Bernstein Comments on Coherent (COHR.US) Q1 Results: No Issues with Fundamentals, Focus is on Timing of Realization, Maintains "Outperform" Rating

According to Wise Finance APP, Bernstein has released its latest research report on Coherent's (COHR.US) quarterly results: the quarterly report itself is fine, the issue lies in “when the goals will be realized.” Coherent’s Q1 FY2027 revenue was $2.37 billion and earnings per share were $2.09, both exceeding consensus by 2% and 4% respectively. The FY2027 guidance forecasts revenue growth of 60%, which is about 10 percentage points higher than consensus. However, the three levers supporting the $350 target price—the first $4 billion revenue in a single quarter, the first 45% gross margin quarter, and scale-up optical ramp—are all expected for late FY2027 to FY2028. The report maintains its rating and target price and points out there is no clear catalyst before the OFC conference in March.

Q1: Data Centers are the Sole Pillar

Total revenue was $2.37 billion, essentially in line with Bernstein’s $2.372 billion and about 2% above consensus at $2.315 billion. There are two main segments: data center & communications at $1.952 billion, surpassing both Bernstein and consensus (about $1.88 billion) by 4%; industrial at $419 million, which is 16% below Bernstein’s $497 million and 8% below consensus at $453 million. Data center & communications now account for 82% of revenue, with industrial weakness entirely offset by AI optical communications—this is why Bernstein did not let the industrial shortfall overturn their view.

Excess Margin Gains Due to Cost Control, with Cash Flow as a Flaw

Non-GAAP gross margin was 40.6%, below Bernstein’s 41.0%, and in line with consensus; operating margin was 23.2%, surpassing Bernstein’s 20.3% and consensus at 22.8%. EPS was $2.09, exceeding Bernstein’s $2.05 and consensus at $2.02. With a lack of surprise in gross margin but significant outperformance in operating margin, it indicates the margin gains are primarily from cost controls, not product mix or capacity efficiency—which directly addresses the long-standing investor dissatisfaction.

The true flaw lies in cash flow: free cash flow was negative $868 million, Bernstein’s original estimate was negative $617 million, while consensus expected a positive $286 million. The market originally anticipated cash generation this quarter, but instead saw significant outflows. Bernstein is relatively lenient: expecting positive cash flow in H2 FY2027 and for all of FY2028; after increased debt in Q1, net debt is still only 1.35 times the annualized adjusted EBITDA; long-term agreements include prepayments that arrive before capital expenditures, thus capital needs are moderate. In an environment where equity financing is heavily punished, relying on debt for incremental funding is more realistic.

Guidance: 60% Growth and Two “Firsts”

Q2 guidance calls for revenue of $2.55–2.75 billion, EPS of $2.28–2.48, gross margin of 40%–42%, non-GAAP operating expenses of $430–450 million, and a tax rate of 18%–20%. At the midpoint, revenue is 2% higher than Bernstein’s initial estimate and 4% above consensus; EPS is 4% above Bernstein’s and 6% higher than consensus.

The full-year guidance is even more significant: FY2027 revenue growth of 60% corresponds to about $11.3 billion, exceeding Bernstein’s previous forecast of $10.8 billion and consensus at $10.7 billion. The company also provided two “firsts”—the first single quarter with $4 billion in revenue in FY2028, and the first 45% gross margin quarter before the end of calendar 2028.

The second commitment is worth calculating. Using the Q2 guidance midpoint of 41% as a starting point, if the 45% margin is achieved in Q4 of calendar 2028 (Q2 FY2029), there are eight quarters in between, which means an average increase of about 50 basis points per quarter. This is slightly faster than the actual lift from Q1 and Q2's guidance (40 basis points per quarter), but roughly consistent with the previous eight quarters’ trend. Coupled with improvements in product mix and internal indium phosphide (InP) capacity ramp-up, Bernstein considers this “credible.” Accordingly, they’ve raised their predictions: Q2 EPS up from $2.28 to $2.45, full-year revenue to $11.5 billion, and EPS to $11.12, an increase of 16.9% from the previous $9.51.

CPO Mathematics: 7.7x Optical Content, $16.2 Billion Opportunity in 2028

The most information-rich section of the report breaks down the optical content in Nvidia Rubin Ultra into a verifiable formula. At the PhotonLink event, Coherent disclosed: a compute or switch chip with 100 Tbps IO and 200 Gbps per channel (500 channels) has an optical content value of about $15,000, i.e., $30 per channel.

According to Nvidia’s OCP design: 8 racks make a scale-up unit, 72 GPUs per rack, two-layer network; leaf nodes within a rack are 144-port NVLink 7 co-packaged optics (CPO) switch trays (2 x 200G channels per port), 72 ports connect to in-rack GPUs, 72 connect to the spine; the spine is deduced to have four spine structures, each spine-leaf link using 18 ports in a bundle. Bernstein assumes GPU-to-leaf uses copper, leaf-to-spine uplink uses CPO, and spine side uses pluggables. Each plane's CPO bandwidth is 230.4 Tbps, with pluggables matching that; Nvidia also disclosed that each Rubin GPU NVLink has a bandwidth of 3TB/s (12 Tbps, 30 x 400G ports), which gives 30 planes totaling 13,824 Tbps.

Keeping scale-out at NLV72’s 32.4 TB/s, the 8-rack unit is 259.2 Tbps. The result: 8-rack optical bandwidth rises from 2,073.6 Tbps (all scale-out) to 15,897.6 Tbps (scale-out 2,073.6, scale-up pluggable 6,912, scale-up CPO 6,912), optical content increases 7.7x; linear value conversion puts an 8-rack unit at about $2.4 million.

Multiply by shipments: Bernstein’s Asia semiconductor team (Mark Li) estimates 1 million R300 units in 2027 and 3.9 million in 2028, corresponding to about 1,800 and 6,800 Rubin Ultra units (each with 576 GPUs), at $2.4 million per unit, presenting a $4.2 billion opportunity in 2027 and $16.2 billion in 2028. Bernstein’s own assumption is 4 million units in 2028, yielding a similar conclusion.

The most sensitive assumption here is that the price per Gbps remains unchanged—in history, this price has consistently declined, and the current relative strength is due to the InP supply bottleneck. Coherent guides that InP capacity will double by late 2027; even if Rubin Ultra accounts for just 10% of total Nvidia GPU shipments, the doubled demand isn’t aggressive; at the same time, VCSEL and micro-LED will take up some demand, but Bernstein still expects supply-demand conditions to remain favorable through 2027. What they’re watching is other ecosystem players’ ramp-up—AXTI, Sumitomo, and Chinese vendors like Yuanjie are not within their coverage.

$20 Billion Long-Term Agreement and “Potential Acquisition Target”

The company emphasizes improved visibility and an expanded proportion of long-term agreements, disclosing for the first time specifics: a $20 billion, five-year, long-term agreement starting in early 2028. This is the second major long-term contract after supplying 400mW continuous-wave lasers for Nvidia’s CPO. Bernstein’s stance is “progress is exciting, but the track record is shaky”—using long-term contracts to hedge cyclical risks hasn’t always succeeded in the past.

Regarding the Marvell and Lumentum merger, the report offers a counterintuitive observation: major M&A usually benefits competitors, as the merging parties focus inward during the integration phase, giving rivals an opportunity to sow uncertainty and doubts among customers. Coherent further emphasizes it does not compete with its customers, while many of Marvell and Lumentum’s customers are also competitors. The report also speculates that Coherent itself may become a more attractive acquisition target, especially for customers concerned about supply; management, however, stresses satisfaction with current strategy.

The same section includes a blunt statement: Bernstein “worries Lumentum might be selling the business before fundamental deterioration,” which would be equally negative for Coherent; they also note that management appears unconcerned about threats from Asian competitors—perhaps overly unconcerned.

Valuation and Rating

The $350 target price is based on two equivalent paths: 31x FY2027 EPS of $11.12, or 19x FY2028 EPS of $18.04. Given the November 4 closing price of $195.09, this implies 79% upside; market cap is $38.425 billion, enterprise value $40.311 billion.

The divergence from consensus is mainly in the out-years: FY2027 EPS forecast at $11.12 is 17% above consensus at $9.50, FY2028’s $18.04 is 26% above consensus at $14.30, and FY2029’s $28.38 is 43% higher than consensus at $19.81. The risks are fourfold: CW+SiPh lasers ease the InP bottleneck and erode pricing power; China introduces InP export controls in response to US tariffs on optical modules; CPO yield issues slow adoption; or a slowdown in AI investment.

The stock price has already declined significantly: down 42.1% in the past month and 40.9% in the past six months, underperforming the S&P 500 by 41.7 and 47.6 percentage points, respectively. Bernstein acknowledges the lack of a short-term catalyst yet still offers a 79% upside—this contradiction is left for the reader to weigh.