Loading
For the last two years, the AI narrative has largely revolved around semiconductors. NVIDIA (NASDAQ: NVDA) became the face of the AI infrastructure boom. Hyperscalers poured hundreds of…
Archive edition · Market data and company circumstances reflect 16 May 2026, when this newsletter was sent.
For the last two years, the AI narrative has largely revolved around semiconductors. NVIDIA (NASDAQ: NVDA) became the face of the AI infrastructure boom. Hyperscalers poured hundreds of billions into GPUs, data centers, and accelerated computing clusters. Investors chased compute power as the defining resource of the AI age.
But a new problem is emerging. The challenge is no longer just generating computing power; it is moving data fast enough between those chips.
As AI models grow larger and training clusters become denser, the amount of data flowing inside data centers is exploding. The existing electrical and copper-based networking systems are beginning to face serious physical and energy limitations. Analysts now think the next major wave of AI infrastructure spending may shift toward optical networking.
Modern AI systems do not operate on a single chip.
Large-scale AI training requires thousands of GPUs working simultaneously across racks, clusters, and entire data centers. The more GPUs involved, the more important connectivity becomes.
Optical networking uses light instead of electrical signals to transfer data. This allows dramatically higher bandwidth, lower latency, and lower power consumption across massive AI clusters. This enables:
In simple terms, optical systems allow AI chips to communicate faster and more efficiently. And in AI infrastructure, communication speed increasingly matters just as much as raw compute power.
The industry is now moving toward ultra-dense AI superclusters where networking is no longer a supporting layer, it becomes a core performance driver.
AI infrastructure spending is rapidly expanding beyond semiconductors alone.
The market is beginning to realize that faster chips alone are not enough. Without high-speed networking systems, many of those compute gains cannot be fully utilized.
This has created growing interest in:
Industry estimates now suggest AI networking infrastructure could become one of the largest beneficiaries of the next wave of AI capital expenditure.
The transition is especially important because AI infrastructure is shifting from:
(Traditional cloud expansion across servers and racks)
(Extremely dense AI clusters requiring ultra-fast internal communication)
This “scale-up” architecture may dramatically increase demand for optical systems over the coming years.
Traditional electrical interconnects were designed for conventional computing workloads, not trillion-parameter AI systems. As AI compute density rises, copper-based systems face several challenges:
These problems intensify as hyperscalers continue building larger AI clusters.
Optical systems solve many of these issues because data is transmitted through photonics rather than electrical pathways.
The AI opportunity is rapidly expanding beyond software and GPUs. As models become larger and AI clusters become denser, the infrastructure required to support them is becoming more complex. Compute, networking, power, cooling, memory, real estate, and software orchestration are now all becoming critical parts of the same AI value chain.
Optical networking sits directly inside this transition. It helps solve one of the biggest challenges in AI infrastructure: Moving massive amounts of data between processors quickly, efficiently, and with lower energy loss.
Every major technology cycle eventually creates secondary infrastructure winners. The internet boom created networking giants. The smartphone era created semiconductor ecosystems. Cloud computing created hyperscaler dominance. AI may now be entering its infrastructure deepening phase.
The first wave rewarded compute providers. The next wave may increasingly reward connectivity providers, power infrastructure, data center operators, cooling systems, and companies enabling efficient AI deployment.
AI is no longer just a software story or a semiconductor story. It is becoming a full-stack infrastructure transformation.
As this shift accelerates, optical networking may move from a niche enabling technology to a foundational requirement of the AI economy.
As AI infrastructure spending broadens beyond semiconductors, several companies operating across networking, photonics, fiber optics, and data center connectivity are increasingly being viewed as potential beneficiaries of the next phase of AI expansion.
This article preserves the analysis in our weekly newsletter sent 16 May 2026. Market prices, forecasts and company circumstances reflect the time of publication and may have changed.
This material is general information, not personal financial advice or a recommendation to trade. Investing and trading involve risk, including loss of capital.