Antiship ballistic missiles have very limited maneuverability in the terminal phase where their onboard sensors come into play. Atmospheric heating and plasma effects reduce the effectiveness of those sensors. This article is a fairly good summary of the current state.
https://www.fpri.org/article/2021/05/chinas-anti-ship-ballis...
All of this stuff is possible in theory. It's just extremely difficult coordinate all the moving pieces together in real time and make it work in an operational environment.
Carriers are large enough that you can start scanning for them before the terminal phase.
The article you linked is overly optimistic. The idea of masking the heat signature of a literal aircraft carrier is completely insane.
I also suggest you do the math on how much you can manoeuvre at a rate of speed of 3Gs at Mach 5 with a manoeuvring distance of 80km. (Hint : 33km radius). This is also an underestimation because it doesn't take into account decrease of transverse velocity.
This stuff was possible in theory 30 years ago. Now it's likely possible in practice.
That drone is most likely to engage an AEGIS missile cruiser (one of the escort ships) providing a protective perimeter around the carrier first. Which has more than enough capabilities to shoot down any "drone" before it reaches the carrier.
The carrier itself also provides multiple AWACS to constantly be scanning for aerial approaches. That's the good thing about a carrier, its got aircraft of its own defending the group.
Can your drone penetrate the scanning area of a E2 Hawkeye?
And yes, a small modern drone can penetrate the scanning area of an E2 Hawkeye. It turns out stealth is a lot easier when your platform is much smaller, and you only care about the frontal aspect. Even moreso when you don't have large-wavelength radar.
Repeated air assaults to repeatedly tear down defenses (ex: target the Cruiser first, removing the first shield of the Carrier. Then target the next cruiser. Then finally work your way to the carrier) is certainly a workable plan. But again, this is a repeated air-assault against a CSG.
My point is that its not going to take "one missile". Its going to take a campaign, at least if the CSG is doing its formation correctly. In the meantime, the CSG is presumably in range of its target, and also launching an all-out aerial assault.
And no, I'm not suggesting lobbing it around the first cruiser. I'm suggesting to approach the cruiser from a different angle, finding another ship of the CSG and build a perimeter, and then strike the center.
This is not a repeated air assault of the CSG. That's an old doctrine. The doctrine is to completely bypass the defences of the CSG using a hypersonic missile.
And no, you don't do this while the CSG is already in range, that's ridiculous. You do this when it's ~2000km away.
The AEGIS Cruiser's tomahawk missiles have 2500km range.
You're not even outranging the escort ships, let alone the carrier at 2000km. That's why two AEGIS Cruisers escort a carrier.
--------
BTW: This magic drone that can escape detection yet provide enough intelligence to pinpoint ships so that missiles can strike their targets on the high-seas sounds a lot like an F35.
And no, the drone we're talking about is nowhere near an F-35. Examples of such drones are the CASC CH-7, the Okhotnik S-70, the GJ-11, the RQ-170 or the XQ-58. Unlike an F-35 these are extremely cheap, unmanned, and have very long loiter times, meaning that they can cover hundreds of kilometers in radius or even more.
Antiship ballistic missiles have strong maneuvering capabilities. You have to look at absolute displacement, not the relative change in velocity. And the sensors required to find a 7000m^2 target are much smaller and limited than you'd think.