This running blog discusses idea/positioning and opinion on the largest market cap companies in US market which hv outsized weightage on the Index due to SPX being market cap weighted index
TLDR – 24/11/2025
My base case: As long as Google is perceived to be the clear AI winner, MSFT / ORCL / NVDA may act as funding shorts to finance long Google exposure.
PT for Google: $350–$400.
META becomes a buy at certain levels. Flow out of NVDA would look for places to go long. Could be low beta defensives stocks (healthcare/utilities/staples/energy) or other part of growth such as google/china tech)
Apple
Perceived as a safe haven and has arguably become cheaper on a price/cash-flow basis due to almost no significant AI capex. Their long-term strategy may be to wait and license cheaper, best-in-class models (from Google or whoever wins) rather than spend heavily now. I am neutral.
Search champ transitioning into AI champ — now perceived as the winner in AI, pivoting from “loser” status as Gemini 3 gained strong traction.
Search box business is stable and cash-flow rich, with relatively low AI capex needs. Buzz is that most of their models run on in-house TPUs (co-developed with Broadcom), reducing dependence on NVIDIA.
Google Cloud (CSP) is firing on all cylinders. Capex remains under control since search profits subsidize compute.
Long-term call options:
- Waymo (self-driving)
- TPU design (much cheaper than GPUs)
- Anthropic stake (application-layer exposure)
Amazon
Latest quarter was solid — retail activity strong and AWS finally accelerating (20% YoY). Neither cheap nor expensive; good long-term buy. Founder selling puts a cap on immediate upside momentum.
Anthropic stake gaining traction inside enterprises. Balanced exposure across Cloud (AWS), commerce (retail/e-com), and AI. Roughly 30× P/E normalized for a ~10% grower. Mildly bullish.
NVIDIA (GPU King)
Currently too concentrated among a few customers. Market perceives risk in OpenAI-linked circular deals and unclear ROI on the next GPU refresh cycle.
Debate intensifying: cheap ASICs vs expensive GPUs, especially with Google running major models on TPUs. Inference may dominate training workloads going forward.
Market sees echoes of Cisco & Sun Microsystems from the dot-com era.
Stock isn’t expensive on classic metrics, but uncertainty remains around GPU pricing/inventory once infra layer companies (CSPs) slow down.
Networking business also interesting. Strong moat from the CUDA software layer.
Broadcom (AVGO)
Google’s partner in TPU design, attracting flows rotating out of NVDA. Strong in data-center networking gear and now software-heavy post-VMware acquisition.
Low beta; dips get bought quickly. I’m neutral — valuation feels expensive.
Microsoft
Price action bearish due to the OpenAI stake, uncertainty around massive AI capex requirements, and unclear CSP (Azure) depreciation cycles. No longer viewed as the AI winner as OpenAI is seen losing to Google / Chinese players. Expensive, and forward path not as obvious.
Could become a funding short to go long Google.
Meta
Strange case: core social media business strong with healthy growth, but Zuckerberg’s heavy AI capex ambitions are hurting sentiment. Latest models haven’t gained traction; CEO appears committed to spending aggressively.
High-beta stock. Selloff driven by governance concerns (non-voting shares) and unlimited CEO control. Very dependent on cash-flow discipline.
Buyer in the $550–600 range for a half position. If $550 doesn’t hold, $380–400 is possible.
Oracle
Business transformed from software/SaaS into a CSP with high topline growth but low margins due to acquisitions. Fate strongly tied to OpenAI. Acts as a proxy trade for long/short OAI. Balance sheet stretched after years of datacenter capex.
No interest in going long except for tactical swings.
Alibaba
The only large-cap Chinese tech stock generating real investor excitement. Retail e-com is soft but Qwen AI models are gaining traction.
Main risk: US sanctions over potential ties to CCP/military development — similar to what happened to Huawei.
Mid-cap Software (IGV)
Perceived as AI losers. Adobe/CRM cheap vs their own history but still expensive on absolute terms. Several names worth scouting for hidden gems:
OKTA, RBRK, BRZE, CRM, ADBE, WIX, MDB, DDOG, ZM, NOW.
Zoom is cheap and has earnings today. I hold several, but OKTA is my top pick here.
Tesla
Extremely expensive for a car company. Trades at a massive premium based on Elon’s status as a tech CEO/influencer/disruptor. Narrative changes monthly — humanoid robots, AI chips, FSD, AI company, etc.
Great trading stock if you catch the swings. Long-term alignment between Elon’s priorities and shareholder interests is unclear. Very divisive with a massive fan base.
Palantir (PLTR)
Expensive, and few understand the actual business. Essentially premium IT consulting wrapped in a platform. Valuation hard to justify. Good trading stock if it recovers moving averages, but long-term I’m skeptical.
IT Services (ACN, INFY, CTSH, WIT, Capgemini)
Sector has become cheap. Market thinks AI will eat them alive, but I’m not convinced. Could see flows if money rotates out of NVDA.
Eli Lilly (LLY)
Money rotating from Tech (NVDA/MSFT) into Healthcare (XLV), and LLY is the biggest component so it gets automatic bid support. Their newest weight-loss drug retatrutide in Phase 3 is extremely exciting.
Restated Base Case
As long as Google is perceived as the AI winner, MSFT / ORCL / NVDA are likely funding shorts, and Google remains the primary long.
PT = $350–$400.
Leave a Reply