I think $MSFT / $NVDA risk / reward pretty mediocre up here.

$Microsoft(MSFT.US)$ is obviously great long-term, but the two main businesses are not exactly accelerating materially or anything. 1) Office decelerated to 15% yy cc last qtr and Co-Pilot is only kind of sort of helping for now (might improve down the road). 2) Azure growth is probably peaking (30% yy cc last qtr and maybe will be 31-32% yy cc) and last qtrs AI contribution was pretty weak (another qtr of that and will be a cold shower on the stock). I'm above street and have high-teens EPS growth going forward. 30x that on CY2025 is only 3% upside, and CY2026 has only 20% upside.

$NVIDIA(NVDA.US)$ is also obviously great. But I think two real observations are that 1) there is obviously a ton of duplicative investment in model training occurring and now at pretty large scale (that will get rationalized at some point), and 2) the Chinese claim their AI is as good as ours using last-gen chips but current open-source models, which suggests there is a ton of room for advancement in the software layer outside NVDA, and the rest of the world might be inefficiently pursuing AI progress by throwing capital at next-gen GPUs. Maybe need to discount Chinese commentary.

Qualitatively, discussion about the company has sort of left earth's orbit and we might be in the honeymoon period where everyone thinks the future is only straight up and to the right.

Street has NVDA growing 97% this year and 33% next year w/ flat margins. No one knows what next year will look like but stock trades at 36x 2025 EPS. So a far cry from the ~25-30x 2024 EPS which is what we had earlier this year when it was pretty interesting.