Amazon - See data hoarder [1]. Other than that, debatable because Amazon organizationally seems to optimize for efficiency (e.g. compressing margins). However, by cracking it into a logistics company and an online retail company, users would benefit from broader access to Amazon's efficient supply chain. Furthermore, the incentive for Amazon to abuse its logistics advantage to muscle out competition (e.g. Amazon Basics + first-party product offerings) would be lessened.
Apple - Probably the most debatable of the group. See first-party app store [2]. The biggest case for Apple would be about how lazy they've recently been for non-margin generating offerings. Since iTunes, and then the app store, they've gotten lazy about relentlessly innovating. You've seen it in the iPhone and the MBP. To say nothing of their server offerings.
Google - See data hoarder [1]. See first-party app store [2].
Facebook - See data hoarder [1]. See first-party app store [2].
Google and Facebook are essentially the Valves of everything. They're spitting out cool things here and there, while rolling in the mountain of cash they're taxing from the rest of the economy. You'd be hard pressed to convince me that cash wouldn't better be spent by other companies.
[1] Data hoarder. By centralizing collection of user / customer data (usually by first-mover advantage, leveraging that to buy any existential threats), these firms have created a dominant market position. They control the platforms, so they collect the data, so they have more data, so they can more efficiently monetize that data, so they can afford to buy competitors (repeat). This cycle is unlikely to be broken without regulatory intervention.
This is firstly a sub-optimal state of affairs because it allows these companies to rent seek from all downstream consumers of this data (most clearly: advertising). By leveraging their position as the sole source of an ongoing data feed for "their" users, they position themselves as the only place to obtain that data in a functional manner. Ergo, all downstream consumers are compelled to purchase it from them, or do without.
The counter-argument is that they perform this function more efficiently than a hypothetical freer market of numerous competitors. However, I leave it to you to decide how often an entity without external competitive pressure has chosen to operate in the most globally efficient manner for their customers.
Secondly, this impacts product development, whereby the data owners may choose at their sole discretion not to share this data with competitors. As a result, there are products that only these parties are capable of creating. Because of the scale of these organizations, there are many opportunities they will not see as worth their time. Consequently, those opportunities will never be pursued, as the smaller companies interested in doing so simply can't without access to that data.
One might argue that that's sub-optimal, and that eventually the data owners would awaken to the profit opportunity in selling that data to the interested companies. To which I would retort something along the lines of "GM, Ford, and Kodak all recognized new market opportunities, and did they pursue them?" Institutional conservatism and inertia at scale is a lazy thing.
[2] First-party app store. By owning the underlying hardware platform / presentation interface (in Facebook's case), there's a clear conflict of interest. First-party app store owners face a substantially lesser incentive to innovate, secure, or respond to customer demands than they would in a free marketplace. As a result, app store management practices and technical functionality rot in comparison to a free market in which they were forced to compete with competitor's app stores.
Without going in depth, there is an obvious financial component, whereby first-party app stores with a monopoly are able to rent seek and extract a greater-than-optimal cut of app revenue. In a free market of alternatives, greater portions of app revenue would flow through to developers (note: in individual, the aggregate case is admittedly more complex).
Additionally, this extends to app censorship and policies. By controlling app stores as the gateway to "their" customers, the underlying owners can shape the kinds of apps they allow to exist. This produces negative outcomes in at least two ways. (1) Apps' responsiveness to actual customer demand is suppressed, as they must first and foremost satisfy app store rules. This can be seen in censorship policies (e.g. issues with distributing open source or "PR sensitive" apps), as well as functional abilities (e.g. the variety of always-on / background execution use cases precluded for all but first-party apps). (2) The app market is distorted to optimize for revenue in the current system, as opposed to revenue from actual customer demand. If there were a freer app store market, I offer that you wouldn't see 1,000 clones of "flashlight w/ ads" apps.