The CPO and Semiconductor Boom: Sustained Growth Amid $690 Billion AI Infrastructure Surge in 2026
Date: May 6, 2026
Introduction
Wall Street has spent much of 2026 fixated on one question: how far can the AI infrastructure boom run? So far, the answer appears to be further than many expected. With major technology companies on track to spend as much as $690 billion in capital expenditures this year, the surge is reshaping the semiconductor industry and pulling a new class of winners into the spotlight.
Among them is Co-Packaged Optics, or CPO, a once-niche technology now becoming essential to the next generation of AI data centers. As demand for computing power climbs, traditional copper interconnects are running into hard physical limits. Optical solutions are moving from the margins to the core of AI infrastructure, helping support a semiconductor market projected to reach $1.3 trillion in revenue by the end of 2026, up 49% from a year earlier.
The CPO Breakthrough: Addressing AI's Energy and Bandwidth Demands
CPO is gaining traction because it addresses two of AI’s toughest constraints at once: bandwidth and power. By placing optical engines directly next to switch ASICs, accelerators, and chiplets, CPO allows data to move faster and more efficiently than conventional copper-based connections can manage.
That shift is not just technical, but economic. At the Optical Fiber Communication Conference in March, Coherent raised its CPO serviceable addressable market estimate to $15 billion, underscoring how quickly the opportunity is expanding. The technology also depends on advanced packaging methods such as 2.5D interposers and hybrid bonding, which are becoming increasingly important as photonic and electronic components are integrated more tightly.
Investors have noticed. Coherent shares were trading at $275.57 in early May, up 41.8% year to date, including gains of 25.3% over the past month and 14.2% over the past week. The rally reflects growing confidence that CPO is no longer experimental. It is becoming a necessary piece of the AI buildout.
NVIDIA Partnership and the $2 Billion Strategic Investment
That shift became even clearer in April, when Coherent signed a Securities Purchase Agreement with NVIDIA. The deal included a private placement of 7,788,161 common shares at $256.80 each, raising $2 billion in cash.
The funding is intended to support research and development, capacity expansion, and operating capabilities as Coherent expands its U.S.-based manufacturing footprint. Just as important, the companies announced a multi-year, non-exclusive strategic partnership that includes a multi-billion-dollar purchase commitment from NVIDIA, along with access and capacity rights for advanced laser and optical networking products.
For investors, the message is straightforward: large chipmakers are no longer treating photonics as a side bet. They are securing supply and technology access early, because they view optical networking as critical to scaling AI systems. NVIDIA has already said its Blackwell and Rubin AI chips represent a $1 trillion revenue opportunity through 2026, up from a previously cited $500 billion.
Massive AI Infrastructure Spending and Semiconductor Market Growth
The backdrop is a historic spending wave. Microsoft, Alphabet, Amazon, Meta, and Oracle are collectively expected to spend between $660 billion and $690 billion on capex in 2026, nearly double 2025 levels. In the first quarter alone, four major technology companies spent $130.65 billion, up 71% from a year earlier, with much of it directed toward data centers.
That spending is flowing across the semiconductor supply chain. Industry sales topped $400 billion in 2025, and 2026 is shaping up to be far larger. AMD’s data-center revenue is projected to rise 73% to $28.7 billion this year, while AI-related data center chips could account for roughly half of all semiconductor industry revenue.
Investment Implications and Forward-Looking Analysis
For retail investors, the opportunity is clear, but so is the bar. Companies exposed to AI infrastructure, especially those solving power and bandwidth bottlenecks, are attracting premium valuations. Coherent trades at about 35 times forward earnings, and many semiconductor names command similar or richer multiples.
That means execution matters. Investors should watch whether companies can expand manufacturing, meet hyperscaler demand, and translate announced spending into real shipments. The AI investment cycle still appears to be in its early stages, but capacity constraints, competition, and lofty expectations could separate leaders from laggards quickly.
Conclusion: A Transformative Infrastructure Buildout
This is starting to look less like a traditional semiconductor upswing and more like a long-term rebuild of the computing stack. CPO has moved from promising concept to strategic necessity, and the companies enabling that transition are becoming central to the AI era.
For investors, the key is not simply chasing the hottest name, but identifying which businesses have the technology, partnerships, and manufacturing discipline to turn this extraordinary demand into lasting growth.