A Goldman Sachs research report published on July 11, 2026, identified optical networking technology as artificial intelligence's next trillion-dollar investment opportunity. The report, which contributed to a 2.48% intraday gain for Goldman Sachs shares, predicts massive capital reallocation from AI compute to AI connectivity hardware. The analysis singles out Lumentum Holdings as a potential primary beneficiary of this structural shift.
Context — why this matters now
The investment thesis arrives as AI model complexity is outstripping traditional data center network capabilities. The last major network infrastructure upgrade cycle was driven by cloud computing around 2017-2019, when data center switch and router markets grew at a 12% compound annual rate. The current AI buildout is creating unprecedented demand for low-latency, high-bandwidth interconnects between GPU clusters.
Goldman’s focus on optics follows a multi-quarter surge in spending on AI accelerators from companies like Nvidia. As of this week, the PHLX Semiconductor Sector Index is up over 40% year-to-date, illustrating intense investor concentration on compute. The bottleneck has now visibly shifted from raw processing power to the speed of moving data between processors.
The immediate catalyst is the release of next-generation AI models requiring exaflop-scale training. These models demand co-location of hundreds of thousands of GPUs, a feat impossible with current copper-based electrical interconnects due to signal degradation and power consumption. Optical interconnects, which use light to transmit data, offer the required bandwidth and energy efficiency, compelling a wholesale architectural change.
Data — what the numbers show
Goldman Sachs’ own share price performance underscores the market's reception to its research. As of 13:27 UTC today, GS traded at $1,055.18, up 2.48% on the day and near the top of its $1,048.01 to $1,067.17 range. This move outperforms the broader financial sector, with the SPDR Financial Select Sector ETF (XLF) up only 0.8% over the same period.
Historical data shows the optical components market was valued at approximately $13 billion in 2025. Analyst projections cited by Goldman suggest this market could expand to a $75-100 billion annual revenue run rate by the end of the decade, representing a near eightfold increase. This growth trajectory is steeper than the semiconductor market's expansion during the initial AI investment wave from 2023 to 2025.
The investment bank's report details a specific performance metric. Replacing electrical links with optical interconnects in AI clusters can reduce power consumption for data transmission by up to 70%. For a large-scale data center, this translates to annual operating expense savings in the hundreds of millions of dollars, creating a powerful economic incentive for adoption.
A comparison of sector valuations reveals the opportunity. The median forward price-to-earnings ratio for AI chip designers exceeds 35x. In contrast, key optical component suppliers trade at a median forward P/E of 18x, suggesting the market has not yet priced in the projected growth surge from AI-driven demand.
Analysis — what it means for markets / sectors / tickers
The direct implication is a major re-rating for optical component and laser manufacturers. Lumentum, highlighted for its expertise in high-performance lasers and 3D sensing, is positioned to supply critical components for co-packaged optics. Other beneficiaries include II-VI Incorporated for its compound semiconductor materials and Fabrinet for its precision manufacturing scale. The shift is negative for incumbent makers of high-speed electrical interconnects and legacy networking gear.
A key limitation is the technological readiness and cost of scaling silicon photonics production. While lab prototypes demonstrate the necessary performance, mass production at a cost-effective yield remains a challenge. Supply chain constraints for specialty materials like indium phosphide could also delay the rollout, potentially capping near-term revenue growth for pure-play optics firms.
Positioning data indicates institutional investors are beginning to rotate capital. Flow analysis shows net inflows into the iShares Semiconductor ETF (SOXX) have plateaued over the last month, while the First Trust Cloud Computing ETF (SKYY), which holds several infrastructure names, has seen accelerating inflows. Hedge fund activity suggests new long positions are being established in mid-cap optical names, paired with short positions in legacy data center REITs with outdated infrastructure.
Outlook — what to watch next
The next significant catalyst is Lumentum’s fiscal fourth-quarter earnings report, scheduled for August 5, 2026. Management commentary on forward guidance for datacom revenue will be the primary focus for investors gauging the early commercial traction of AI optical links. Secondly, the OFC Conference in March 2027 will serve as the key industry event for product announcements and technology roadmaps from key vendors.
Critical levels to monitor include the $150 share price zone for Lumentum, which represents a key technical resistance level dating back to its 2021 peak. A sustained break above this level on high volume would confirm the bullish thesis is gaining broad market adoption. For the broader sector, watch the SOX/SOXL relative strength ratio against the Nasdaq-100; a breakout would signal sector leadership rotation.
Market direction will be conditional on these events. Should Lumentum’s guidance meet or exceed elevated expectations, a sector-wide rally in optical stocks is probable. Conversely, a delay in product timelines or a miss on guidance would likely trigger a sharp contraction in the elevated valuations of smaller, speculative component makers, while leaving the long-term thesis for market leaders intact.
Frequently Asked Questions
What is optical networking and why is it important for AI?
Optical networking uses light pulses transmitted through fiber-optic cables to send data. For AI, its importance stems from the need to connect thousands of GPUs within a data center with extreme speed and low latency. Electrical copper cables face physical limitations in bandwidth and power use over short distances, creating a bottleneck. Optical links can move terabytes of data per second with far less energy, enabling the scale required for next-generation AI model training.
How does this Goldman Sachs report compare to prior predictions about AI infrastructure?
Previous major infrastructure predictions centered on the demand for semiconductors, particularly GPUs and custom AI accelerators. This report represents a pivot, identifying the connectivity between those chips as the next critical, under-invested layer. The scale of the opportunity is framed similarly, with a projected addressable market in the hundreds of billions, but it targets a different segment of the hardware stack that has not yet seen a comparable valuation surge.
Which publicly traded companies are leaders in optical components for data centers?
