Been digging into the energy sector lately and there's something worth paying attention to. With AI consuming massive amounts of power and Goldman Sachs projecting a 160% surge in data center needs by 2030, energy demand is about to hit different. Throw in the revenge travel trend pushing crude demand up by 3.5 million barrels daily through summer, plus OPEC keeping production cuts tight at 5.8 million barrels a day, and you've got a pretty solid fundamental backdrop.

What caught my eye though is that small energy companies are getting overlooked while everyone's chasing mega-caps. There are actually some really interesting small-cap plays here that are trading at crazy valuations and still throwing off decent dividend yields.

Teekay Tankers is one that stands out. Canadian shipping company, up 25% YTD, trading at a forward P/E of just 4.31 with a 1.61% dividend. Analysts are calling it a Strong Buy with 20% upside potential. Then there's Riley Exploration Permian, an Oklahoma-based oil and gas explorer focused on the Permian Basin. Only 649 million market cap but yielding 4.77% in dividends and analysts see 63% upside.

Scorpio Tankers is another one worth looking at. Monaco-based, specializes in refined product tankers, up 23.9% already but still trading at a forward P/E of 5.41. Crescent Energy out of Houston is down 8.8% YTD but that's actually created a buying opportunity - 3.95% dividend yield and trading at a 6.34 forward P/E with 42% upside according to consensus.

Finally, Northern Oil and Gas has an interesting model where they buy minority interests in wells rather than operating them directly. Up 11.5% on the year, 3.84% dividend, and analysts see another 21% upside.

The thing about small energy companies right now is they're still cheap relative to their growth drivers and most are paying real dividends. If the energy thesis plays out - and the AI power demand alone seems pretty hard to dismiss - these small-cap energy plays could surprise people who've written off the sector entirely. Worth doing your own research obviously, but the risk-reward on several of these looks interesting from where I'm sitting.