$NVDA - NVIDIA is currently capturing significant value from AI infrastructure spending, with better margins than other layers of the stack. The hardware layer is accruing substantial value in the current AI buildout phase.
$ANTHROPIC - Anthropic has achieved massive revenue growth by focusing on coding as a use case, going from $9B run rate to $47B run rate, demonstrating strong product-market fit in software development.
$AILABS - Foundation model companies are raising unprecedented amounts of capital and driving the frontier forward in AI development, despite questions about long-term commoditization. The software productivity market alone could be worth a trillion dollars.
Bearish:
$OPENAI - OpenAI has shown strategic inconsistency, initially trying everything (ads, e-commerce, browser, social video) before pivoting to coding. Questions remain about their ability to bridge the gap between power users and the 40% of weekly users who haven't achieved daily engagement.
$AILABS - Foundation models face structural challenges to capturing value: no clear network effects, no sustainable differentiation, chatbots are limited V1 UIs, and models appear to be commoditizing infrastructure similar to ISPs and telecom networks that built expensive infrastructure but captured little value.
$TELCO - Telecom operators serve as a cautionary tale for AI infrastructure providers - they built expensive global infrastructure with enormous usage growth but captured no value as it moved up the stack, with stocks flat for twenty years despite massive CapEx.
$INFRA - AI infrastructure faces a structural pricing problem with massive disequilibrium between supply, demand, and pricing. With $700B-$2T in CapEx coming and models becoming 100-200x more efficient annually, pricing power will erode as multiple companies compete selling commoditized compute.