The AI infrastructure buildout is a multi-stage process that has evolved from a focus on processing chips to now include networking and optical components.
Hyperscalers are the primary drivers of AI demand, and their increasing development of custom in-house chips poses a significant competitive threat to established chipmakers like NVIDIA.
The memory chip market has fundamentally shifted from a cyclical consumer-driven model to one with long-term visibility and sustained demand due to data center buildouts.
Traditional valuation metrics are being challenged by the AI boom, with some stocks trading at extreme multiples while analysts argue for fundamentally higher valuations for others like Micron.
Major tech companies are making significant strategic shifts to capitalize on AI, including acquisitions (Cisco), job cuts to refocus resources (Cisco), and developing in-house components (Apple).
Initial Phase
Vlastelica's commentary focuses on the aggressive AI-related spending plans announced by hyperscalers like Microsoft, Amazon, Alphabet, and Meta, establishing them as the primary drivers of the AI boom.
Market Divergence
He identifies a significant market shift, noting that the correlation between software and chip stocks turned negative for the first time in history, suggesting a rotation of investor capital and focus.
Broadening Infrastructure Focus
The discourse expands from processing chips to include networking and memory chips as the next critical areas of investment, highlighting companies like Cisco, Micron, and SK Hynix as key beneficiaries.
May
He reports on a period of explosive growth for memory chip stocks, with Micron jumping more than 70% in this month alone, its largest one-month gain since 1987.
Valuation Reassessment
Following massive stock gains, Vlastelica notes that analysts are drastically re-evaluating companies, with UBS, for example, tripling its price target for Micron and suggesting a threefold increase in its fair valuation multiple.
Strategic Maturation
Vlastelica reports on strategic shifts from major tech companies, such as Cisco's job cuts to refocus on AI and Alphabet's plan to sell its custom TPU chips, signaling a new phase of market maturity and competition.
▶The AI Infrastructure Gold RushMay 2026
Vlastelica details a massive, multi-phased buildout of AI infrastructure driven by hyperscalers. This has moved beyond just processing chips to encompass networking components from companies like Cisco and optical products, indicating a broadening of the market.
Investors should look beyond the obvious AI chip plays to second and third-tier beneficiaries in networking and components, as the infrastructure buildout matures and creates new bottlenecks and opportunities.
▶The New Memory Chip ParadigmJul 2026
He argues that the memory chip market has fundamentally shifted from a cyclical, consumer-driven model to one with long-term visibility and sustained demand. This is fueled by data center needs for AI, leading to supply shortages and unprecedented growth for companies like Micron and SK Hynix.
The structural shift in demand suggests that historical valuation models for memory chip companies may no longer be appropriate, justifying the significantly higher multiples and price targets being proposed by analysts.
▶Hyperscaler Dominance and DiversificationMay 2026
Vlastelica emphasizes the central role of hyperscalers (Microsoft, Amazon, Alphabet, Meta) as the primary customers driving AI demand. He also highlights their strategic move to develop custom in-house AI chips, which poses a long-term competitive threat to established players like NVIDIA.
While hyperscaler spending is a massive tailwind for the semiconductor industry, their push for custom silicon creates a complex dynamic where they are simultaneously the biggest customers and future competitors.
▶Valuation vs. Reality in the AI BoomJul 2026
He frequently discusses the tension between soaring stock prices and traditional valuation metrics. While some stocks like Intel are at all-time high valuation multiples, others like Micron are seeing analyst ratings skyrocket, indicating a market struggling to accurately price in the AI revolution.
The market is bifurcated, with some AI-related stocks potentially overvalued while others, particularly in newly critical sub-sectors like memory, may still have room to grow as their fundamental business models transform.