As earnings season once again puts the largest technology companies under the microscope, the conversation around the "Magnificent Seven" has taken on a sharper edge. Five of these mega-cap names are reporting this week, and the question is no longer simply whether they will beat consensus. The deeper question is whether the staggering levels of capital expenditure being poured into artificial intelligence infrastructure will translate into durable revenue, sustained margin, and continued multiple expansion. The picture that emerges is uneven: some names appear poised to deliver, one in particular is clouded by uncertainty, and the entire group is operating in a market that has effectively become an oligopoly defined by the cost of admission.
Amazon: A Solid Bet Anchored by Backlog and Custom Silicon
Among the hyperscalers, Amazon stands out as one of the more attractive setups coming out of this earnings cycle. Projected growth sits at 25.6%, and the company recently struck a $100 billion infrastructure deal with Anthropic that materially extends its AI runway. While there is some skepticism about whether that growth profile is achievable, the skepticism is not unique to Amazon — it applies to all hyperscalers wrestling with the question of AI monetization.
What makes the Amazon case compelling is the underlying support. The company carries roughly $244 billion in contracted backlog, an enormous foundation of locked-in future revenue. That backlog is then complemented by high-margin contributions from its custom silicon: Trainium for AI training workloads and Graviton for general-purpose compute. These chips give Amazon a margin lever that competitors cannot easily replicate, and they reduce dependence on third-party accelerator supply. For investors looking across the hyperscaler complex, this combination of growth, backlog, and proprietary silicon makes Amazon a name with a particularly favorable risk-reward.
This view is reinforced by recent portfolio shifts among prominent growth investors. One well-known manager has reportedly sold a sizable position in AMD — at the low end, around $75 million — and rotated into Amazon as well as X Energy, a signal that some sophisticated capital is rotating out of merchant silicon and into the platforms that consume it.
Microsoft and Alphabet: Underappreciated Cloud Engines
Microsoft and Alphabet deserve close attention, primarily because the market is not yet giving their cloud businesses the credit they merit. Azure now represents roughly a third of Microsoft's revenue and is likely growing in the upper 30s, possibly approaching 40%. That growth rate is the engine that drives Microsoft's overall topline expansion of around 17% — a figure that runs slightly above consensus expectations. The remarkable thing is that investors are not paying much for that profile: the price-to-earnings multiple sits in the low 20s, an unusually modest valuation for a business with this kind of growth and incumbency.
Alphabet presents a similar story, though it carries a somewhat richer multiple. Its search business remains on fire, defying the narrative that generative AI would erode traditional search economics. At the same time, its cloud business is likely growing in the 50s this quarter — an extraordinary figure for a unit of its scale. Both companies are exceptionally well positioned to be primary beneficiaries of the continued generative AI trend, and both offer long-term growth opportunities with relatively low downside given current multiples.
The Capex Question: How Much Is Too Much?
Perhaps the single most important variable hanging over this earnings season is capital expenditure. Aggregate hyperscaler capex is approaching $680 to $700 billion this year, well above the $600 billion mark that already seemed staggering. Microsoft alone is now tracking toward roughly $200 billion in planned capex. Some sell-side voices have begun to flag that this figure may be too high for a single company to absorb gracefully.
The first time one of these hyperscalers — Meta, in particular — printed a capex number that floored the market, it triggered a wave of anxiety. The anxiety abated only when the rest of the group came out with similarly large numbers, normalizing what had felt like an outlier. That sequence has now repeated and intensified.
Yet the framework for evaluating capex is not absolute. When revenue growth is accelerating and capex is accelerating only slightly faster, that remains acceptable, especially if capex can roll over in subsequent years. The complication is that in the current AI environment, sustained high capex looks like the new normal rather than a transient spike. It is the price of admission for being a hyperscaler at all.
That price of admission has a structural consequence: the market has become an oligopoly. The capital intensity required to compete effectively prevents most new entrants from breaking in. There are interesting smaller players — Nebius and CoreWeave, for example — that are doing impressive optimization work, but their most plausible long-term path may be as acquisition targets rather than independent challengers.
The capex story is also being underwritten by a rapidly expanding total addressable market. By 2026, the aggregate addressable market across these businesses is projected to reach roughly $5 trillion, encompassing cloud computing, infrastructure, content advertising, and broader AI-enabled services. With contractual obligations locking in cloud revenue and adjacent businesses providing additional surface area, the runway looks solid across the group.
The Strategic Logic of Connected Software and Infrastructure
A key insight from the current landscape is the strategic value of owning both software and infrastructure. When a company's software is directly connected to its own cloud infrastructure — as is the case with Google, Microsoft, and Amazon — that vertical integration becomes a powerful competitive moat. It ties customers more deeply, captures more economic value across the stack, and creates a flywheel effect where AI workloads run on infrastructure the same company also monetizes. All three of these companies should be on the radar of long-term investors as growth opportunities with relatively contained downside.
Apple: The Question Mark of the Group
Apple is the source of meaningful uncertainty heading into earnings, and that uncertainty is enough to make a basket trade across all Mag Seven names — such as buying a Mag Seven ETF ahead of reports — unattractive. The story under new leadership might eventually become interesting, but the near-term setup has real headwinds.
Smartphone sales are being pressured by memory pricing dynamics, and those dynamics are themselves a knock-on effect of the AI boom. The insatiable demand for memory inside AI infrastructure is driving memory component costs higher across the industry, and Apple, as a major buyer of memory for its devices, is squeezed on the input side of its hardware business. This is one of the more underappreciated cross-currents of the AI cycle: the very trend that is lifting the cloud names is creating cost pressure for the consumer device franchise, even as it sells the same end-customer experience layer. Apple is the one Mag Seven name visibly underperforming the group on the day, and that underperformance reflects this confluence of pressures.
Conclusion: An Oligopoly Worth Owning Selectively
The takeaway from this earnings preview is that the Mag Seven is no longer best understood as a monolithic basket. The hyperscaler trio of Amazon, Microsoft, and Alphabet appears to offer a particularly compelling combination of growth, backlog, integrated software-and-infrastructure positioning, and reasonable multiples. Capex is high and rising, but it is increasingly justified by the size of the addressable opportunity and by the contractual revenue already booked against it. Apple is a separate case, weighed down by memory cost pressures and smartphone demand softness even as the rest of the cohort rides the AI wave. For investors, this is a moment that rewards selectivity over indexing — picking the companies whose capex is unmistakably translating into durable, contracted, high-margin revenue, and being patient with those whose path is less clear.