Marvell Shares Slide as Google AI Deal Payoff Moves Out

Marvell Technology’s shares fell more than 8% after a strong results cycle because investors wanted the Google custom-chip deal to translate into revenue sooner. The reaction exposes the harder phase of the AI chip boom: converting major design wins into revenue quickly enough to satisfy elevated expectations.

Strong Forecasts Could Not Outrun the Timeline

In its August 28 report, Reuters said Marvell shares dropped to $221.60 in early trading as investors looked past solid results for evidence of longer-term growth. The company’s shares had nearly tripled during the year, while Big Tech’s AI spending was expected to top $740 billion. That backdrop raised the bar: a beat is no longer enough when investors are underwriting several years of accelerating AI demand.

Marvell’s own outlook remained constructive. Reuters reported that the company expects revenue to grow about 45% in fiscal 2027 and reach roughly $18 billion in fiscal 2028. The forecasts raise a harder question: how much growth is priced in, and how much depends on programs not yet in production?

The Google Deal Is Valuable but Back-Loaded

The pressure centres on Marvell’s relationship with Google. In a separate Reuters report on the custom-chip agreement, the deal was described as potentially generating up to $120 billion in revenue through fiscal 2033. Google also received a warrant to buy up to 58.97 million Marvell shares at $206.58 apiece, a potential $12.18 billion stake.

Those figures describe the opportunity, not near-term revenue. CEO Matt Murphy told investors that custom-revenue targets through fiscal 2028 already include some Google-related contribution, but that the relationship becomes much more meaningful in fiscal 2029. That sequencing helps explain the selloff: the commercial prize can be enormous while the earnings impact remains distant enough to challenge short-term valuation assumptions.

Marvell Is Selling More Than a Chip

Marvell’s role covers the processors, storage-management components, and networking technology that work alongside Google’s Tensor Processing Units. Its company platform and product portfolio reflect the broader shift in custom silicon: cloud operators want differentiated systems that can improve cost and performance for specific workloads, especially inference.

That opportunity is strategically important because the custom-chip market is not a simple contest between one merchant GPU and another. Google’s TPU platform represents the cloud buyer’s push toward workload-specific accelerators. NVIDIA’s data-center platform represents the scale and software advantages of a general-purpose accelerated ecosystem, while AMD Instinct adds another competitive path. Marvell’s value depends on becoming an essential design and connectivity layer across that changing mix.

Why the Stock Reaction Matters for AI Infrastructure

Reuters said Friday’s decline could erase more than $17.4 billion from Marvell’s market value if losses held. At least eight brokerages raised their price targets after the results, yet the stock still sold off. That combination is a warning about expectation risk: analysts can become more positive on a company’s long-run position while the market reprices the wait for actual revenue.

Cloud accelerated computing demand remains real, but deployment cycles are lumpy. Customers must qualify designs, secure capacity, integrate software, and ramp production. Independent market analysis similarly focused on the gap between the Google deal’s headline potential and the timing of sales.

The Investment Case Now Requires Patience

The Marvell story is therefore less about whether AI spending exists than about when custom silicon converts into reported revenue. Our coverage of NVIDIA and AWS’s two-million-GPU expansion shows the scale of competing infrastructure commitments, while physical-AI compute points to another source of demand. AMD’s inference-silicon moves and AI data-center financing underline the supply response.

Google’s custom-chip relationship could become one of Marvell’s defining growth engines, but fiscal 2029 is a long way from a Friday earnings reaction. Enterprise AI partnerships can create demand pipelines, yet investors ultimately need production schedules, customer concentration visibility, and margin evidence. Marvell has the design win; the market is waiting for the revenue clock to catch up.