“The Bottleneck Isn’t the Chips Anymore”: Goldman Sachs Says These Are AI’s Next Big Winners

A Goldman Sachs executive says the AI trade is quietly shifting away from chips toward a different hardware category entirely, and two companies are already racing to fill a supply gap that could take years to close.

Published August 11, 2026, 6:25am ET · 3 min read

This post may contain links from our sponsors and affiliates, and Flywheel Publishing may receive compensation for actions taken through them.

Underwater scene featuring multiple dark, ribbed fiber optic cables running along a rocky, dark blue ocean floor. Bright blue and white light streaks erupt from open ends of the cables, resembling bursting data streams. Several translucent, glowing blue jellyfish and small fish swim in the deep blue water, illuminated by shafts of light from the surface. The scene conveys advanced technology and global connectivity in an aquatic environment.
Underwater fiber optic cables, like those depicted, symbolize the vital infrastructure driving the surge in optics stocks and the opportunities in new related ETFs. © 24/7 Wall St.

Editor’s note: This article has been updated to clarify the sourcing of Sung Cho’s comments. Cho’s remarks were made during an August 10, 2026 interview on CNBC, not in an interview with 24/7 Wall St. A link to the CNBC appearance has been added and the attribution in the text has been revised accordingly. Updated August 31, 2026.

Goldman Sachs Asset Management’s Sung Cho, co-head of public technology investing, argues the AI trade is rotating away from the graphics processor cycle toward companies wiring AI together, specifically optical and fiber equipment makers. Cho laid out that view in an August 10, 2026 interview with CNBC, and all of his comments below are drawn from that appearance. His picks: Lumentum (NASDAQ:LITE | LITE Price Prediction) and Coherent (NYSE:COHR).

Cho’s case starts with a shift in workload mix. “One of the most important trends that we’re seeing in the market today is the shift from AI training, driving most of the compute, to AI inference, driving most of the compute,” he said in the CNBC interview. “Underneath that architecture is a completely different set of architecture, a completely different set of chips, optical equipment. And so the leadership is going to evolve and change as this transition happens.”

Why Connectivity Is the New Constraint

The physical footprint changes with inference. “As you go into inference, you need a lot more data centers that are closer to the customers that they’re serving. As a result, we’re just going to have to connect a lot of data centers,” Cho told CNBC. Inside those buildings, the wiring becomes the ceiling. “What’s happening is that compute speeds, the processor speeds are no longer the bottleneck. What the bottleneck is, is actually the ability to be able to have chip to chip communication, server to server communication. And right now, a lot of those connections are happening via copper. And that’s going to be replaced by optical as well,” he added.

The supply-side setup gives the trade duration. “One of the unique aspects of optical is that it’s extremely hard to bring new capacity online. And so the demand for optical and fiber is moving at an accelerating rate as a result of this transition. But the ability for the industry to bring capacity online is going to be somewhat limited and keep that duration of that trade,” Cho said. A semiconductor ETF is up 80% year to date but down 20% from its 52-week highs, while memory stocks tripled over the last couple of years despite similar bottleneck labels.

Lumentum: Margin Expansion Backing the Story

Lumentum’s fiscal Q3 2026 report, filed May 5, 2026, validated the thesis. Revenue reached $808.4 million, up 90.1% year over year, with non-GAAP EPS of $2.37 and non-GAAP operating margin expanding 700 basis points sequentially to 32.2%. CEO Michael Hurlston flagged optical circuit switches with backlog above $400 million and co-packaged optics with an incremental multi-hundred-million-dollar order deliverable in first half calendar 2027 booked. Details are in the company’s 8-K filing. Shares closed at $813.51 on August 10, up 120.71% year to date and 599.67% over one year.

LITE earnings explorer

Coherent: Scaling Capacity to Meet AI Demand

Coherent, now an S&P 500 constituent, posted Q3 FY2026 revenue of $1.81 billion, up 20.5% year over year, with Datacenter & Communications contributing $1.36 billion, up 40.6% YoY and now 75% of total revenue. Non-GAAP EPS was $1.41, with non-GAAP operating margin at 20.3%. CEO Jim Anderson said Coherent is on track to double internal InP output by year-end 2026 and more than double again by 2027. NVIDIA’s $2 billion investment anchors a partnership around laser and optical networking gear, with new engines such as CPO/NPO, optical circuit switches, and multi-rail solutions adding $20+ billion in incremental serviceable addressable market by calendar 2030. Shares finished at $325.15, up 76.17% year to date.

COHR earnings explorer

What to Watch Next

Sell-side positioning tracks Cho’s thesis. Lumentum carries 5 strong buy and 16 buy ratings with an analyst target of $1,125.93, while Coherent shows 4 strong buys and 13 buys against a $394.62 target. Reddit sentiment for LITE swung to very bullish score of 82 on August 4. Monitor InP capacity ramps, CPO order flow into calendar 2027, and whether hyperscaler capex holds through the inference build-out Cho describes.

Contact [email protected] for any questions or corrections.

AJ Tiarsmith

AJ has spent the past 10 years writing about financial markets at The Motley Fool. His coverage centers on technology stocks and the broader macroeconomic trends, from interest rates to geopolitics,  that shape where markets are headed next. AJ is drawn to the stories where big-picture economics and individual companies collide.

All articles →