Like much blockchain technology, the Hyperledger ecosystem is relatively new, and many projects have not even hit a full 1.0 release yet. While there is a large amount of development activity and multiple working systems already in production use, the system as a whole is fractured. For instance, Sawtooth is written in Python, Fabric in Go, Quilt in Java, and so on. Even staying inside the Hyperledger family, it would be difficult to use a homogeneous set of technologies for end-to-end implementations.
Moreover, Hyperledger's focus on private networks is a problem for the projects that may wish to have a public component. One of the appeals of blockchain technology is transparency. A project that seeks maximum transparency through public usage of their technology may need to look elsewhere or find a way to bridge between Hyperledger and public networks—possibly by using Quilt and ILP.
Similarly, projects looking to raise funds through an ICO should probably look elsewhere. Few projects have tried to use Hyperledger as part of an ICO, and, as far as we know, none of those have actually succeeded in fundraising. Hyperledger remains oriented strongly toward private networks—where it has succeeded tremendously.