There is a lot of space between a frame being safe to ride and being the quality you see even on second tier frames from the big brands.
It’s worth thinking about the economics of a replica frame selling for $500 painted up as a Factor or Pinarello. I know that it would cost me about $2000-$3000 USD to get a contract manufacturer to make, design a layup for and test a handlebar. I would guess that a frame would be on the order $10k-$12k. It’s probably a breakeven at 50 units at a factory cost of $300 and 30 at less than $200 to make money on a replica frame. To sell 30 frames, you probably need them to pass basic fatigue and safety tests to make sure you can sell even that many. Your don't have much additional budget for development so other than a 3d scan of each frame size to get the shape right and some basic cad work to smooth out the details from the relatively low resolution of that scan, and then work to get the layup good enough to pass the testing. There isn't much budget for additional work. While external tube shape is the majority of the strength of a tube, thickness still matters for stiffness. Because you are using lower quality materials you will end up with thicker tubes, a really stiff bike, and a heavier frame. Its probably been tested enough that it wont break, but will also likely inherit the flaws of the original design because on the outside, it needs to look like the same.
In the case of Factor, they replaced their headset assemblies and changed the design since the Ostro VAM was pulled from competition in 2021 due to sheared steerer tubes. Are you getting the corrected versions of the parts on the Ostro VAM replica? Are the parts even of sufficient quality to avoid the problems that led to the re-design? Can these replica makers get the details of a C shaped steerer right enough to be reliable in the long run? What about the headset and derailleur hanger hardware? Potential headset issues also track with the flaws seen on Cervelo replicas which also use a non-round steerer.
The Pinarello Dogma F suffers from a surprisingly high incidence of cracked seatstays and chainstays as a result of the asymmetric design. The Velobuild version also seems to have this problem. Plus you get bonus headset problems.
An open mold is usually a little more expensive. But, because it's a clean sheet design, they are going to design it in CAD and run the basic FEA stuff and use the results to iterate the design. They have both a good understanding of where / how to adjust the layup and the freedom to change tube shapes to get the performance they are looking for. It looks more generic, but the engineer and designers are incentivized to maximize compatibility with 3rd party components and make the best riding bike they can within their budget because they are looking to secure sales for dozens of bikes with multiple small brands. As a result, these bikes almost always use forks with round steerers and common headset sizes that are easier to design, to make to a high standard, and possible to source from 3rd parties.
A branded version of one of these open mold frames from a brand that does development like a Cinelli or a Guerciotti might ask for layup adjustment, will do some additional QC and swap out the generic hardware for branded FSA or Deda versions. They might stick them in a wind tunnel to help find the right handlebar and wheels to sell with the frame. They might also swap the fork. These bikes win big races at the U23 and junior levels, cost 3x as much, but generally speaking most reviewers will say they fall short of the mid-range bikes from the big brands.
Is the replica safe to ride? Probably, but the commercial motivations are totally different and there is a reason why the open mold is more likely to have the details sorted. So yes, there is an inherent reason why the replica is likely to be a worse bike.
