What do you define as "herd immunity"? My understanding is that it is defined as a large enough pool of infected but recovered people, that the disease finds it difficult to find uninfected people to spread to. So if they are highly underestimating infection, that would be a good thing?
The goal is to avoid fatalities, and also unnecessary and prolonged suffering, whether in hospitals or at home. The goal isn't necessarily to prevent infections, in fact, infections are inevitable, the only conceivable way to prevent infections is to stay locked down forever (even that probably will not work, like I said before, it's just a modern-day mega-scale "Parikshit syndrome"). No govt. goes all out to prevent common cold infections, because the death rate is minimal (still not zero of course). There's no herd immunity for the common cold either, because people keep getting reinfected, "cured" patients don't retain immunity for more than a few months.
If the fatalities in Sweden are at a plateau, that is the desirable end state (of course, also have to look at how many people are being subjected to suffering, regardless of recovery). And fatalities do seem to be slowing down. Infection numbers by themselves are meaningless.
Sweden being sparsely populated - have you looked at the population densities of their cities? Average population density of a country isn't so meaningful, you have to look at clusters like cities or towns. That's what determines disease spread. Sweden as a whole being 22 people/ km^2 isn't so meaningful, Ladakh is way less than that, even Arunachal Pradesh seems less than that.
For herd immunity to be effective for disease like COVID19, we would need a large % of population infected like 60-70%. By that number is reached, already a lot of damage would have happened. Even 1% of those 60% die, in India's case it would mean 0.6% of total population or close to 7 Million deaths in very short time. And that is an underestimate from all current data. The moment we see saturation in hospitals the death rate is bound to spike.
Most cases are in Stockholm. And its not like they are not doing anything and still their death rate is going down. The WFH has increased quite a lot in last few weeks, many companies have started furloughs. People there are maintaining self quarantine even if they have any fever or cold (they won't be tested, but are advised to remain in home quarantine and they are following it quite diligently. Even though govt had not enforced, many are following Social distancing to the letter, unlike in some other Countries. Of course, maintaining Social distancing is rather easy in Sweden even under normal circumstances, even in their most dense population centers. Try doing that in any Indian City while maintaining daily activity. You don't have to take population density literally, if you want apple to apple comparison compare Stockholm to Mumbai. Its just a rough parameter to give an idea on the order of magnitude.
The underestimation is still nowhere close to the levels where herd immunity will start playing a role. We simply cannot conclude herd immunity is doing any magic here. For all we know, its the social distancing that is playing a role here. And unlike in case of Italy, Sweden is much well prepared to handle the severe cases, they had a slow start (now we should wonder why a relatively slow increase in infection rate for Sweden, compared to even their neighbors) which gave them time to increase existing preparedness.
While its reasonable to assume this virus will work like other viruses and the recuperated persons should have at least temporary immunity of a few months, I am yet to see any conclusive proof showing this. Has anyone tried to reinfect any recovered volunteers under any control study..?? I think not. One Chinese study I saw, found the anti-bodies waning rather quickly from the cured people. We will need a lot of study and data first to establish immunity is good enough and holds at least a few months, so the herd immunity can work even in theory, before we can use it to make any policy decisions. To make matters worse we are seeing cases of resurgence. In SKorea so far 2% of cured people are found positive again, as per a report I saw. Have they tested all cured patients repeatedly to see if they have captured all resurgence cases..? I think not. Do we have some of the asymptomatic patients with resurgence to enough viral load in their bodies so they become infectious again for extended time period..?? We don't know. (though its reasonable to assume asymptomatic people would see their bodies eliminating the virus soon enough). Too many unknowns. Do we want to conduct an experiment on our population by erring on the wrong side in National decisions..? May be some countries can afford to do so as they are better equipped to deal with the fall-out. Others have thin margin of error.