The latest US inflation data released on September 11 showed both headline and core CPI rising 0.4% and 0.3% month-over-month respectively, with core inflation exceeding expectations. Despite this, the S&P 500 and Nasdaq Composite still climbed 0.86% and 0.96% respectively, as falling oil prices and strong earnings from AI-related tech companies continued to provide bullish support. This highlights how the market remains balanced between inflationary pressures from geopolitical conflicts and the trajectory of earnings growth. Wall Street's bullish thesis for Marvell Technology (MRVL.US), which focuses on AI ASICs and datacenter optical interconnect chips, embodies the latter logic—namely that the earnings trajectory driven by the AI computing theme overrides all negative narratives.

Bank of America has maintained its $365 price target for Marvell, citing the company's ability to expand revenue share per AI system through custom AI accelerators (AI ASICs) designed for massive AI agent inference demands, along with complementary optical interconnect chips. Relative to the September 11 closing price of $236.10, this target implies roughly 55% potential upside. However, the projected $40-45 billion revenue opportunity by 2030 is a scenario-based estimate, with actual realization depending on customer projects, product share, and profitability.

Why AI compute demand continues to expand the value of datacenter AI chips and optical interconnects

From an engineering and business model perspective, Marvell's core benefit from AI expansion lies in "custom AI acceleration computing plus system-level optical interconnect/optical communication." ASICs designed for specific workloads can improve energy efficiency and reduce large-scale deployment costs by optimizing compute units, memory access, and data flow. Meanwhile, more accelerators urgently require comprehensive supporting networks, memory interface chips, and optical-based high-speed datacenter connectivity chips. Marvell refers to these supporting businesses as "XPU companion chips," covering PCIe retimers, coprocessors, CXL controllers, high-speed optical interconnect chips, and key co-packaged optics (CPO) components.

According to Bank of America's latest estimates, each custom processor paired with one to two companion chips priced between $500 and $1,500 can generate multiple revenue increments for the company as customers scale up. The complexity of high-speed analog circuits, interface intellectual property, and system verification also creates barriers that make it difficult for customers to replicate these products on their own.

Marvell's second growth curve—the "optical interconnect chip business" beyond co-developing AI ASICs with cloud giants like Amazon—is fundamentally driven by the fact that effective compute power in AI clusters increasingly depends on cross-node communication efficiency. Expert parallelism in mixture-of-experts models, distributed inference, and systems separating pre-fill and decode phases all increase data exchange. Network congestion can cause expensive accelerators to wait for data, reducing overall cluster utilization. As transmission rates and distances increase, the losses, signal compensation, and power consumption pressures of copper connections intensify, pushing more links toward optical connectivity. Marvell participates through optical communication digital signal processors, high-speed serial interfaces, drivers, and related interconnect products. The company's already-in-production Ara platform and expanded 1.6T product portfolio directly serve this upgrade cycle.

Lumentum's demand outlook provides corroborating evidence from another part of this supply chain. Management commentary captured in Citi meeting notes indicates that shipments of optical module components (i.e., optical modules for tightly coupled AI accelerator interconnects) are expected to roughly double from 2026 to 2027, with EML and CW laser shortages projected to persist through 2027. Improved visibility in OCS and NPO businesses supports an upward revision of fiscal 2028 EPS targets to $40. The two companies benefit from different segments: Marvell focuses on semiconductors that process, transmit, and connect data, while Lumentum supplies laser sources, optical components, and optical switching systems. Consequently, the growth logic for AI chips and optical interconnects still has a foundation to continue, but sustained stock price gains require ongoing delivery on orders, capacity certifications, gross margins, and cash flow.

Bank of America sends a strong bullish signal to Marvell stock investors

According to Seeking Alpha data, shares of this AI chip giant have surged more than 160% over the past six months, with demand for its datacenter chips fueling market optimism. In a report shared with TheStreet, Bank of America analyst Vivek Arya, after meeting with CEO Matt Murphy and CFO Dan Durn, believes there are ample reasons to remain firmly bullish on the stock. Wall Street's overall assessment of Marvell leans positive, though price targets vary.

As of September 11, StockAnalysis showed a consensus rating of "Strong Buy" from 45 S&P Global analysts, with an average one-year price target of $284.64, implying about 20.6% upside from the $236.10 closing price, while negative targets suggest roughly 20% downside. The highest target in that sample is $400, corresponding to approximately 69.4% potential upside. Public rating records show KeyBanc analyst John Vinh maintained an "Overweight" rating and $400 price target on August 28. Bank of America's $365 target likewise sits well above the market average. The common bullish direction across Wall Street is the expansion of AI datacenter businesses, with divergence centered on the speed of growth realization and the valuation multiple the market is willing to grant.

As of September 11, the stock was up nearly 12%, while large semiconductor peers such as NVIDIA (NVDA.US) fell more than 5% over the past week. Specifically, Marvell builds the hardware that powers AI operations. The company designs custom processors as well as high-speed optical interconnect chips that transmit data between processors, enabling massive computing systems to work together efficiently. Its expanded partnership with Google is particularly notable, covering chips that support the search giant's AI infrastructure. Additionally, that agreement gives Google the right to purchase Marvell shares at specific prices depending on future purchasing volumes. But Arya's latest views, shared after lunch with management, go beyond any single customer. Bank of America emphasizes that Marvell's broadening product portfolio could further deepen its role within AI systems. With the Analyst Day approaching on October 6, investors may soon gain clearer insight into just how large this opportunity truly is.

Bank of America maintains a "Buy" rating and a high $365 price target for Marvell stock, implying nearly 55% upside from its September 11 close of approximately $236.10. The analyst team led by Vivek Arya bases its confidence on Marvell's potential to sell more technology into each AI system. Arya and his team believe Marvell's biggest opportunity lies in companion chips that connect processors, manage memory, and transmit data. Compared to the processors themselves, this business is far more difficult for customers to replicate. Marvell already delivers these products to all four US hyperscale cloud providers, and each of its custom processors requires one to two companion chips priced between $500 and $1,500. This creates new opportunities for Marvell to increase revenue from these customers as they expand compute capacity.

Bank of America estimates that by 2030, this companion chip market alone could exceed $60-65 billion. Based on a projected 40-50% market share, Marvell's annual sales opportunity could reach $30 billion, compared to management's fiscal 2028 outlook of $3-4+ billion. Additionally, custom processors provide another significant growth driver. Bank of America believes Marvell's potential sales in this area could reach $15 billion by 2030, bringing the combined opportunity to $40-45 billion. However, it's important to note that these are model-based projections, not confirmed orders. The earnings potential also explains this optimism. Analyst Arya estimates Marvell's fiscal 2028 EPS potential at nearly $14, above a $11 baseline level; every additional $1 billion in sales could add $0.30-0.35 to EPS. His price target applies a 33x multiple to baseline earnings excluding stock-based compensation. Thus, the bullish scenario relies heavily on strong growth as well as investors continuing to grant a premium valuation while Marvell executes its plans. For reference, according to Seeking Alpha data, the stock's non-GAAP price-to-earnings ratio over the past five years is 42x, so the 33x baseline valuation requirement is not especially demanding.

What could interrupt Marvell's uptrend?

The biggest risk at the stock level is the gap between the massive market opportunity and actual realized earnings. The $40-45 billion sales scenario envisioned by Bank of America relies heavily on market expansion, customer spending, and significant stock price appreciation. These sales revenues have yet to be realized. Customer project execution is arguably the first test. Bank of America notes uncertainty surrounding next-generation chip projects at Amazon and Microsoft. Product launch delays could push sales revenue out while R&D costs continue to accrue, undermining the earnings growth investors expect. Competition is likely to further amplify this risk. Broadcom (AVGO.US), the share leader in AI ASICs, may remain the strongest incumbent in custom silicon, while standardized AI accelerators dominated by NVIDIA and AMD also compete for the same spending budget. Marvell's broad interconnect portfolio provides multiple growth paths, but winning business does not necessarily guarantee pricing power or the margins required by Bank of America's earnings assumptions.

Additionally, the Google agreement deserves close scrutiny. The much-discussed $120 billion figure involves a framework tied to purchasing, not guaranteed orders. As Peace Longe pointed out in raising concerns, nearly 59 million shares could potentially be issued under that warrant, bringing both commercial opportunity and potential dilution. Valuation is another pressure point. Bank of America's price target corresponds to roughly 33x fiscal 2028 earnings, significantly above the 26x historical median it cites. This high target therefore implies investors must continue granting Marvell a premium. Slower growth could simultaneously hurt both earnings expectations and the valuation multiple applied to those earnings. Therefore, when evaluating the October 6 Analyst Day, key focus areas should include delivery timelines, margin expectations, and the ramp-up progress of customer projects. After such a substantial rally, merely expanding market size projections provides limited confidence support. The strongest confirmation signal at that point would be clearer evidence that sales growth translates into sustainable actual EPS and cash flow per share.