Most investors looking for top AI stocks automatically gravitate toward Nvidia (NASDAQ:NVDA), Microsoft (NASDAQ:MSFT), Broadcom (NASDAQ:AVGO), and perhaps few others. Yet some of the most attractive AI opportunities may lie among lesser-known companies supplying the infrastructure, connectivity, and computing capacity powering the AI revolution.
The following five companies are not household names, but they are increasingly appearing on institutional investors’ watchlists as potential best AI stock candidates for the next phase of the AI buildout. While all AI stocks carry risk, these businesses combine strong growth prospects with exposure to structural AI spending trends. The list also offers diversification beyond the crowded mega-cap trade.
1. Astera Labs (NASDAQ:ALAB)
What does it do?
Astera Labs develops semiconductor connectivity solutions that help AI accelerators, CPUs, networking chips, and memory systems communicate efficiently within AI datacentres.
As AI clusters become larger and more complex, Astera’s products are becoming increasingly important. Analysts expect the company’s addressable market to expand rapidly through the remainder of the decade.
Investment snapshot
| Metric | Value |
| Market focus | AI connectivity infrastructure |
| 1-year share price performance | Approx. +100%+ |
| 5-year performance | N/A (recent listing) |
| 2026 revenue growth forecast | ~80-100% |
| Forward PE | ~70x |
| PEG ratio | ~1.0-1.3x |
| Balance sheet | Net cash |
Why it could be a best AI stock
Many investors focus on GPUs while overlooking the networking bottlenecks inside AI systems. Astera benefits regardless of whether Nvidia, Advanced Micro Devices (NASDAQ), or custom AI chips win market share.
2. Nebius (NASDAQ:NBIS)
What does it do?
Nebius operates AI cloud infrastructure and large-scale GPU clusters for enterprises and AI developers.
The company is one of the fastest-growing businesses anywhere in the AI ecosystem. Analysts expect revenue to increase dramatically as new data-centre capacity comes online.
Investment snapshot
| Metric | Value |
| Market focus | AI cloud infrastructure |
| 1-year share price performance | +500%+ |
| 5-year performance | N/A |
| 2026 revenue growth forecast | 500%+ |
| Forward P/S | ~15-20x |
| P/E | Not meaningful |
| Net cash position | Strong |
Why it could be among the top AI stocks
Few companies outside hyperscalers are growing as quickly. The challenge is valuation risk, but if management delivers on guidance, Nebius could become one of the defining AI infrastructure winners of the decade.
3. Celestica (NYSE:CLS)
What does it do?
Historically viewed as an electronics manufacturer, Celestica has transformed itself into a supplier of networking, cloud, and hyperscale datacentre infrastructure.
Many investors still do not classify it among AI stocks, creating a potentially attractive valuation opportunity.
Investment snapshot
| Metric | Value |
| Market focus | Data-centre networking |
| 1-year share price performance | +150%+ |
| 5-year share price performance | +1,000%+ |
| 2026 earnings growth forecast | 25-35% |
| Forward P/E | ~22x |
| PEG ratio | ~0.8x |
| Free cash flow | Strong |
Why it could be a best AI stock
Compared with many AI stocks, Celestica still trades on a relatively modest earnings multiple despite substantial exposure to hyperscale AI infrastructure spending.
4. Marvell Technology (NASDAQ:MRVL)
What does it do?
Marvell designs custom AI chips, optical networking components, and data-centre semiconductors.
The company has emerged as a major beneficiary of hyperscalers developing custom AI processors rather than relying exclusively on Nvidia GPUs. Management expects its custom-chip business to exceed $10 billion by fiscal 2029.
Investment snapshot
| Metric | Value |
| Market focus | Custom AI silicon |
| 1-year share price performance | +70%+ |
| 5-year share price performance | +500%+ |
| Revenue growth forecast | 25-35% |
| Forward P/E | ~35x |
| PEG ratio | ~0.9x |
| Gross margin | ~60% |
Why it remains one of the top AI stocks
The market is only beginning to appreciate the size of the custom-silicon opportunity. If hyperscalers continue developing proprietary AI chips, Marvell could enjoy years of growth.
5. STMicroelectronics (NYSE:STM)
What does it do?
STMicroelectronics is best known for industrial and automotive semiconductors, but its AI infrastructure exposure is growing rapidly.
Management recently raised datacentre revenue targets as AI-related demand accelerated.
Investment snapshot
| Metric | Value |
| Market focus | Power and infrastructure semiconductors |
| 1-year share price performance | +40%+ |
| 5-year share price performance | +150%+ |
| Revenue growth forecast | 15-20% |
| Forward P/E | ~18x |
| Dividend yield | ~1% |
| Net debt | Low |
Why it could be the best AI stock for value investors
Unlike many richly valued AI stocks, STMicroelectronics offers direct exposure to AI infrastructure while trading at a traditional semiconductor valuation.
How they compare
| Company | AI Theme | Forward PE | Forecast Revenue Growth | Valuation Appeal |
| Astera Labs | Connectivity | ~70x | Very High | Medium |
| Nebius | AI Cloud | N/A | Exceptional | Speculative |
| Celestica | Networking Infrastructure | ~22x | High | High |
| Marvell | Custom AI Chips | ~35x | High | Attractive |
| STMicroelectronics | AI Infrastructure Components | ~18x | Moderate | Very Attractive |
Investor verdict
Some of the most attractive AI opportunities may lie among lesser-known companies. Investors seeking top AI stocks beyond the obvious mega-cap winners, Celestica and Marvell currently offer the strongest blend of growth and valuation.
For maximum upside potential, Nebius may be the most explosive best AI stock candidate, although execution risk is significant.
Those investors seeking a more balanced approach, Astera Labs and STMicroelectronics provide differentiated exposure to the AI buildout at different points on the risk spectrum.
The key takeaway is that the next generation of winning AI stocks may not be the companies building AI models, but rather the businesses supplying the infrastructure, networking, power management, and connectivity needed to make AI work at scale. As AI spending continues to expand through the current AI investment supercycle, these under-the-radar names could join the ranks of tomorrow’s top AI stocks.
You might also like:
