructure optimized specifically for high-performance computing.
That can include specialized accelerators, advanced cooling technologies, faster networking equipment and high-capacity electrical systems.
Some of the world’s largest technology companies are committing substantial capital to building and expanding data centers capable of supporting AI.
The scale of that investment illustrates why Wall Street analysts are increasingly focused on the companies supplying the underlying infrastructure.
The Semiconductor Industry Remains at the Center
Semiconductors are arguably the most important physical component of the AI revolution.
Advanced artificial intelligence systems rely heavily on specialized chips capable of processing large numbers of calculations simultaneously.
Graphics processing units, or GPUs, became particularly important because their parallel processing architecture proved well suited for machine-learning workloads.
But GPUs represent only part of the semiconductor opportunity.
AI servers require CPUs, memory chips, networking components, storage controllers and other specialized processors.
High-bandwidth memory has become especially important because advanced AI accelerators need enormous quantities of data delivered rapidly to maintain performance.
This creates opportunities across multiple parts of the semiconductor market.
It also creates challenges.
Manufacturing advanced chips is technically difficult and requires highly specialized factories, equipment and supply chains.
A shortage in any critical component can potentially limit the availability of entire AI systems.
Memory Is Becoming Increasingly Important
Memory technology is one area receiving growing attention as AI systems become more powerful.
A processor may be capable of performing enormous numbers of calculations, but its performance can be limited if data cannot move into and out of the chip quickly enough.
Advanced memory helps address that bottleneck.
AI systems therefore require not only greater amounts of memory but also faster and more energy-efficient memory.
This dynamic has increased the strategic importance of companies involved in high-performance memory production.
The transition also demonstrates how the AI boom can affect areas of the semiconductor industry that may initially appear less visible than processors.
As computing architectures evolve, the value of individual components can change substantially.
For investors, understanding these dependencies can be just as important as identifying the companies designing the most powerful chips.
Networking Is Another Critical Bottleneck
Thousands of processors working individually are far less useful than thousands of processors capable of communicating efficiently.
That makes networking infrastructure another essential component of large-scale AI systems.
Modern AI data centers require extremely high-speed connections between servers.
The faster information can move between processors, the more efficiently large computing clusters can operate.
Networking performance therefore directly affects the productivity of expensive AI infrastructure.
As companies spend billions of dollars on processors, they also have strong incentives to invest in networking systems capable of maximizing the performance of those processors.
This creates another layer of potential beneficiaries from the AI infrastructure boom.
Efficiency Could Become as Important as Raw Performance
The first phase of the AI computing race focused heavily on performance.
Companies wanted the fastest processors and the largest clusters possible.
But efficiency is becoming increasingly important.
Advanced AI infrastructure consumes enormous amounts of electricity.
Data centers also generate substantial heat, requiring sophisticated cooling systems.
As computing demand continues to increase, power availability could become one of the industry’s most significant constraints.
This means future AI infrastructure investment may increasingly focus on performance per watt rather than simply maximum computing capability.
Companies capable of delivering more computing output while consuming less electricity could gain a significant competitive advantage.
The same applies to cooling technologies, power-management equipment and data-center design.
AI infrastructure is therefore becoming as much an energy and engineering challenge as a semiconductor challenge.
Two Stocks Could Be Particularly Well Positioned
Piper Sandler’s analysis highlights two technology stocks as particularly attractive beneficiaries of the accelerating demand for computing infrastructure.
The companies are positioned within the hardware ecosystem that supports AI workloads and could benefit as businesses continue investing in processors, memory and data-center infrastructure.
The significance of Piper Sandler’s argument extends beyond the performance of any two individual stocks.
It reflects a broader investment strategy centered on the infrastructure required to make artificial intelligence possible.
Instead of attempting to predict which AI application will eventually dominate a particular industry, investors can look toward companies providing the underlying tools used by many competing AI platforms.
This approach is sometimes compared to supplying essential infrastructure during a major economic expansion.
Regardless of which individual AI models ultimately become market leaders, all of them require computing resources.
Cloud Computing Is Accelerating the Trend
Cloud computing companies are among the largest buyers of AI infrastructure.
Major cloud providers increasingly offer customers access to specialized AI computing resources without requiring businesses to purchase their own hardware.
This model makes advanced AI technology accessible to companies that could never afford to construct large computing clusters independently.
As more organizations adopt AI through cloud platforms, cloud providers must continue expanding their infrastructure.
That creates a reinforcing cycle.
Greater AI adoption increases cloud demand. Cloud providers purchase more computing equipment. Additional computing capacity makes AI services more accessible, which can encourage further adoption.
If that cycle continues, demand for high-performance hardware could remain elevated for years.
Autonomous Systems Could Expand Computing Requirements Further
Generative AI is currently receiving much of the attention, but other forms of artificial intelligence could create additional computing demand.
Autonomous vehicles, robotics and industrial automation all require sophisticated processing capabilities.
Some of that computing happens in centralized data centers, while other workloads must be processed directly on devices.
This creates opportunities for both cloud-based AI hardware and so-called edge computing.
As autonomous systems become more advanced, they may require increasingly powerful processors capable of analyzing large amounts of information in real time.
That could expand the AI semiconductor opportunity beyond traditional data centers.
Global Supply Chains Face Pressure
The extraordinary demand for AI hardware also introduces significant risks.
Advanced semiconductor production is concentrated among a relatively small number of manufacturers and geographic regions.
This creates potential vulnerabilities.
Geopolitical tensions, manufacturing disruptions, trade restrictions or shortages of critical materials could affect the availability of advanced computing hardware.
Technology companies are therefore attempting to diversify supply chains and secure long-term access to key components.
Governments have also increased efforts to expand domestic semiconductor production because advanced chips are increasingly viewed as strategically important infrastructure.
The AI boom is consequently influencing not only corporate investment but also industrial policy and international trade.
Investors Still Need to Consider Valuation
Strong industry demand does not automatically make every AI-related stock an attractive investment.
Expectations for future AI growth are already reflected in the valuations of many technology companies.
That creates an important distinction between identifying a successful industry and identifying a successful investment.
A company can experience rapid revenue growth while its stock performs poorly if investors previously expected even stronger results.
Competition also remains intense.
Technology changes quickly, and today’s dominant hardware architecture may eventually face competition from more efficient alternatives.
Investors evaluating AI infrastructure companies therefore need to consider revenue growth, profitability, competitive advantages, customer concentration and valuation alongside the broader industry opportunity.
AI Infrastructure Could Remain a Long-Term Investment Theme
Despite those risks, Piper Sandler’s description of “insatiable demand” captures one of the defining trends shaping the technology industry.
Artificial intelligence is fundamentally dependent on computing power.
As models become larger, applications become more sophisticated and adoption spreads across industries, the infrastructure required to support AI is expanding with them.
Processors, memory, networking systems and data centers all stand to play critical roles in that transformation.
The investment opportunity may therefore extend well beyond the companies creating the most recognizable AI applications.
The businesses providing the underlying computing infrastructure could become some of the most important beneficiaries of the entire AI cycle.
Piper Sandler’s focus on two particular technology stocks reflects that broader thesis: the AI revolution cannot continue without enormous amounts of computing power.
And if demand continues growing at the pace analysts expect, the companies capable of supplying that infrastructure could remain at the center of technology investment for years to come.
