The market is not only technically overbought at this time, but arguably on shaky fundamental ground. High inflation is slowly chipping away at the economy, and the steep valuations of artificial intelligence stocks that performed so well when the AI revolution was still young are now being questioned.
Nevertheless, compelling growth stocks exist for investors willing to look past current market noise. These aren’t companies riding temporary hype cycles. They’re positioned to benefit from secular trends that will unfold over the next decade and beyond.
Secular growth trends emerge when structural, long-term forces reshape entire industries. Unlike cyclical trends that peak and reverse, secular trends persist through multiple economic cycles because they reflect fundamental changes in how people work, consume energy, or solve problems. Technology adoption cycles follow predictable patterns—early skepticism, followed by accelerating adoption as costs decline and benefits become undeniable, eventually reaching mainstream acceptance.
Three growth stocks exemplify this secular opportunity. Each trades well below recent highs, creating entry points for investors with conviction in multi-decade growth narratives. One operates in hydrogen fuel cells finally reaching mainstream adoption after 29 years of patient development. Another provides enterprise automation software that AI threatens to disrupt but customers continue renewing at 97%+ rates. A third manufactures specialized chips for artificial intelligence infrastructure at a time when demand is accelerating and diversifying beyond a single provider.
Plug Power Inc. (PLUG) trades around $2.63 and represents the early stages of hydrogen technology reaching mainstream adoption. For the entirety of Plug Power’s 29-year existence, it’s been unprofitable, and investors have repeatedly questioned why the company persisted with hydrogen fuel cells despite decades of losses.
Now we know. This technology is finally moving into the mainstream, offering the company a realistic path to profitability.
Hydrogen fuel cells work by splitting hydrogen molecules into charged protons and electrons that generate electricity. These cells can power anything from small vehicles to buildings, including data centers. For most of this technology’s existence, its stumbling block was lack of acceptance stemming from lack of understanding, plus the fact that pure hydrogen wasn’t cheap or easy to procure.
That’s changing. Plug Power is addressing the hydrogen industry’s biggest impediments head-on. The company now manufactures and markets electrolyzers that split ordinary water into oxygen and hydrogen. It also sells hydrogen directly and sells electricity produced by its own equipment. This diversification creates multiple revenue streams and reduces dependence on any single application.
This approach is working. Although Plug Power won’t reach profitability in the immediate future, last year’s net loss was 20% less than 2024’s loss even as 2025 revenue grew 13%. Credit higher-margin profit centers like power purchase agreements and raw hydrogen sales, which are expanding to comprise more of total revenue. At its current rate, the company expects to swing to profit by late 2028.
The underlying secular tailwind is substantial. Precedence Research predicts that the global hydrogen business will double in size by 2035, while the fuel cell market itself could grow at an average annual pace of 25% in the same timeframe. This isn’t speculative technology—it’s secular growth backed by industrial demand and government policy supporting decarbonization.
Hydrogen serves industries that can’t easily electrify. Heavy trucking, aviation, industrial heat applications, and data center power all require energy-dense solutions that batteries can’t provide. Hydrogen fills that gap, creating a decades-long growth opportunity as the world transitions away from fossil fuels.
ServiceNow Inc. (NOW) trades around $111 and has been halved over the past year amid concerns that artificial intelligence threatens its business model. The logic seems straightforward: if anyone can use AI to create automation solutions for free, why pay ServiceNow for coded applications?
But this analysis misses crucial reality. Many AI-generated solutions don’t provide the same reliability or functionality as ServiceNow’s applications, which were engineered from the ground up to excel at particular tasks. Enterprise customers have discovered this distinction through direct experience.
The irony is that most investors express skepticism about AI-powered coding agents, yet most of ServiceNow’s paying customers clearly disagree. First-quarter non-GAAP revenue of almost $3.7 billion was up 19% year-over-year, with the bulk coming from subscriptions boasting renewal rates regularly at or above 97% through Q1.
Ninety-seven percent renewal rates in enterprise software represent exceptional stickiness. Customers aren’t replacing ServiceNow with free AI tools—they’re renewing and expanding usage. This suggests that ServiceNow’s value proposition extends far beyond basic code generation. Customers derive value from integration, reliability, security, and purpose-built functionality that generic AI tools can’t replicate.
The company is looking for similar results through the remainder of the year. Given that ServiceNow is regularly rated as a leader in Gartner’s rankings of enterprise application developers, this double-digit growth pace could easily persist well into the future. The secular trend toward workflow automation and enterprise digitization remains intact regardless of AI sentiment shifts.
Analyst sentiment supports this view. Although investors as a whole clearly aren’t hopeful, the vast majority of analysts covering this stock currently rate it as a strong buy, with a consensus target of $140.38—more than 30% above current prices. This divergence between investor pessimism and analyst conviction suggests the market has overshot on the downside.
ServiceNow’s secular growth driver is straightforward: enterprises across industries must digitize workflows to compete. The software enables that transformation by automating and streamlining business processes. This need exists regardless of whether individual customers consider AI existential threats to specific vendor categories.
Marvell Technology Inc. (MRVL) trades around $235 and has declined 20% from late June peaks despite benefiting from the artificial intelligence infrastructure buildout. The company manufactures computing hardware largely for data centers, including switches, Ethernet controllers, digital signal processors, storage interfaces, and increasingly, computing processors.
Marvell competes with Broadcom on the networking front and Nvidia in AI compute, but the market is slowly opening to other options. Due to a combination of costs and need for more specific solutions, newcomers are winning contracts.
A particularly relevant development: after developing high-performance processors for internal use, e-commerce giant Amazon is now entertaining selling these Trainium chips to third-party customers outside of its AI data center ecosystem. This represents a significant market expansion. Precedence Research believes the custom AI chip market will grow at an average annual pace of 25% through 2035, when it will be worth $550 billion. Marvell helped design Amazon’s Trainium processors, positioning it directly in this emerging market segment.
This exemplifies how Marvell can capitalize on the evolving AI market. The company isn’t dependent on a single AI chip design winning. Instead, Marvell provides infrastructure components that support diverse AI architectures—networking, storage, processing. As AI infrastructure grows more sophisticated and specialized, companies need the exact capabilities Marvell provides.
Investors won’t necessarily need to wait for specialty design work to start paying off. Last year’s top line improved by more than 40%, pushing the company out of the red and well into the black. Analysts are looking for similar revenue growth this year and next, more than doubling per-share profits in the process.
Despite pessimistic rhetoric surrounding AI practical utility, nobody seems to think demand for more AI infrastructure will slow down. The secular tailwind supporting semiconductor demand for AI infrastructure persists regardless of short-term sentiment shifts. As enterprises deploy AI across increasingly diverse applications, demand for the infrastructure enabling those deployments—switches, processors, storage interfaces—continues accelerating.
