The company is the sum of it's parts: The employees. Employees exist without the founder or the company. Thus, the value still exists within the economy, it is just the distribution is different as said employees are producing value in other sectors of the economy.
When a founder forms a company he is not creating value. He is redeploying existing labor to do OTHER things. Whether that creates new value is a separate issue as said labor could create new value wherever it was before. Additionally, the founder is also creating a new corporate pyramid structure that funnels a lot of the redeployed wealth from this labor towards the people sitting on top: Whether that be himself or shareholders.
Let's say there are two companies. Company A and Company B. Company A is a established company and Company B is a startup. Lets also say there is one worker, named Chris. When Chris works for company A his work output is 20$ a day.
Let's say Chris changes his job and moves to company B. You're telling me that Company B by virtue of being a startup makes chris's work output suddenly equivalent to 100$? Tell me what magic makes this the case? What is it that makes it so new value is created out of thin air? It's not physically possible. In fact it's stupid to think that just by creating a startup you're creating value. You're just redeploying labor.
So where does new value come from if not just from a company by virtue of being a startup? The answer: When new technology or new innovations are created that fundamentally change the infrastructure of the economy so that it can be more efficient. It can happen from ANY company and it can even happen from government, it does not have to come from a startup. Examples: Apple: iphone, military: internet.
To tell you the truth, most startups amount to shit. If anything the majority of startups redeploy labor onto endeavors that amount to nothing. Usually big corporations more effectively redeploy labor. Could any random startup do what apple did to the phone industry? Probably, but waay less likely.