Anurag Rana

AI infrastructure demand is creating a supply crunch for semiconductors and driving memory prices higher; with companies like Google raising capital to invest heavily in AI, the easy trade is to go long semiconductors and AI infrastructure hardware, getting out of software and services. This trade is supported by prices rising for memory chips (Micron up) and the overall chip shortage.

Speaker states Amazon's AI revenue run rate is $5B in Q1, AWS growth is improving due to itself and Anthropic, and capex will increase. Notes differentiation via in-house chips reduces Nvidia costs. Strong AI demand and proprietary chip development translate to improved cloud growth rates and better economics, justifying the market's positive reaction to increased investment. LONG due to accelerating growth in the core AWS business driven by AI adoption and strategic vertical integration. AI revenue growth fails to materialize at projected rates, or capex intensity outweighs profitability gains.

Over the past three years, a massive shift of free cash flow from software to chips and hardware has occurred, driven by cloud providers needing more servers and memory as prices rise, and there are no signs of this trend slowing, benefiting ASML, TSMC, and memory producers.

Over the past three years, a massive shift of free cash flow from software to chips and hardware has occurred, driven by cloud providers needing more servers and memory as prices rise, and there are no signs of this trend slowing, benefiting ASML, TSMC, and memory producers.

AI infrastructure demand is creating a supply crunch for semiconductors and driving memory prices higher; with companies like Google raising capital to invest heavily in AI, the easy trade is to go long semiconductors and AI infrastructure hardware, getting out of software and services. This trade is supported by prices rising for memory chips (Micron up) and the overall chip shortage.

AI infrastructure demand is creating a supply crunch for semiconductors and driving memory prices higher; with companies like Google raising capital to invest heavily in AI, the easy trade is to go long semiconductors and AI infrastructure hardware, getting out of software and services. This trade is supported by prices rising for memory chips (Micron up) and the overall chip shortage.

While broad software and services are under pressure, large-cap software leaders with dominant market share – ServiceNow, Workday, SAP – are still growing and insulated from the AI-driven budget shift. Their commanding positions protect them from the disruption facing smaller firms and consulting companies.

While broad software and services are under pressure, large-cap software leaders with dominant market share – ServiceNow, Workday, SAP – are still growing and insulated from the AI-driven budget shift. Their commanding positions protect them from the disruption facing smaller firms and consulting companies.

While broad software and services are under pressure, large-cap software leaders with dominant market share – ServiceNow, Workday, SAP – are still growing and insulated from the AI-driven budget shift. Their commanding positions protect them from the disruption facing smaller firms and consulting companies.

Salesforce's acquisition of Fin is a smart move both offensively and defensively. Fin provides a generic AI agent that can work across any software platform, filling a gap where Salesforce's own Agentforce was limited to Salesforce's package. This acquisition also blocks standalone AI agent companies like Fin, Intercom, and Sierra from eventually storing customer data and becoming the system of record, which would displace Salesforce's core bread-and-butter. The deal signals Salesforce is serious about integrating AI and accelerates its agent practice.

The selloff in AI infrastructure stocks (NVIDIA, Microsoft, Amazon) due to OpenAI missing internal targets is overdone because these companies have diversified customer bases beyond OpenAI, including hyperscalers with strong demand for AI compute. The dip presents a buying opportunity, except for Oracle which is overly dependent on OpenAI.

The selloff in AI infrastructure stocks (NVIDIA, Microsoft, Amazon) due to OpenAI missing internal targets is overdone because these companies have diversified customer bases beyond OpenAI, including hyperscalers with strong demand for AI compute. The dip presents a buying opportunity, except for Oracle which is overly dependent on OpenAI.

They added $29 billion more of RPO on the balance sheet but they are not going to raise the 2026 capex. By holding capex steady at $50 billion while significantly growing backlog and revenue visibility, Oracle proves it can capture AI market share without destroying its balance sheet or free cash flow. This alleviates the market's primary fear regarding the stock's debt load and cost of capital. LONG. Oracle is successfully transitioning into a top-tier cloud hyperscaler with disciplined capital allocation. High weighted average cost of capital (12%) and negative free cash flow could become problematic if AI demand unexpectedly decelerates.