A concentrated bet on AI infrastructure companies has dramatically outperformed the broader market, with a portfolio of eight stocks gaining 30.2% between January and April 2026, while the S&P 500 declined slightly. The outperformance wasn't random. The three companies with the strongest alternative data signals at the start of the period, Nebius, Vertiv, and Marvell Technology, delivered the largest gains, each rising more than 67% .

This pattern reveals something important about how AI infrastructure demand is unfolding. The companies that hired most aggressively, posted the most job openings in specialized roles, and showed the strongest employee sentiment were the ones that subsequently announced major deals and expanded capacity. In other words, the hiring signals preceded the public announcements by months, suggesting that alternative data sources can identify infrastructure winners before traditional financial metrics catch up .

Which AI Infrastructure Companies Showed the Strongest Signals?

Nebius, a cloud infrastructure provider, gained 74.1% over the three-month period. The company's LinkedIn headcount had grown 58% in the prior year, concentrated in infrastructure and compute capacity roles. That aggressive hiring pattern was visible in the data before the company announced a $27 billion infrastructure deal with Meta Platforms, a $2 billion investment from NVIDIA, and nine new data center sites across the United States and Europe .

Vertiv, which specializes in cooling systems and thermal management for data centers, gained 69.5%. In January, the company had posted 31% more job openings than the prior year, with most positions focused on liquid cooling and thermal management. That hiring surge preceded the announcement of a 22.7% year-over-year revenue increase in Q4, with orders up 252% year over year. Vertiv also joined the S&P 500 in March and continued expanding its U.S. manufacturing footprint .

Marvell Technology gained 67.4%, driven by job postings in optical interconnects and silicon photonics architecture that signaled an active product cycle before it became visible to the market. NVIDIA subsequently made a $2 billion equity investment in Marvell in late March as part of the NVLink Fusion partnership, validating the positioning that the alternative data had already identified .

How to Identify AI Infrastructure Opportunities Before They Hit Headlines?

• Monitor Hiring Trends: Track job posting increases and headcount growth on platforms like LinkedIn, especially in specialized roles like liquid cooling engineers, silicon photonics architects, and infrastructure capacity planners. Companies hiring aggressively in these areas are often building ahead of announced demand.
• Watch Employee Sentiment Metrics: Employee business outlook scores and internal sentiment readings can signal confidence at companies scaling rapidly. High employee ratings often precede public announcements of major deals or capacity expansions.
• Analyze Web Traffic and Community Mentions: Changes in monthly website visits and mentions on forums like Reddit can indicate growing awareness and retail interest before analyst consensus catches up. A 275% spike in Reddit mentions, for example, may signal emerging investor attention.
• Track Analyst Price Targets: Compare current stock prices to average analyst targets. Companies with 24% to 30% upside implied by analyst consensus may still have room to run if the underlying business fundamentals remain intact.

The broader portfolio also included more stable, structural names that delivered positive returns ranging from 2.1% to 11.0%. Taiwan Semiconductor gained 11.0%, Arista Networks gained 8.8%, Amazon gained 4.7%, NVIDIA gained 4.7%, and Constellation Energy gained 2.1%. These companies benefited from sustained AI buildout over time rather than single catalyst events .

What's the Outlook for These Companies Going Forward?

Nebius continues to show strong signals. Employee business outlook reads 91 out of 100, an all-time high for the company, reflecting internal confidence as Nebius scales faster than almost any other public AI infrastructure company. Web traffic is running at 573,000 monthly visits, up substantially from earlier in the year. Goldman Sachs has a price target of $205, roughly 24% above current levels .

Marvell Technology has 449 active job postings and a business outlook score of 87, both above-average signals for a semiconductor company at this stage of the cycle. The NVIDIA partnership through NVLink Fusion is not a one-time event but rather embeds Marvell deeper into the AI interconnect stack for the foreseeable future. Analyst rating is 93% buy, with average price targets implying roughly 28% additional upside. Google is reportedly in talks with Marvell about custom TPU (Tensor Processing Unit) silicon, which would add a second hyperscaler relationship to the existing AWS Trainium engagement .

Vertiv reported 1,814 active job postings and 17,430 employees according to LinkedIn, with both numbers still increasing and pointing to a business in active expansion mode. The company acquired BMarko Structures to expand capacity for manufactured and converged infrastructure solutions. Analyst rating is 100% buy across 19 analysts tracked. The company reports Q1 2026 earnings on April 22, which will be an early read on whether the 252% order growth from Q4 continues .

Constellation Energy presents a more complicated picture. Job postings declined roughly 22% in late March, and the stock remains below its January levels despite recovering from a sharp March selloff. However, the structural case for nuclear power as the preferred energy source for AI data centers has not changed. Constellation Energy signed a 380-megawatt deal with CyrusOne in March adjacent to its Freestone Energy Center. The average analyst target of $383 implies roughly 30% upside from current levels, though the question remains whether the Three Mile Island reconnection delay resolves on the current timeline .

NVIDIA, the portfolio's most debated name, has traded sideways in a year where it might reasonably have been expected to continue its 2025 run. Export controls, custom silicon competition from hyperscalers, and IT sector selling pressure have all weighed on sentiment. However, the underlying demand signals remain intact. Jensen Huang cited $1 trillion in Blackwell and Vera Rubin orders through 2027 at GTC in March, suggesting sustained demand for the company's flagship AI chips .

The key takeaway is that alternative data signals, when properly interpreted, can identify infrastructure winners months before traditional financial metrics reveal the story. Companies that are hiring aggressively in specialized roles, showing strong employee sentiment, and generating growing web traffic are often the ones about to announce major deals or capacity expansions. For investors tracking AI infrastructure, these signals may provide an edge in identifying opportunities before they become consensus trades.