Will the AI boom in semiconductor stocks continue?
A critical question for the rally is whether the “base effect” will trip up chip stocks.
Key takeaways
- Semiconductor stocks have surged, driven by a historic increase in capital spending on AI infrastructure, like data centers and factories.
- While AI spending on chips is expected to continue rising, the risk is that the pace of growth slows.
- Morningstar analysts don’t expect demand for AI products and technology to fade away any time soon, though the rate of new investment in AI could decelerate.
It’s shaping up to be another blockbuster year for semiconductor stocks, thanks to the spending boom on artificial intelligence technologies. After enduring deeper losses than the broader market in April, the sector roared back to life this summer, with the Morningstar Global Semiconductors Index up34%in 2025 so far. That’s more than double the return of the entire US market. Some stocks have seen even more dramatic gains, with Broadcom AVGO up 47% this year. Industry leader Nvidia NVDA is up 33%.
But are investors getting out over their skis in a volatile industry that could be vulnerable to a pullback? Lurking under the surface of the big bull market in semi stocks are worries about whether growth projections for AI demand are overly optimistic and inflating valuations.
While there is little fear that the money pouring into AI will dry up, there’s another risk for investors to consider. Could the pace of growth merely slow below what investors currently expect, leading to a steep fall in semiconductor stocks as their earning potential is marked down?
AI demand drives semiconductor stock growth
Semiconductors are computer chips that power a wide range of electronics, from cellphones to cars to appliances. The massive rally over the past two years has come because cutting-edge chips are essential components for AI processes, which are highly data-intensive and demand an enormous amount of power and resources.
Across the industry, companies occupy different roles in the production chain. Some, like Taiwan Semiconductor TSM, manufacture the chips. Others, like Nvidia, Broadcom, and Advanced Micro Devices AMD, are primarily chip designers that outsource their manufacturing to other firms. These design companies have seen some of the best growth in recent years, according to Morningstar analysts. The seemingly insatiable appetite for AI technology over the past two and a half years has fueled a boom in capital spending by semiconductor firms and their customers.
Central to this demand are companies known as hyperscalers, which provide the cloud computing power, data centers, storage capacity, and other infrastructure tools needed for AI applications. In a new industry report, Morningstar analysts project the annual collective investment in AI by hyperscalers—Microsoft MSFT, Alphabet GOOG/GOOGL, Amazon AMZN, Meta Platforms META and Oracle ORCL—to triple to $450 billion by 2027 from $150 billion in 2023.
Private markets are getting in on the action too. Data analyzed by Apollo chief economist Torsten Slok shows that AI and machine learning are now the primary drivers of all venture capital investment in North America.
The summer surge
After huge gains on many semiconductor stocks in 2023 and 2024, a suddenly uncertain outlook for the economy and the threat of tariffs on chips pushed many of these names into bear market territory (losses greater than 20%) this spring. But the past few months have been a different story. The top holdings in the Global Semiconductors Index—Nvidia, Broadcom, TSMC, and AMD—have all returned 30% or more this year.
A recent blockbuster earnings report showing a 63% annual increase in AI chip revenue sent shares of Broadcom sharply higher over the past few weeks, and the firm’s year-to-date return is now approaching 50%.
Shares of Oracle surged 30% this month after its quarterly earnings report showed enormous demand for data centers and cloud AI services from hyperscaler customers. Morningstar raised its fair value estimate for Oracle from $205 per share to $330.
Meanwhile, shares of Nvidia and Intel INTC shot higher last week after Nvidia announced a $5 billion investment in Intel, which has lagged far behind in the race for AI dominance, to jointly develop data centers and chips. The announcement of a $100 billion partnership between Nvidia and ChatGPT maker OpenAI sent Nvidia stock higher still.
AI will continue to drive growth
The gains in these stocks are rooted in investor optimism that the massive investment in the chips needed to run AI technologies will continue. For now, Morningstar analysts say the train shows no signs of slowing. “We’re still bullish on AI and semiconductors in general, led by Nvidia,” says senior equity analyst Brian Colello, as “AI demand still seems to be exceeding supply.” At the same time, he adds, “we’ve seen no signs of a slowdown in hyperscaler spending. Oracle’s and Meta’s plans, in particular, have been eye-popping.”
Morningstar analysts expect AI chip revenue to increase roughly fourfold over the next few years, making the new technology the largest growth driver in the industry. “AI is set to create an entirely new data center ecosystem, creating massive demand tailwinds across the technology sector,” they write.
Morningstar analysts describe cloud hyperscalers like Microsoft and Meta as the “backbone of today’s AI boom.” Capex growth here is “unprecedented and show[s] no signs of abating any time soon.” Across the industry, they’re expecting revenue from AI chips to grow at a 40% compound annual rate through 2028, with Nvidia as the clear market leader already deriving the majority of its value from AI. Colello calls that an “outstanding” pace for an industry that would otherwise see annual growth of mid-to-high single digits.
Broadcom will be a “strong second player” after Nvidia, the analysts write. They expect AI to account for 40%-50% of revenue for firms like Broadcom, AMD, and Marvell MRVL. They also expect a broad range of semiconductor firms to benefit from AI-related demand and see “exceptional growth” in the years ahead. They point to TSMC, Cadence CDNS, Synopsys SNPS, ARM Holdings ARM, ASM International ASMXF, and ASML ASML as major beneficiaries.
Base effects may come into play
One potential snag in the growth prospects of these stocks comes from what Wall Street calls a base effect. Spending on AI could grow so much that it can’t increase at the same rate. If that happens, growth in investments in percentage terms would likely fall.
Looking at the multi-year horizon, Morningstar equity analyst Phelix Lee says semiconductor manufacturers’ growth in AI spending could begin to slow. “AI capex is already in hundreds of billions, so its sheer scale makes it less possible to grow at the same breakneck pace,” he says. “This means earnings of AI stocks are more likely to slow, which isn’t good news for stock prices.” He cautions that he does not expect that growth to decline over time, but rather increase at a reduced rate.
“We think the market is too upbeat on long-term AI investment growth after cloud service providers hiked capital spending for 2025, and on the profitability of Chinese players that make advanced AI chips,” Lee wrote in a recent report covering semiconductor manufacturers. Among those companies, he likes TSMC and GlobalWafers 6488 for their strong margins and large exposure to US markets.
The base effect phenomenon is also often discussed around Nvidia’s earnings, which reliably blow by analysts’ expectations. But as the company gets bigger and more robust, the scale of growth becomes smaller relative to its size. Still, Colello says it’s more important to the outlook for these stocks that the dollars invested in AI keep growing, even if the rate of growth slows. “All signals suggest this is the case,” he says. He points to bullish spending plans from the likes of Meta, Oracle, and OpenAI.
AI and the cyclicality of chip stocks
It might not appear that way right now, given the exponential growth of the past few years, but semiconductors are a cyclical industry. “There is a boom/bust cycle that lasts approximately four years, which coincides with the time it typically takes to respond to a shortage (plan, build, and ramp up new capacity), followed by inevitable changes in demand and potential overbuilding,” our analysts explain.
AI is changing the calculus a little. Analysts expect to see underlying growth of about 7.1%-9.6% per year continue through 2030, thanks in part to demand for the new technology. Under the hood, the AI boom has created exponentially higher growth for certain products and markets, like AI-optimized servers.
But overall, analysts don’t expect AI to smooth out the industry’s cycles. They are warning investors of a “bumpy ride” over the next few years, even as growth trends higher. Throughout those cycles, analysts expect firms with wide moats to see less volatility in earnings and share prices.
