Google Looks Smart. Can't Say the Same for META, MSFT and ORCL.
META: $30 billion public bond sale and $27 billion private capital deal for AI/LLM-related infrastructure, yet less than $5 billion in LLM-related revenue. META will have egg on its face when this AI trade fully plays out. The best and brightest AI researchers aren’t going to META.
MSFT: Microsoft has taken CapEx from 13% of Revenue in FY’23 to 18% in FY’24 to 23% in FY’25 much of it in support of LLM-related initiatives primarily with OpenAI and to a lesser extent with Anthropic and internally developed LLMs. MSFT Co-Pilot has been a dud. MSFT does own a 27% stake in OpenAI, but my guess is that OpenAI would experience valuation multiple compression were it to try to execute another private round, which is why it will float shares to the public (easier to trick public company investors). MSFT will have egg on its face when the AI trade has fully played out.
ORCL: ORCL shares have another leg down. ORCL hopes to see Cloud revenue accelerate as evidenced by its bogus long-term OpenAI “contract”, but my guess is that the combination of Google (GOOGL), Anthropic and capital constraints will smash OpenAI over the next several years. Oracle picked the wrong LLM partner.
GOOGL: Google has so fully embedded its multi-modal model - Gemini - into its business that it will be impossible to tease out the revenue lift that Gemini provides Google Cloud, Google Workspace, YouTube and Google hardware products. Google’s holistic approach to Gen AI is the only smart, strategic approach I can point to among the cloud service providers. If you haven’t tried Gemini 3.0, Nano Banana or Antigravity, give them a shot.