If you'd like something that's set based and will most likely perform a bit better than looking at each individual character give the following a shot... Like the above I made an assumption that the lastname starts with a capital, no requirement that the firstname be capitalized.
Also you'll see the number 1000 used in my character expressions, you can adjust that down to the limit of your column left or actually do the Len calculation if you like... Just remember to Keep the Collation as that's what gives you the case sensative bit...
CREATE TABLE #test (
INSERT INTO [#test]
SELECT 'RobertJain' UNION ALL
SELECT 'JansonStuart' UNION ALL
SELECT 'IsmailSummit' UNION ALL
LEFT([fullname], PATINDEX ('%[A-Z]%', SUBSTRING([fullname], 2, 1000) COLLATE Latin1_General_BIN)) AS FirstName,
SUBSTRING(fullname, PATINDEX ('%[A-Z]%', SUBSTRING([fullname], 2, 1000) COLLATE Latin1_General_BIN)+1, 1000) AS LastName
DROP TABLE [#test]
To help us help you read this
For better help with performance problems please read this