There are many arguments against replacing Trident with a new submarine-launched nuclear weapons system. Critics like me have argued that its ridiculously expensive, that its proponents can't say when it should be used, that its use would be illegal under international law and that its a Cold War system that's irrelevant to today's geopolitics.
We also note that our nuclear weapons did not deter the Russians from annexing Crimea or DAESH from conquering half of Iraq.
But here's another one - it may be too easy to detect by the time its ready.
Writing in New Scientist (27 May, p37-39) David Hambling reminds us that nuclear submarines are big and release a lot of heat. Therefore a big nuclear sub makes noise, disturbs the water and leaves traces. Hambling reports evidence that such traces may be detectable for hours to days after the sub has passed; much longer than generally thought. The best research on this is, of course, secret.
There's another reason. The development of AI and low-cost drones has vastly reduced the cost and increased the effectiveness of airborne surveillance. This is also being extended to underwater surveillance. By the time we launch our new Dreadnought submarines the skies and seas may be full of nosy drones happily reporting their positions to Washington, Moscow, Beijing - or even, via a new app, to a phone near you!
£100B would be a lot to pay for a secure nuclear deterrent whose position isn't actually secret.