Oracles can be crucial for certain applications that are and will be build on the Tezos blockchain platform. Being fully dependent on Chainlink or other oracles could not be the best choice long term. Yes, we know, Chainlink has a working product, but that doesn't mean no other oracle networks will exist. Tezos has an ever growing community and will be able to pull off a decentralized oracle structure of its own. Maybe some Chainlink integration could happen. But Tezos will have an oracle structure of its own. It's inevitable and healthy for the Tezos ecosystem. At this point of time, there are two initiatives known for Tezos-based oracles. (But most likely, there will eventually be some interlinked structure that benefit all Tezos-based projects.)
The TQ Tezos innitiative:
First, we have TQ Tezos, which is a "Technology company building Tezos-based solutions and open source software for enterprises and developers."
They have an initial setup ready: the 'Oracle quick start".
The Tezsure initiative:
What is Tezsure? Tezsure is an insurance platform that is build on the Tezos blockchain. Tezsure is founded by Bernd Oostrum, who worked as the main blockchain architect/developer for one of the biggest insurance companies in the Netherlands. He has extensive experience building blockchain solutions for insurance companies and the Dutch government. Tezsure is funded by both silicon valley and blockchain valley: through Draper U venture, the Tezos Foundation and CV Labs Zug. More about Tezsure here and here.
An insurance application would have to rely heavily on off chain data that can be provided by an oracle setup. Tezsure has already build an oracle setup in the beginning of 2019 to showcase their first pilot: Crop insurance. At this point, Tezsure is taking the next step and asks for input from the Tezos dev community to get a feel for what devellopers would like to see in an oracle network that can help their business in the (near) future.
Tezsure is not trying to reinvent the wheel and will use simular architectures as the existing oracle solutions. With one key difference: there will be no use for a new custom token to create incentive for the network. The token to be used will be Tezos' native token: XTZ. Since the orcale network will only be providing data for Tezos-based applications, XTZ will work fine as token for the network:
Keep it respectful
Some of the Chainlink supporters have gone pretty nuts over these developments on twitter. So a polite request: please refrain from any obscenities and insults in the comments. I don't have all the answers and the project is in development.