Dom Rizzo runs T. Rowe Price’s Global Technology Fund, and the numbers he brings to the table force even skeptics to sit up. His fund has compounded at more than 40% a year over the past three and a half years, beating its benchmark by more than 500 basis points, largely on a fat overweight in semiconductors. On CNBC this week, he made a simple claim. Even after the run, the chips are still where you want to be.
The dispersion trade nobody wants to talk about
Rizzo’s opening frame is a widening gap inside tech itself. “We’ve seen semiconductors up 100% and software down roughly 14%,” he said, and the tape backs him up. Advanced Micro Devices (NASDAQ:AMD | AMD Price Prediction) is up 142% year to date, Intel (NASDAQ:INTC) has run 222%, and Micron has done 227%. His fund is heavy in CPUs, owning both AMD and Intel, so this is not an academic observation for him.
The dispersion matters because most retail portfolios are still calibrated to a 2023 world where software ate everything. Rizzo’s argument is that the compute layer flipped from commodity to bottleneck, and the market has repriced accordingly. When the fund manager beating the tape by 500 basis points a year tells you sector selection inside tech is driving returns, you listen.
A hardware cycle rolling into 2027
“I think we’re going to see a major, major hardware cycle heading into next year,” Rizzo said, referencing meetings with more than 20 tech companies at a recent West Coast summit. His read on spend was uniform across memory, CPU, optics, routers, and networking switches. The infrastructure demand, in his words, “almost doesn’t really matter where you look.”
Recent results support the cycle. NVIDIA (NASDAQ:NVDA) posted $1.87 in non-GAAP EPS on $81.61 billion of revenue for its April quarter, with data center up 92% and networking up 199%, per the company’s Q1 FY27 filing.
Broadcom (NASDAQ:AVGO) guided next-quarter AI semiconductor revenue to $16 billion, up more than 200% year over year. Micron’s fiscal Q3 revenue landed at $41.46 billion, up 345.7% year over year, with gross margins expanding to 84.6%. Those are cycle-peak numbers being posted as guidance keeps stepping higher.
Rizzo thinks the Street is still low. He expects roughly 75% capex growth from the big five hyperscalers next year and would not be surprised by north of 80%. The unlock, he argues, is agentic coding and reasoning workloads, which are far more CPU and memory intensive than the training-heavy phase everyone got used to. His proof point is Anthropic, which went from roughly $5 billion in run-rate revenue last summer to close to $50 billion. If that curve is directionally right, the DRAM math alone is difficult to argue with.
The valuation argument that keeps him long
Rizzo’s valuation math cuts against the run-up narrative. “I look at NVIDIA just 14 times earnings here as we head into the Vera Rubin cycle… I still think semis are the place to be even after this big run.” NVIDIA trades at a forward P/E of 23 on today’s numbers, but Rizzo is looking through to Rubin-era earnings power. He also flagged SK Hynix at just 4 times earnings, which frames how the memory complex is priced in Asia versus the enthusiasm around Micron.
The foundry underneath all of this remains Taiwan Semiconductor, where CEO C.C. Wei has guided full-year revenue growth above 30% and where 3nm now accounts for more than half of wafer revenue. That is the physical substrate Rizzo’s thesis rests on.
What to watch next
The near-term risk is optical. Chips sold off Tuesday, with AMD down 5.7%, Micron down 9.74%, and Taiwan Semi down 6.8%, a session that shakes conviction. Rizzo’s frame gives you a way to think through it. If hyperscaler capex compounds at 75% again next year, the multiples on this group are lower a year from now even if the stocks stand still. That is the trade he is making, and his track record earns the benefit of the doubt.
Contact [email protected] for any questions or corrections.
Top Gaining Stocks
Top Losing Stocks