Buy These 2 Technology Services Stocks to Tap Recent Northbound Trend

Coherent forecasts fiscal Q1 2027 revenue of $2.2-$2.4 billion, with the $2.3 billion midpoint implying 45.6% year-over-year growth, alongside adjusted EPS guidance of $1.85-$2.05 and a $2 billion NVIDIA investment agreement through 2030. Cerebras reported Q2 core revenue growth of 103% to $209.9 million, raised 2026 core-revenue guidance to $880-$890 million, and holds $25.4 billion in remaining performance obligations. The article highlights both AI infrastructure companies as Zacks Rank #2 buys, citing average analyst-target upside of 29.0% for Coherent and 38.5% for Cerebras, though Cerebras remains projected to post negative operating margins in 2026.

Analysis

COHR is the cleaner expression of the AI networking bottleneck: compute scaling raises the value of bandwidth, switching and thermal management faster than it raises the value of commodity server components. The key underwriting question is not the headline revenue ramp but whether mix can lift gross margin materially beyond the guided near-term range; without that, the stock risks being valued as a cyclical optical-component supplier despite AI exposure. A deeper NVIDIA ecosystem relationship can also disadvantage more interchangeable optics vendors, including LITE and AAOI, if design wins migrate toward integrated photonics and co-packaged-optics architectures.

CBRS has a more asymmetric but materially less de-risked setup. Backlog/RPO and contracted power do not equal recognized, cash-generative revenue: conversion depends on customer deployment schedules, power interconnection, financing and utilization, while the operating loss profile makes any delay a valuation and capital-needs issue. Its avoidance of constrained HBM/advanced packaging is a real supply-chain advantage, but it also leaves execution risk around software compatibility and whether inference customers prioritize lowest total cost per token over incumbent CUDA portability.

Over the next 1-3 months, COHR's report and evidence of AI-product mix expansion are the relevant catalysts; analyst targets are not. Over 6-18 months, data-center power scarcity shifts value toward suppliers that improve compute utilization per MW—optics and inference architectures both qualify—but it constrains CBRS's revenue realization more directly than COHR's component sales. Consensus is likely underweighting the possibility that hyperscalers vertically integrate optical networking or standardize around alternative interconnects, which would cap COHR's terminal multiple even if current demand remains strong.

Key Decisions for Investors

• Initiate a 6-12 month long COHR / short LITE pair, sized beta-neutral. COHR has superior exposure to higher-value AI optical content; exit if COHR fails to show sequential gross-margin expansion or reduces its AI/datacom outlook at the next two reports.
• Accumulate COHR only on post-earnings volatility rather than chase momentum; use a 12-15% stop from entry. Base-case upside requires sustained estimate revisions and a mix-driven multiple re-rating, while downside is a cyclical optics de-rating if margins remain flat.
• Keep CBRS on an event-driven watchlist, not a core long, until management discloses contracted-capacity conversion, customer concentration, capex/financing requirements and cloud utilization. A long becomes actionable after evidence that service revenue growth converts into narrowing operating losses; failure to improve operating margin through 2027 is thesis-invalidating.
• For a higher-risk tactical expression, buy CBRS call spreads 6-9 months out rather than equity: cap premium at risk while retaining exposure to capacity activation or a major inference-customer announcement. Avoid naked short exposure given potential backlog-validation and strategic-partnership squeezes.
• Monitor NVDA's next platform and networking disclosures for CPO adoption timing. A delayed CPO transition favors conventional transceiver suppliers near term but weakens the strategic premium assigned to COHR's longer-duration photonics optionality.