To The Moon

The AI investment story is gradually moving beyond GPUs and data-centre processors. As AI clusters become larger and more complex, moving data between thousands of processors is becoming just as important as the computing power itself.

This is creating new opportunities for companies supplying optical networking components, lasers, digital signal processors (DSPs) and high-speed interconnect technology.

Marvell is positioning itself to capture more of this market. However, its latest share-price movement highlights an important concern: is the company strengthening its competitive position through a potential acquisition, or taking on too much financial and integration risk?

Marvell shares fell 13.24% on Thursday, reversing the previous session’s 3.32% gain. Investors are weighing the financing, potential shareholder dilution and integration challenges associated with a possible acquisition of Lumentum.

The strategic logic is understandable. Combining Marvell’s networking chips and DSP technology with Lumentum’s optical and laser capabilities could create a more comprehensive portfolio for AI data-centre customers.

But strategic fit and shareholder value are two different things.

1. Why optical networking is becoming important in AI infrastructure

Most attention in the AI hardware market has focused on GPUs from companies such as Nvidia.

However, building larger AI data centres creates another challenge: transferring enormous amounts of information between processors, memory systems and storage.

Traditional electrical connections face increasing limitations involving power consumption, signal integrity and transmission distance.

Optical technology offers a potential solution by using light to transmit data at high speeds with improved efficiency over certain distances.

This is why technologies such as optical transceivers, lasers, silicon photonics and digital signal processors are becoming increasingly important.

As AI infrastructure expands, demand for faster and more efficient connectivity could become a significant investment theme.

The opportunity is not simply about selling more chips. It is about providing the infrastructure that allows the entire AI system to operate efficiently.

2. Marvell: Building a broader AI infrastructure business

Marvell has established itself in data-centre networking, custom silicon and high-speed connectivity.

Its products support the movement of data across increasingly complex computing environments.

A potential acquisition of Lumentum could strengthen its position by adding more optical components and laser technology to its existing portfolio.

The potential advantage is integration.

Rather than relying on separate suppliers for different parts of the connectivity system, customers could have access to a broader range of complementary technologies.

This could improve customer relationships, increase opportunities for cross-selling and potentially create a more competitive offering.

However, there are several risks.

An acquisition requires financing, and issuing additional shares could dilute existing shareholders. Taking on debt could also increase interest expenses and reduce financial flexibility.

There is also the challenge of integrating different businesses, technologies and customer relationships.

Most importantly, investors need evidence that the acquisition will generate enough additional cash flow to justify its cost.

Buying a complementary business does not automatically create shareholder value.

3. Lumentum: Positioned in the optical infrastructure opportunity

Lumentum is interesting because its business is more directly exposed to optical communications technology.

As data-centre operators increase bandwidth requirements, demand for advanced optical components and laser technology could benefit.

Its capabilities could become particularly valuable as customers move towards higher-speed networking architectures.

However, investors should consider the cyclical nature of the optical components industry.

Customer investment programmes can fluctuate, product transitions may create uneven order patterns, and manufacturing capacity must be managed carefully.

If demand is strong today, companies still need to demonstrate that orders can translate into sustainable revenue, operating margins and free cash flow.

For Lumentum shareholders, a potential acquisition also introduces another consideration: the price and terms offered matter just as much as the strategic benefits.

4. Coherent: Another way to participate in optical networking growth

Coherent offers another perspective on the optical infrastructure market.

Its portfolio includes optical communications components, lasers and photonics technologies used across several industries.

This broader exposure could provide different sources of demand beyond AI data centres.

However, the same diversification also means investors need to examine how much of its growth is actually coming from AI-related infrastructure rather than its other end markets.

I would pay particular attention to optical communications revenue, gross margins, order growth and cash generation.

Coherent’s performance could help investors assess whether the optical opportunity is translating into broader industry demand rather than being driven by a small number of major customers.

5. The biggest risk: AI demand does not guarantee acquisition success

I think this is the most important part of the Marvell story.

The AI infrastructure opportunity is real, but investors should not assume that every company expanding its product portfolio will automatically benefit.

There are several questions worth considering:

• Acquisition valuation: Is Marvell paying a reasonable price for Lumentum’s technology and future earnings?

• Shareholder dilution: How much additional equity might be required to finance the transaction?

• Debt and cash flow: Will the financing structure reduce financial flexibility?

• Integration: Can the companies combine their technologies and customer relationships without disrupting existing operations?

• Customer demand: Are customers committing to long-term orders, or could spending fluctuate between product generations?

Another important consideration is competition.

The optical networking market is attracting investment because AI data centres require faster connectivity. However, technology evolves quickly, and today’s competitive advantage may not remain permanent.

Companies must continue investing in research, manufacturing and product development to maintain their positions.

6. What I would watch before making an investment decision

Rather than focusing only on the potential acquisition, I would compare the three companies using several financial indicators:

Revenue growth: Is demand increasing across the existing business, even without acquisition-related contributions?

Gross margins: Can companies maintain pricing power as production volumes increase?

Free cash flow: Are growing orders translating into cash after capital expenditure and working-capital requirements?

Customer concentration: How dependent are they on a small number of large data-centre customers?

Valuation: Does the current share price already reflect substantial future AI growth?

For Marvell, I would also want to see clear acquisition terms, realistic synergy targets and a credible path towards improving earnings and cash flow per share.

My investment perspective

I find the optical networking theme attractive because it addresses a genuine bottleneck in AI infrastructure.

The market has spent considerable time focusing on computing power, but connectivity, bandwidth and power efficiency are equally important as AI systems scale.

However, I would separate the investment opportunity from the acquisition announcement.

Marvell may eventually benefit from owning a broader optical and networking portfolio, but the price paid and financing structure will determine whether shareholders actually gain value.

Lumentum offers more direct exposure to optical technology, while Coherent provides another route through its broader photonics and communications business.

I would be particularly interested in seeing whether all three companies can demonstrate stronger standalone demand and cash-flow generation before assigning significant value to potential acquisition synergies.

My biggest takeaway is that the next phase of the AI infrastructure race may be about removing connectivity bottlenecks rather than simply increasing computing capacity.

The optical networking opportunity could be substantial, but I believe companies should prove that their existing businesses can generate sustainable cash flow before investors give them full credit for future expansion.

For discussion purposes only, not financial advice.