· 个股论点
我不会将全栈(Full-stack)和垂直整合(Vertically integrated)混为一谈，我在谈论 $ORCL 时措辞非常精确。 $NBIS 采用混合模式以扩大利润率。 $IREN 的主要问题在于他们正在做全栈业务，其中软件编排(Software orchestration)是一个令人难以置信且被误解的护城河，但这将执行风险放大了数个数量级。 如果你做对了，你将拥有行业内最高的利润率/收入。但 $CRWV 花了数年和数十亿美元通过收购才做对，其CEO在电话会议上表示软件是他们的护城河。 很难让市场相信，$IREN 作为一个全新转型进入高性能计算(HPC) AI 云的公司，其全栈能力会优于 $CRWV 花数年构建的能力，甚至优于市值5000多亿美元的 $ORCL 至今未能做到的程度。这就是导致我做空(DG) $IREN 的持续执行风险。 但当然，如果他们能精准做到（这很难），那么其回报潜力将极其巨大。
I don't use full-stack and vertically integrated interchangeably, my wording is very precise when I talk about $ORCL. $NBIS does mixed to expand margins. The main issue with $IREN is that they're doing full-stack where software orchestration is an incredible and misunderstood moat and it increases execution risk by magnitudes. If you get it right you have the highest margins/revenue in the industry. But $CRWV spent years and billions in acquisitions to get it right and the CEO said on call that software was their moat. It's hard to convince the market that $IREN, which is doing a brand new pivot into HPC AI cloud, will have better full-stack that $CRWV spent years building and that $ORCL a $500B+ company failed ot do so far. This is the lingering execution risk that led me to DG $IREN. But of course if they get it spot-on (which is difficult), then it has incredibly high return potential.