Building Modern Railroads
Platinum Performers and the Industrial-Scale AI Buildout
And of course infrastrucure also includes everything inside: racks of computer equipment, routers/multiplexers, ethernet hubs, cables and connectors, air conditioning and water cooling. You will also frequently see in the press that these companies each have a moat. Which is what we used to call 'barriers to entry'. In a particular industry, the scale of the spend makes it difficult if not impossible for anyone else to partipate, and hence they are unable to compete. It's generally agreed that these are the five we will hear most often about.
There is a useful comparison to the 1920s, but it shouldn't be taken too literally. The industrial expansion in that period was shaped by building and extending the electrical grid and affordable automobiles. Radio allowed market participants to reach a vastly wider audience than telephone or telegraph. Automobiles fuled the growth of suburbs, and less dependence on rail for long trips. Add commerce/advertising/entertainment to the mix. How about live radio broadcasts of the home baseball team with commercials from local car dealers. Use cases in other words.
At the industry sector level, AI has a similar feel in one respect: enabling infrastructure is being built before the full economic impact can be measured. In the AI buildout, the modern “rails” of the railroads are cloud infrastructure, AI accelerators, and developer ecosystems. That is why Alphabet, Amazon, Microsoft, Meta, and Oracle belong in this category with their moats. They are building an environment where AI will be trained, deployed, and monetized. Read on - for some highlights about These Five.
Alphabet Inc., Class C (NASDAQ: GOOG)
The latest earnings cycle reinforces the views we're discussing in this blog. Alphabet reported Q1 2026 revenue of $109.9 billion, up 22% year over year, with diluted EPS of $5.11. Google Cloud revenue increased 63% to $20.0 billion, led by enterprise AI infrastructure and core Google Cloud Platform services. Alphabet also disclosed that net gains on equity securities increased diluted EPS by $2.35, a reminder that reported GAAP earnings need to be read carefully in this quarter’s comparison.
Investor Relations | SEC Filings | Updated Quarterly
Amazon.com, Inc. (NASDAQ: AMZN)
Amazon reported Q1 2026 net sales of $181.5 billion, up 17% year over year, with diluted EPS of $2.78 compared with $1.59 in the prior-year quarter. Amazon also disclosed that first-quarter net income included $16.8 billion of pre-tax gains from investments in Anthropic. The operating story remains broader than a single EPS number: AWS, cloud services, and logistics automation all sit inside Amazon’s larger industrial-scale platform.
Investor Relations | SEC Filings | Updated Quarterly
Microsoft Corporation (NASDAQ: MSFT)
Microsoft reported FY26 Q3 revenue of $82.9 billion, up 18% year over year, and diluted EPS of $4.27. Microsoft’s fiscal year ends June 30, but the March 31 report still gives a direct view into the same AI build-calendar. Microsoft reported non-GAAP diluted EPS of $4.27.
Investor Relations | SEC Filings | Updated Quarterly
Meta Platforms, Inc. (NASDAQ: META)
Meta reported Q1 2026 revenue of $56.3 billion, up 33% year over year, with diluted EPS of $10.44. Meta also disclosed that diluted EPS would have been $3.13 lower without an income tax benefit reported. That adjustment matters, but it does not change the strategic point: Meta continues to direct enormous resources toward AI infrastructure and their advertising platform. All of the Hyperscalers are talking about their CapEx spending, but that's another NEA post/article in entirety, to quantify the trends in real dollars and consider whether the industry can afford the buildout.
Investor Relations | SEC Filings | Updated Quarterly
Oracle Corporation (NASDAQ: ORCL)
Oracle's fiscal year ends May 31, which doesn't fall at the end of a calendar quarter either, but it belongs in the Platinum discussion because its cloud infrastructure business has become increasingly tied to AI demand. And they have an enormous embedded base of Java and Oracle Database customers. Billions of lines of code in the aggregate that need to be updated with AI features. Python is popular in the news and with programmers, but not as capable as Java. Oracle reported fiscal 2026 third-quarter revenue of $17.2 billion, with cloud revenue of $7.6 billion.
Investor Relations | SEC Filings | Updated Quarterly
This is important: each company is using its own model. Alphabet with search and cloud; Amazon with AWS and logistics; Microsoft with Azure and enterprise software; Meta with advertising and consumer AI; Oracle with database customers and cloud infrastructure. The common thread is capital formation around AI infrastructure. Railroads, engines and boxcars were mainly built in the hundred years prior to the 1920's. It won't take nearly that long to build the AI industrial base.
New Edge Analytics traditionally looks for proxies: smaller or more specialized companies positioned to benefit from a large technology trend. But in this cycle, the largest companies are not merely beneficiaries. They are the builders. Their capital spending defines the direction of the broader industry. The Hyperscalers are the companies assuming the risk.
The 1920s comparison is useful only to the extent that it reminds us how infrastructure revolutions develop. First comes capital investment. Then comes experimentation, more recently called 'platformization'. Then comes consolidation and productivity. AI is still early in that process. The Platinum Performers are important because they are funding the buildout before the final use cases are fully known. For now I'd put a personal bet on Big Pharma and Biotech, and all of the Financial Services industry - Finance, Insurance and Real Estate. Banks will probably be the first to show real dollar benefits from the use of AI.
