$150 in "profit" per tablet is beyond ridiculous. I have no idea of where you could pull a number like that from. The common wisdom is that the gross margin on Nexus hardware is around 0. Google's cut from what little content they manage to sell can't amount to much. And the incremental ad revenue from the average Android tablet user isn't going to be more than a few bucks a year either.
Fortunately for you I generally can show my reasoning on the things I put out there :-) In this case there was data from the Apple vs Samsung trial [1] which if you recall said 775k Galaxy Tab's generated 325M$ in revenue. Doing that math gives you about $448 per tablet in revenue. iSuppli calculated the original Galaxy Tab cost Samsung about $214 to make. That would suggest a gross profit in the neighborhood of $234 per tablet. I was assuming that there was some sort of "OS/License" fee on top of that which is why I didn't suggest it might be as high as $200 each.
[1] "Samsung sold 725,000 units and generated $325 million." - http://news.cnet.com/8301-13579_3-57490740-37/samsung-tablet...
[2] "When the $9.35 manufacturing cost of the Galaxy Tab is added in, the combined materials and production cost amounts to $214.57." - http://www.isuppli.com/Teardowns/News/Pages/Samsung-Galaxy-T...
That's a very different case to the much more aggressive pricing used for the Nexus 7 (publically stated to be selling with no margins) and the Galaxy Nexus. And the new Chromebook too, come to think of it. I guess it's possible that they're switching strategies again. So yeah, if these are priced at $600, I'm sure there's a healthy margin on the few units they manage to sell. But until prices are announced, I see little reason to think that there is such a strategy change.
While all SD cards are SPI-compatible no one really accesses them that way (it's way too slow). Almost all SD cards in electronics today use the proprietary SDIO bus.
SD cards use a protocol called 'SPI', its a serial protocol and requires a total of 5 I/O lines to the processor. The connector [1] is less then 10 cents. I can't really speak to the cost of milling a slot in the side of the case. So we're not talking a large expense, or even a nominal expense. There are consumers for whom its a deal breaker that it isn't there, and few of whom which won't buy a tablet if it has a card reading slot. By that reasoning I conclude that it only expands the market (and doesn't contract it) assuming a 5x markup cost->market (50 cents), Lets assume Google makes $150 in 'profit' per tablet, so selling one to me with a card reader pays for 299 (($150 / .5) - 1) card readers in tablets where the customer didn't care if it existed. 1/299 is 0.3%. So if adding a card reader expanded the sales by more than 0.3% it would be a net gain for Google. That seems pretty achievable.
Now lets talk about "but I never needed any more memory." (more of incision's point than yours) there are two benefits to a card, one is expandable space, but more is 'rapidly replaceable content' I can take a 16GB micro SD, or a 64GB SDXHC card, and swap it out in a 3 seconds. That is equivalent to a network between 5.3 and 21.3 GB per second (53 or 213 gbps) If I rip a few movies into a card to carry with me on the plane where there is no decent internet, and then when I arrive I can pop them out and put my books back in, or what ever content or app data that I need locally. That is a user model that may eventually change (although keeping a copy of a movie in the cloud to watch on your tablet probably gets you in trouble) but for now its being held hostage by a 10 cent connector.
[1] http://www.alibaba.com/product-gs/656158231/ST9S008V4AR1500_...