And all of this sits on the assumption that these hyperscalers survive. Several smart people took a closer look on the balance sheet on and off the record and there seems to be a lot of hope as business model.
E.g. Morgan Stanley found ca. $620 B off the balance sheet commitments among the 5 big ones (This was in Jan-2026). Those are "RVG´s" or backstops for data center builders (and leasers) which can be hidden in the GAAP and therefore window dress the balance.
Moodys did the same deep dive in May and said this:
"If a company concludes a lease renewal is likely to be
exercised, but not reasonably certain, it can avoid
classifying both the lease renewal periods and the
residual value guarantee (RVG) as liabilities."
That is a lot of money not showing up in any calculation and adding to CAPEX for all the other purposes. Michael Burry has a good summary on it.
What if the next winner in nuclear isn't the company with the best reactor design, but the company that can deploy commercial power first?
AI demand appears to be arriving now. Many advanced nuclear projects are still targeting the 2030s.
That timing mismatch is one reason I have been researching newly public Deep Fission ($FISN). Their thesis is not better reactor physics. Their thesis is faster deployment.
If the bottleneck becomes deployment rather than demand, the market may begin valuing nuclear companies very differently.
And all of this sits on the assumption that these hyperscalers survive. Several smart people took a closer look on the balance sheet on and off the record and there seems to be a lot of hope as business model.
E.g. Morgan Stanley found ca. $620 B off the balance sheet commitments among the 5 big ones (This was in Jan-2026). Those are "RVG´s" or backstops for data center builders (and leasers) which can be hidden in the GAAP and therefore window dress the balance.
Moodys did the same deep dive in May and said this:
"If a company concludes a lease renewal is likely to be
exercised, but not reasonably certain, it can avoid
classifying both the lease renewal periods and the
residual value guarantee (RVG) as liabilities."
That is a lot of money not showing up in any calculation and adding to CAPEX for all the other purposes. Michael Burry has a good summary on it.
https://substack.com/@michaeljburry/note/c-278923362?r=5klxyn
Looking forward to your SMR suggestions...
One question I've been thinking about:
What if the next winner in nuclear isn't the company with the best reactor design, but the company that can deploy commercial power first?
AI demand appears to be arriving now. Many advanced nuclear projects are still targeting the 2030s.
That timing mismatch is one reason I have been researching newly public Deep Fission ($FISN). Their thesis is not better reactor physics. Their thesis is faster deployment.
If the bottleneck becomes deployment rather than demand, the market may begin valuing nuclear companies very differently.