Marvell Technology Stock And The Forecast It Kept Outgrowing

The chipmaker kept telling investors its own AI targets were too low, but the message took a while to land.

When a stock like Marvell Technology (MRVL) puts up a 189% gain in a year, the natural question is: what did we miss? How could anyone have seen that coming? The honest answer is that you rarely can, at least not with certainty. Markets are messy, and perfect foresight is a myth.

But sometimes, a company’s signals are loud and clear. It broadcasts its trajectory in plain sight. In the months leading up to its surge, Marvell’s management was exceptionally bullish, repeatedly and explicitly telling the market that their own well-publicized targets for the AI boom were already obsolete.

How Early Did They Start Raising The Bar?

The pattern began to emerge clearly on the company’s third-quarter fiscal 2025 earnings call in December 2024. Amid a strong report, management stated they were “clearly set to significantly exceed the full year AI revenue target of $1.5 billion.” This was a direct signal that their own benchmark, set just months earlier, was too low. The previous day, they announced a multi-generational five year agreement with Amazon Web Services covering everything from custom AI products to optical DSPs, cementing a key relationship for the years ahead.

What About The Target For The Following Year?

One strong year is one thing, but markets pay for sustained growth. Three months later, on their March 2025 call, Marvell’s leadership went a step further. They expanded their bullish outlook to include the next one. Management said they expected to “very significantly exceed our $2.5 billion target in fiscal 2026.” They were now telegraphing that the AI demand they were seeing was so strong that it was already outpacing the forecast for the year that hadn’t even started yet. This escalating confidence suggested Marvell’s bet on AI was even bigger than many realized, a dynamic explored in other analyses at the time.

This Wasn’t Just A One-Year Ramp?

The final piece of the puzzle was durability. In both its March and May 2025 calls, the company hammered home the long-term nature of its custom silicon programs. Management was clear, stating they expected revenue from their key custom XPUs to grow in fiscal 2026 and to “continue to grow next year fiscal 2027 and beyond.” By the May call, they confirmed they had already “secured three-nanometer wafer and advanced packaging capacity” for the next generation. They were building a multi-year pipeline, securing the supply for it, and telling anyone who would listen.

Interestingly, the options market seemed to be looking the other way. In the weeks before the run, implied volatility actually eased from the 71st to the 53rd percentile of its annual range, suggesting traders were pricing in less volatility, not more. It’s a useful reminder that even when a story is being laid out, it’s not always the one the market chooses to hear.

This was the sound of a business that couldn’t set targets fast enough.

How Do You Get Ahead Of The Next One?

Not by listening harder; no individual can monitor thousands of companies that closely. The practical edge is to watch the one signal that is public and trackable: a guidance raise. Our Guidance Momentum rankings show which S&P 500 names are flashing it now, with momentum to match. But guidance is one signal among many. And if it is exposure to semiconductor as a whole you want, rather than hunting the next single name to surge, a semiconductor ETF like SOXQ covers that single sector. Going broader than any one sector, to a quality-first mix across the whole market, is the natural next step. The Trefis High Quality (HQ) Portfolio weighs the full picture of quality across thousands of names, holds the 30 strongest, re-balances them with rules, and has outpaced a benchmark that combines the three major indices – the S&P 500, S&P Mid-cap, and Russell 2000.