This has been a debate question still keeps on running between Microsoft HYPER-V team and customers and IT Pros. When we want to put production servers into virtualization we always look to avoid any single hardware failures (HDD failures, power failures…etc) so network card failures are among key component where failure is not acceptable. So to avoid that we normally can combine 2 or more network card and combine them as a NIC team or sometimes referred as Link aggregation. So having multiple physical NICs, teamed together, to provide additional redundancy, but at the same time, improving link speed in many cases. Wikipedia has a good explanation.
Now the problem start with HYPER-V is this is not officially supported by the Microsoft. They state this is related to the hardware side so vendors need to support that. Currently Broadcom, Intel and other vendors working on their own solution for that. In fact if you have been handling HP product you much be familiar with the HP NIC teaming software very much. But still due to the complication we have the doubt whether to go ahead and do NIC teaming on HYPER-V.
Look forward to see a firm answer from Microsoft HYPER-V team soon. Apart from that most of the features available in the HYPER-V is outstanding. (Compared as a free virtualization product)