Hedge fund Altimeter Capital is pivoting its AI investments, divesting from Nvidia, Micron, and TSMC to place a nearly $489 million bet on cloud infrastructure provider CoreWeave. This strategic shift anticipates future growth in AI compute utilization over hardware manufacturing, but raises questions about CoreWeave’s soaring valuation.
In the feverish race for artificial intelligence supremacy, where titans like Nvidia, Taiwan Semiconductor Manufacturing (TSMC), and Micron Technology have long reigned supreme, a subtle yet significant shift is underway.
The astute eye of Brad Gerstner, the founder and CEO of hedge fund Altimeter Capital, appears to be spotting new horizons.
He is making moves that challenge conventional wisdom and point to an evolving investment thesis in the AI landscape.
Altimeter’s recent portfolio adjustments—trimming Nvidia, and outright divesting from Micron and TSMC—are not merely tactical shifts.
They signal a deeper read on the future of AI infrastructure.
This culminates in a hefty, almost $489 million stake in a lesser-known entity: CoreWeave, hailed as the hottest AI IPO stock of 2025.
Gerstner is no stranger to prescient bets.
His track record includes early investments in data cloud disruptor Snowflake and Asian ride-hailing giant Grab, demonstrating a keen ability to identify nascent trends and back companies poised for exponential growth.
This history lends considerable weight to Altimeter’s latest pivot, prompting a deeper inquiry into the rationale behind shedding established AI giants for a new contender.
On the surface, divesting from Nvidia, Micron, and TSMC seems counterintuitive.
Nvidia commands an overwhelming 90% or more of the data center GPU market, its chips the very backbone of the AI revolution.
Micron specializes in memory storage, an indispensable component for data centers.
TSMC, meanwhile, is the world’s most advanced chip fabricator, the foundry where the designs of Nvidia, AMD, and countless others come to life.
Their collective dominance in the AI hardware supply chain is undeniable.
So, why the cold shoulder from Altimeter?
The answer, as often is the case in high finance, lies in anticipating future headwinds and identifying emerging vulnerabilities.
For Nvidia, despite its formidable lead, risks are accumulating.
Its revenue is heavily concentrated among a handful of cloud hyperscalers like Amazon, Alphabet, and Microsoft.
These tech giants, while current customers, are simultaneously investing heavily in developing their own custom AI chips, a clear signal of their long-term intent to reduce reliance on external providers.
This internal chip development, coupled with Advanced Micro Devices (AMD) steadily gaining ground in the data center arena through key deals with Oracle, Microsoft, and Meta Platforms, suggests that Nvidia’s meteoric growth trajectory might face a deceleration.
Furthermore, the geopolitical shadow cast by U.S. export controls and potential tariffs from a Trump administration on sales to China looms large, threatening a significant chunk of Nvidia’s international business.
Micron’s position, while vital, is also fraught with challenges.
Memory chips, despite their necessity, are largely commoditized.
This makes them susceptible to pricing pressures and cyclical demand.
More critically, the ongoing shift towards cloud-based AI infrastructure could serve as a structural headwind for Micron’s hardware-centric chip memory business.
As AI workloads increasingly move to the cloud, the dynamics of memory consumption and demand could evolve in ways that disadvantage traditional hardware providers.
TSMC, the linchpin of global chip manufacturing, faces a different set of anxieties.
While demand for the chips it produces remains robust, any slowdown in sales from its major customers, particularly Nvidia, would inevitably ripple through its own business.
Beyond market dynamics, TSMC’s operational footprint is overwhelmingly concentrated in Taiwan, a region perpetually under geopolitical strain from China.
This inherent risk could compel U.S. chip designers like AMD and Nvidia to diversify their foundry partners, potentially turning to alternatives such as Intel, thereby eroding TSMC’s near-monopoly.
Enter CoreWeave.
Altimeter’s significant bet on this cloud computing infrastructure provider reveals a strategic shift towards a more nimble, less capital-intensive segment of the AI value chain.
CoreWeave doesn’t design or manufacture chips; instead, it offers clients access to Nvidia GPUs and other chip integrations through its cloud platform.
This positioning sidesteps the immense time and capital required for hardware development and fabrication.
In essence, CoreWeave is a hyperscaler that fills the critical gap between chip production and the optimized AI cloud infrastructure needed for training and inferencing AI models.
It capitalizes on the booming demand for AI compute power without being burdened by the cyclicality and geopolitical risks inherent in hardware manufacturing.
This model allows CoreWeave to scale with remarkable agility, positioning it as a key enabler of the AI revolution rather than a direct supplier of its foundational components.
While Nvidia, Micron, and TSMC might be reaching a phase of maturity in their respective markets, CoreWeave’s business model appears to be in the nascent stages of exponential expansion, capturing value from the utilization of AI hardware rather than its creation.
However, the journalist’s role demands a balanced perspective.
While CoreWeave’s business proposition is compelling, its valuation demands scrutiny.
Since its initial public offering earlier this year, CoreWeave has experienced a notable valuation expansion, with its price-to-sales (P/S) ratio soaring.
This rapid ascent, driven by what appears to be outsized market momentum, raises a cautionary flag.
CoreWeave’s P/S multiple is nearly four times that of Oracle, a company that also provides core data center infrastructure services but is a mature, profitable enterprise.
CoreWeave, by contrast, is a high-cash-burn operation, typical of rapidly scaling startups but a stark contrast to Oracle’s financial stability.
While the fundamental thesis behind CoreWeave’s value proposition in the AI landscape is sound, the current market enthusiasm might have pushed its stock into overbought territory.
Investing at its current valuation carries the risk of becoming an “unsuspecting bag holder” should the momentum wane.
For now, a prudent investor might do well to observe CoreWeave from the sidelines, closely monitoring its growth trajectory and operational efficiency, rather than chasing its current elevated price.
Altimeter Capital’s bold re-calibration of its AI portfolio is more than just an investment decision; it’s a commentary on the evolving dynamics of the AI industry.
It underscores a belief that while the architects and manufacturers of AI hardware remain crucial, the true exponential growth might now reside in the agile, infrastructure-as-a-service providers who bridge the gap between silicon and intelligence, allowing the AI revolution to truly take flight.
It’s a reminder that even in established bull markets, the smart money is always looking for the next wave, the next inflection point, and the next, perhaps less obvious, opportunity.
The expansion introduces agentic decision flows to bolster investigative capabilities for a national security agency in the EMEA region.
Zoho founder Sridhar Vembu warns that companies must prioritize data sovereignty by building internal AI models to protect sensitive intellectual property.
Cities are deploying artificial intelligence on sanitation vehicles to monitor private property, raising complex questions about the future of urban privacy and municipal oversight.
