I can't give you a firm answer about the people in question, but I have spent some time working with census aggregates. So some general considerations:

It might be an error. Here's how the census was done at the time:

A guy would come up to your house. He had a big ledger with him full of pre-printed census forms. As I'm sure you know from looking at it, the 1840 census asked a fair number of questions. The census taker would do the tallies and mark out the form himself. If you've ever worked on a large spreadsheet, you know how easy it is to migrate a row or column or two over without realizing it. This is the most likely explanation for the incredible (not just surprisingly large, but literally unbelievable) number of free black Americans recorded as mentally ill in 1840. It was so nuts that John C. Calhoun, of all people, admitted to every error pointed out to him. However, he refused to redo the count (as Secretary of State at the time he was in charge of such things.) because "everybody knew" that freedom made black people crazy. He's still John C. Calhoun. It wouldn't be a big surprise to find out that individual tallies got messed up elsewhere.

However, it might be legit. The standard for who was a member of the household amounted to whoever lived there regularly on census day. If he boarded servants and apprentices they would normally be counted as part of his family. This was also true of slaves. Family, in a census context, means the same as household does to us rather than requiring a biological or married relation. The longterm residents of a boarding house and inmates at asylums were members of the owners' or manager's census families.

With regard to the manager vs. owner distinction. Errors are always possible, but if the owner of the Iron Works didn't live on-site then the manager would probably be listed as the head of household. There are issues similar in regard large plantations, which is where I learned all of this. The really big planters often had operations in multiple states. They themselves might appear in both the censuses of, say, South Carolina and Georgia due to that property. But the census worker was supposed to list the people present and/or primarily resident at the time Legal ownership wasn't necessarily the controlling factor. Given that plantation overseers often lived on site, and plantation owners might have a separate residence where they spent most of their time (This is especially common in the South Carolina lowcountry.), the overseer is the obvious candidate for head of household when the census person comes by. He lives there and is generally in charge and responsible for the property.

The way to check this for sure would be to look for a separate entry for the owner. Has he got a house in town somewhere? If there's a full accounting of him elsewhere, then I'd be pretty confident that placing the overseer as head of household for the iron works was deliberate rather than a mistake. I lean towards it being so anyway, but the separate census entry would really seal the deal.

I assume that all these folks were the employees of the ironworks. They would be living on site. If the owner of the factory lived elsewhere, he wouldn't be included within this household. It sounds like the census makes a clear distinction between the biological family of the manager and other employees. These employees might be apprentices or they might be full journeymen or waged employees. I doubt you'd find indentured servitude in 1840.

Make sure you check the census of 1830 and 1850 to trace the growth of this endeavor.