diff --git a/modules/angular2/src/compiler/html_lexer.ts b/modules/angular2/src/compiler/html_lexer.ts
index e3a725ce9d..f410fbdb73 100644
--- a/modules/angular2/src/compiler/html_lexer.ts
+++ b/modules/angular2/src/compiler/html_lexer.ts
@@ -378,7 +378,9 @@ class _HtmlTokenizer {
     let savedPos = this._savePosition();
     let lowercaseTagName;
     try {
-      this._attemptUntilFn(isNotWhitespace);
+      if (!isAsciiLetter(this.peek)) {
+        throw this._createError(unexpectedCharacterErrorMsg(this.peek), this._getLocation());
+      }
       var nameStart = this.index;
       this._consumeTagOpenStart(start);
       lowercaseTagName = this.inputLowercase.substring(nameStart, this.index);
diff --git a/modules/angular2/test/compiler/html_lexer_spec.ts b/modules/angular2/test/compiler/html_lexer_spec.ts
index 5d3f741156..5e61887eef 100644
--- a/modules/angular2/test/compiler/html_lexer_spec.ts
+++ b/modules/angular2/test/compiler/html_lexer_spec.ts
@@ -174,8 +174,8 @@ export function main() {
             ]);
       });
 
-      it('should allow whitespace', () => {
-        expect(tokenizeAndHumanizeParts('< test >'))
+      it('should allow whitespace after the tag name', () => {
+        expect(tokenizeAndHumanizeParts('<test >'))
             .toEqual([
               [HtmlTokenType.TAG_OPEN_START, null, 'test'],
               [HtmlTokenType.TAG_OPEN_END],
@@ -438,6 +438,9 @@ export function main() {
               [HtmlTokenType.TAG_CLOSE, '</p>'],
               [HtmlTokenType.EOF, ''],
             ]);
+
+        expect(tokenizeAndHumanizeParts('< a>'))
+            .toEqual([[HtmlTokenType.TEXT, '< a>'], [HtmlTokenType.EOF]]);
       });
 
       // TODO(vicb): make the lexer aware of Angular expressions