Lumentum Holdings, a key optical communications supplier, has provided a definitive stance on the industry's most debated architectural question. The company clarifies that NPO and CPO are not a zero-sum game, but rather overlapping incremental opportunities, with both tracks forming the core drivers of its future revenue growth.

According to Citi Research's review of Lumentum's management commentary following its June quarter earnings call, management explicitly stated that NPO deployment intensity from non-Nvidia customers is already significantly higher than current order demand from Nvidia. This makes the emerging NPO opportunity exhibit a higher optical intensity multiplier effect on an aggregate basis. Concurrently, management refuted the market's tendency to tie the company's scale-up shipment cadence directly to Nvidia's Kyber rack timeline, reiterating that there is no direct dependency between the two.

This statement suggests that the earlier market concern about "NPO's rise cannibalizing CPO demand" was a misjudgment. For investors, Lumentum's potential total addressable market may be underestimated. Citi Research maintains a Buy rating on the stock with a price target of $1100, implying approximately 34% upside from the current share price of $820.59.

NPO and CPO Advance in Parallel, Shipment Timelines Become Clearer

Management provided more specific guidance on the shipment timelines for both architectures. On the CPO front, north-south NPO opportunities related to customers like Nvidia are expected to begin shipping in the fourth quarter of 2027, with volume ramping into early 2028. In-rack CPO connections, which are vertical interconnects related to scale-up applications, maintain prior guidance, with shipments expected in the second half of 2028 and deployment in 2029.

Management specifically emphasized that the very existence of multi-rack configurations constitutes a sufficient condition for north-south optical interconnect demand between racks, and this opportunity is independent of the specific launch date of the Kyber rack. This statement directly addresses market concerns that delays in Kyber could negatively impact Lumentum's shipment cadence.

Notably, management also confirmed that current scale-out CPO shipments are already underway, scale-up demand continues to exceed supply, and no signs of hesitation have been observed from lead customers.

Non-Nvidia NPO Customers Provide Additional Incremental Growth, with Higher Optical Intensity

When explaining the NPO opportunity, management introduced an analytical framework of "revenue opportunity multiplied by optical intensity," a perspective that reveals the unique value of non-Nvidia NPO customers. For Nvidia, only a smaller portion of CPO and scale-up opportunities is expected to adopt an external laser source (ELS) approach. For other customers, ELS represents a much larger share of the total opportunity, covering all demand in some cases.

Based on this, management concluded that even if the scale advantage of recent Nvidia orders is more significant, the attractiveness of non-Nvidia NPO customers on an aggregate basis should not be overlooked. Additionally, management pointed out that some non-Nvidia NPO customers choose to embed lasers directly inside the optical engine rather than using an external laser source. This technical path drives demand for mid-power embedded lasers (120mW to 150mW), creating an additional incremental customer base beyond Nvidia's clientele.

Mid-Power Embedded Lasers: No Direct Competition Yet, but a Later Timeline

In the niche market for 120mW to 150mW low-power embedded lasers, management stated they have not yet seen direct competitors, while acknowledging that this area is still in its early stages. From a technical barrier perspective, management believes that adjusting a 400mW laser downward to 120mW-150mW is structurally more feasible than expanding a 70mW-100mW low-power product upward to 120mW. This is because the reliability requirements for the high-power end are comparable to or even stricter than the target power range, providing a technical derivative advantage.

Regarding production ramp-up, the shipment timeline for this niche is later than for ELS-driven CPO and NPO opportunities. Management expects shipments could occur in the first half of 2028, with volume production ramping up in the second half of the year. Management admitted that the company has not yet formally secured design wins in this embedded engine category, but the shipment timeline implies these projects are targeting scale-up applications.

Japanese Fab Capacity Reallocation: Strategically Capturing CW Incremental Revenue

Management disclosed a significant capacity allocation decision: reallocating 10% to 20% of excess capacity at its Japanese wafer fab from EML (electro-absorption modulated laser) to CW (continuous wave) laser production. This decision appears somewhat counterintuitive on the surface, as EML has higher gross margins and the company's EML supply remains significantly constrained.

Management provided two strategic justifications for this move. First, following chip size redesign and production line modifications, the profitability of CW lasers has improved significantly, narrowing the gap with EML. Second, as most transceivers transition to the silicon photonics path, the CW market size is expected to expand substantially before 2027. Management emphasized that this capacity reallocation does not reflect any change in EML demand outlook. Instead, it is a strategic choice to actively capture CW incremental revenue while the EML supply-demand tension persists.

Citi Research values Lumentum using a 40x price-to-earnings multiple, representing a premium of approximately 14% over the current 35x forward P/E average for AI optical peers. Citi cites the rationale that Lumentum's earnings per share growth rate from 2025 to 2028 is expected to be 2 to 3 times that of its peers. Under a relative valuation framework adjusted for earnings growth, this premium is considered reasonable. The price target remains at $1,100.