How does he view the timelines for the new big Athabasca mines? When can we expect new meaningful supply from them?

Could be a 10-bagger within a year if things go well. Future nationalization can be a risk although it seems lower at the moment, but selling around production start might be an option to lower that risk, at least for traders like me.
Just my own thoughts, please do your own DD.

$GLO, it is Niger with jurisdiction risk, but the only big greenfield project under construction, 4 Mlbs/year expected. With debt financing or JV in place, it should pop, and keep rising significantly until production start.

$RSH, top exploration project in the US with zero dilution as Myriad does all the financing.
$WUC, US miner with tiny mcap, expected to sign agreement to deliver ore to the White Mesa any day.

I continue being invested in high torque companies, trading the dips, ETF rebalances, etc., and my portfolio has been doing quite well despite the overall downfall. My current top picks are:

But in summary, fundamentals keep being better than ever and the uranium prices have to go much much higher to incentivize new production and exploration. It must happen soon, otherwise some nuclear plants will have to shut down, and we can forget about SMRs.

Access to capital is scarce, and the companies, especially those moving projects forward, have high costs and therefore keep diluting shareholders. Many retail investors, after adding at every dip, are out of money.

Supply chain issues, inflation, workforce issues, etc. will all make it take so much more time and money than expected.
The biggest problem as I see it for the uranium companies and investors is that the thesis has taken longer time to play out than expected.

They need to build a mine, a huge mill, airport, housing, etc. And they are not even permitted yet. Same for all the new Athabasca projects, they all have in common that there is no infrastructure nearby and they need to build it all from scratch.

On the other hand, new projects keep being delayed and ramp-up slower than expected. Just look at US producers like EU, UEC, URG, PEN, etc., everything is slower than expected and production numbers are really low. NexGen will need a miracle to enter full production by 2029.

It seems evident those numbers were bigger than at least I had expected. Another factor is that conversion/enrichment capacity must be expanded, but it takes some time before that converts into demand for u3o8. But you can't go on like this forever, re-stocking must soon start to happen.

But since all the time more uranium is used than is produced, it means inventories are used. And that is often the unknown number in the equation: we know the production numbers and the consumption of the reactors is easy to calculate, but mostly we don't know how big inventories utilities have.

Utilities decided to not buy on the spot market, and since there is always some volumes of uranium entering the spot market, spot will trickle downwards if there are no buyers. Investor sentiment was not good enough to get SPUT into a premium to NAV to support the spot market.

I think what held us back after the spike early 2024, was that production from Cameco and Kazatomprom has been quite in line with expectations, so investors who expected worse numbers lost a trigger.

Downblending of nuclear weapons would be nice, but still very early talks, and I don't expect there is enough trust between the super powers to really agree on this. And even if they do, the quantities would be far from as big as last time around, and would not be able to bridge the supply gap.

I think the truth is quite simple. The fundamentals, that is the supply/demand picture for uranium, has not changed much. It continues to be the best seen in decades. DeepSeek may have shown AI can be more energy effecient, but that is not expected to have any particular effect on future demand.