Two AI Infrastructure Plays for the 2026 Growth Wave

The scale of the AI capital expenditure wave is nothing short of generational. For 2026, the Big Five hyperscalers-Amazon, Microsoft, Google, Meta, and Oracle-are projected to spend over $600 billion on infrastructure, a 36% increase from 2025. This isn't just a spike; it's a sustained acceleration from a base that already doubled in 2025. The composition of that spending reveals the core driver: roughly 75% ($450B) targets AI infrastructure. This concentration means the demand for GPUs, custom processors, data center construction, and the networking that ties it all together is now the primary growth engine for these tech giants.

Funding this unprecedented buildout has forced a fundamental shift in capital intensity. Hyperscalers are now spending 45-57% of revenue on capex, a level historically seen in industrial or utility companies, not software firms. To bridge the gap between spending and internal cash flow, they have turned to the debt markets with abandon. In 2025 alone, the Big Five raised $108 billion in debt, a figure that is 3.4 times the historical average. Projections suggest the technology sector may need to issue $1.5 trillion in new debt over the coming years to finance the AI infrastructure buildout.

This creates a massive, scalable market for infrastructure providers. The $450 billion dedicated to AI infrastructure is a direct pipeline for companies like Celestica, which designs and builds the data center networking switches that enable fast connectivity in these AI systems. The sheer volume of spending, coupled with the necessity to finance it through debt, signals a multi-year investment cycle. For players embedded in this supply chain, the opportunity is to capture a growing share of this expanding pie. The cycle is set to continue, with Goldman Sachs projecting total hyperscaler capex from 2025-2027 to reach $1.15 trillion-more than double the spending of the previous three years. This isn't a fleeting trend; it's a structural reordering of capital allocation that defines the next era of technology growth.

Stock Pick 1: Celestica (CLS) - The System Integrator Play

Celestica's growth story is a textbook case of a company positioned exactly where the AI wave is hitting. The company operates as a system integrator, providing the critical design, engineering, and manufacturing for data center networking switches that enable the high-speed connectivity required by AI systems. This isn't a peripheral role; it's a core component of the infrastructure buildout. As hyperscalers race to deploy custom AI processors, they need partners like Celestica to rapidly scale the networking hardware that ties it all together.

The strength of its position is underscored by its client base. Celestica counts major players in the AI chip industry among its partners, including custom AI chip leader Broadcom and companies like Marvell Technology, AMD, and Intel. This direct access to the ecosystem driving AI hardware demand provides a built-in growth engine. The results are already impressive: its connectivity and cloud solutions (CCS) segment, which makes up 76% of revenue, saw revenue jump 43% year over year in Q3 2025 to $2.4 billion. The company is gaining market share in this fast-growing space, with its estimate of controlling 55% of custom Ethernet switches last year, up from 40% in 2024.

The scalability of Celestica's model is a key advantage. By serving multiple hyperscalers and chipmakers, it mitigates the risk of over-reliance on any single customer. This multi-client approach is essential for capturing a larger share of the massive hyperscaler capex wave. The company's growth narrative is clear: it is positioned to benefit from the projected 78% increase in the custom AI processor market this year. Its 2025 revenue guidance of $12.2 billion implies a 26% jump, with even steeper earnings growth expected. For 2026, management anticipates a 31% top-line jump.

The financial upside is substantial. If Celestica achieves its ambitious goal of $16 billion in revenue this year and trades at a multiple in line with the broader market, its market cap could hit $88 billion. That would represent more than double its current valuation. The company is already seeing new opportunities, with a hyperscaler approaching it to design rack-scale networking solutions, with mass production expected in 2026. In essence, Celestica is a scalable play on the AI infrastructure buildout, leveraging its system integrator model to capture growth across multiple clients in a market that is only beginning to ramp.

Stock Pick 2: Advanced Micro Devices (AMD) - The Diversified Hardware Challenger

While Nvidia dominates the current GPU-centric AI market, Advanced Micro Devices is carving out a formidable position as a diversified challenger with a scalable platform for the next phase of growth. Its strategy is built on two pillars: aggressive competition in the AI hardware race and a broad portfolio that provides a durable, multi-year growth trajectory beyond any single segment.

AMD's competitive positioning is defined by its accelerating investment in AI infrastructure. The company's revenue has grown at an annualized rate of over 20% over the past two years, and management believes this growth will accelerate significantly as AI spending ramps. This isn't just a story of one product line. Its diversified portfolio powers an AI-driven economy, spanning data center CPUs, GPUs, and adaptive computing chips like FPGAs. This breadth provides a scalable platform to capture growth across different parts of the AI stack, from massive data centers to edge devices.

Recent financial results show this momentum gaining speed. Last quarter, AMD's revenue grew 24% year-over-year, outpacing its 18% top-line growth rate. The data center segment is the engine, with revenue hitting a record $4.3 billion in the third quarter. Management expects this business to grow at an annualized rate of over 60% in the next five years. The client segment, driven by Ryzen processors, also saw a robust 46% year-over-year increase to a record $2.8 billion. This multi-pronged growth is critical for sustainability, as it reduces reliance on any single market cycle.

The long-term outlook is where AMD's scalability becomes most apparent. The company's own forecast calls for its annual revenue to grow at a 35% compound annual basis over the next three to five years. This ambitious target is supported by a clear path: continued market share gains in PCs and gaming, a massive expansion in data center AI, and the commercialization of its adaptive computing portfolio. For a growth investor, this represents a company not just riding the AI wave, but building a fleet of boats to navigate the entire ocean of technological change.

Valuation, Catalysts, and Risks

The growth thesis for AI infrastructure plays rests on a foundation of massive, scalable demand. The total addressable market is vast, with the global AI market projected to grow at a CAGR of 26.60% to $2.48 trillion by 2034. This isn't a niche opportunity; it's a multi-decade expansion of the economic landscape. For companies like Celestica and AMD, the key is capturing a growing share of the core infrastructure buildout, which is already a $450 billion annual market within the hyperscaler spending wave.

The primary near-term catalyst is the execution of the $602 billion capex plan. Any upward revisions to consensus estimates, which have historically been too low, will be positive for sentiment and valuation. The trend is already moving in that direction, with the consensus estimate for 2026 capital spending climbing to $527 billion from $465 billion at the start of the third-quarter earnings season. This divergence in performance among AI stocks shows investors are becoming selective, rewarding those with a clear link between spending and future revenue. The catalyst is therefore not just the spending itself, but the proof that it translates into sustainable growth for the suppliers.

Yet, the path is not without significant risks. The first is customer concentration. The entire buildout is driven by a handful of hyperscalers-Amazon, Microsoft, Google, Meta, and Oracle. Their collective spending creates immense opportunity, but it also concentrates risk. A slowdown in any one of these giants' AI ambitions could ripple through the supply chain. This is the inherent vulnerability of being a system integrator or a hardware supplier to a few dominant clients.

The second, more profound risk is the potential deflation of the AI bubble. As MIT Sloan experts predict, the AI bubble will deflate, and the economy will suffer. This doesn't mean AI stops being important, but it signals a shift from the current phase of massive experimentation and investment to one focused on operational integration and productivity. The market may re-rate companies from those spending heavily on capex to those demonstrating clear returns on that investment. For infrastructure providers, this could mean a period of margin pressure as hyperscalers demand more efficiency, or a slowdown if the integration phase is longer than expected.

In essence, the setup is a high-growth race against a backdrop of volatility. The TAM is enormous, and the capex catalyst is real and accelerating. But the winners will be those who can navigate the concentration risk and adapt to a market that may soon demand more than just spending-it will demand proof of value.