1
00:00:00,040 --> 00:00:09,920
so I want to buy a GPU let's say a GTX 1660 how is this the top page of results

2
00:00:06,399 --> 00:00:16,960
on newegg.com I mean seriously who is

3
00:00:09,920 --> 00:00:19,080
sirona MIL C manie 51 risk canology I

4
00:00:16,960 --> 00:00:24,400
haven't even heard of half of these Brands and Amazon is flooded with them

5
00:00:21,519 --> 00:00:30,039
too I mean the prices look great we're talking as much as 40% off of a

6
00:00:27,920 --> 00:00:35,480
recognized brand and they're still shipped by Newegg but the last time I

7
00:00:32,599 --> 00:00:40,680
checked NVIDIA isn't selling gpus at a discount so what are they refurbished

8
00:00:37,800 --> 00:00:44,520
crypto gpus are they outright scams like we've covered in the past where the

9
00:00:42,239 --> 00:00:49,960
company pops up defrauds a bunch of customers then vanishes only to appear

10
00:00:46,879 --> 00:00:51,879
again under a new equally random name in

11
00:00:49,960 --> 00:00:57,000
either case with such prominent placement on reputable websites Shoppers

12
00:00:54,879 --> 00:01:02,440
could easily be enticed by the savings and choose one of these over a more

13
00:00:59,199 --> 00:01:04,839
established br brand which I don't know

14
00:01:02,440 --> 00:01:10,320
might not be such a bad thing unless it's a scam I was determined to get to

15
00:01:07,439 --> 00:01:15,159
the bottom of this so I tasked our Labs team with buying and testing as many of

16
00:01:13,439 --> 00:01:18,960
these Rando gpus as they were comfortable with and what they

17
00:01:16,479 --> 00:01:24,680
discovered well let's just say the answers aren't as straightforward as

18
00:01:21,360 --> 00:01:26,520
this segue to our sponsor build Redux

19
00:01:24,680 --> 00:01:30,840
they'll put together and ship your customized fully Decked Out rig right to

20
00:01:28,680 --> 00:01:36,200
your doorstep because your PC can never be too fast or Too Furious insert family

21
00:01:33,960 --> 00:01:39,680
joke check out build Redux at the link below the script doesn't actually call

22
00:01:38,119 --> 00:01:44,439
for me to react to the packaging here but some of these are too funny to

23
00:01:41,600 --> 00:01:47,880
ignore this one is upside down if you turn it around this is just a low

24
00:01:46,240 --> 00:01:52,360
resolution ripoff of NVIDIA's old Founders Edition packaging along those

25
00:01:49,799 --> 00:01:57,680
same lines keny's artwork is original but these talking points are absolutely

26
00:01:54,960 --> 00:02:02,360
ancient 3D Vision this GPU doesn't support SLI and in spite of being one of

27
00:02:00,439 --> 00:02:06,560
the two Brands I do recognize here this isn't exactly confidence inspiring I

28
00:02:04,399 --> 00:02:11,760
have so many questions right now but one thing we know for sure is that based on

29
00:02:08,879 --> 00:02:15,879
the featured status of these gpus they must be as profitable for New Egg as

30
00:02:13,920 --> 00:02:21,840
selling a more expensive card from a premier brand but how could that be well

31
00:02:19,200 --> 00:02:27,400
here's the thing shipped by New Egg isn't the same thing as sold by Newegg

32
00:02:25,440 --> 00:02:32,000
like many other online retailers these days newegg.com is actually a

33
00:02:29,920 --> 00:02:35,879
marketplace now that allows other businesses to sell items through their

34
00:02:34,360 --> 00:02:38,920
platform New Egg collects and distributes payments manages

35
00:02:37,440 --> 00:02:43,560
Communications between buyers and sellers and perhaps most importantly

36
00:02:41,480 --> 00:02:47,599
trades on their position as a known entity effectively lending their

37
00:02:45,640 --> 00:02:52,280
credibility to little sellers that you've never heard of some marketplaces

38
00:02:49,560 --> 00:02:55,840
also offer warehousing shipping and even customer support so that sellers don't

39
00:02:54,000 --> 00:03:00,800
need to maintain their own facilities or staff the fees that these marketplaces

40
00:02:58,280 --> 00:03:04,159
charge for these services quickly add up to more than the profit that they would

41
00:03:02,519 --> 00:03:09,400
earn by selling the same product themselves want proof look no further

42
00:03:06,440 --> 00:03:18,799
than our sample search we go sold by New Egg W we're literally three of the more

43
00:03:14,720 --> 00:03:22,000
than 15 results are actually sold by New

44
00:03:18,799 --> 00:03:24,239
Egg themselves and I'm not counting this

45
00:03:22,000 --> 00:03:27,760
same card but with a game bundle now obviously New Egg still wants to move

46
00:03:26,080 --> 00:03:31,280
their own stock out of the warehouse before it gets too old so those three

47
00:03:29,360 --> 00:03:36,400
list things do appear toward the start of the results but this is nothing like

48
00:03:34,319 --> 00:03:41,200
the old days when they would stock loads of gpus from all the major manufacturers

49
00:03:39,000 --> 00:03:47,000
to ensure a consistent customer experience in total then we ordered

50
00:03:44,040 --> 00:03:50,280
eight cards from these unfamiliar Brands as usual we're going to have them Linked

51
00:03:48,239 --> 00:03:53,480
In the video description below but I would watch the rest of the video before

52
00:03:51,680 --> 00:03:59,079
you run out and buy them you see in spite of our search being GTX 1660 we

53
00:03:56,840 --> 00:04:06,519
ended up with only one of those since New Egg mostly returned results for 1660

54
00:04:02,280 --> 00:04:08,360
supers and also this lowprofile 1650 so

55
00:04:06,519 --> 00:04:12,000
we picked that up as well just to see what was going on with that of these

56
00:04:10,000 --> 00:04:17,239
cards most of them arrived as quickly as you would expect with the exception

57
00:04:13,840 --> 00:04:19,239
being this 1660 super from Elsa that

58
00:04:17,239 --> 00:04:23,919
took so long to arrive that we actually opened a support case looking for a

59
00:04:20,880 --> 00:04:26,160
refund it did get here but not in time

60
00:04:23,919 --> 00:04:30,000
for the full test Gauntlet since it was actually only a few hours ago that might

61
00:04:28,000 --> 00:04:34,280
not matter though as it turns out more on that later our first stage was the

62
00:04:32,280 --> 00:04:38,120
physical inspection the lab's team checked the cards over for signs of

63
00:04:36,000 --> 00:04:42,840
where with special attention paid to the PCIe connectors the fan bearings and the

64
00:04:40,960 --> 00:04:46,960
thermal pads along with a quick evaluation of other factors like heat

65
00:04:44,840 --> 00:04:51,080
sink design and overall cleanliness based on that physical examination we

66
00:04:48,919 --> 00:04:56,759
found that there were only three cards here that we can confidently say are new

67
00:04:53,360 --> 00:05:00,240
those are the 1660 supers from yes and

68
00:04:56,759 --> 00:05:02,919
milc along with our 1650 from from

69
00:05:00,240 --> 00:05:07,680
sirona these all show no signs of previous use they have good strong fan

70
00:05:05,280 --> 00:05:12,639
bearings and nice new thermal pads the 1660 supers both included warranty cards

71
00:05:10,320 --> 00:05:16,440
and the 1650 even had a surprisingly good installation manual in the box now

72
00:05:14,680 --> 00:05:20,280
I had a quick look at the Elsa before we started filming and I didn't see any

73
00:05:18,280 --> 00:05:24,039
obvious signs of where but it wasn't looked over by the Labs team so take

74
00:05:22,360 --> 00:05:28,600
that for what it is as for the rest of them it turns out that none of these are

75
00:05:26,520 --> 00:05:33,840
quite what they claim to be first off we've got the kology 1660 super that

76
00:05:31,560 --> 00:05:40,039
appears to be a totally different design from the corn brand one that we ordered

77
00:05:37,080 --> 00:05:44,919
though we did spot the same canology stickers on the corn product page make

78
00:05:42,919 --> 00:05:48,600
of that what you will but maybe it doesn't matter because the Labs believes

79
00:05:46,639 --> 00:05:52,840
that this card may have been refurbished due to wear on the PCIe pins oily

80
00:05:50,840 --> 00:05:58,160
residue from the thermal pads and what seems to be some uneven wear in the fan

81
00:05:55,319 --> 00:06:05,680
bearings next we've got a pair of cards from manna jie a base model 1660 and a

82
00:06:02,199 --> 00:06:07,639
1660 super the base 1660 in particular

83
00:06:05,680 --> 00:06:13,199
has some concerning indicators this sticker here above the PCIe connector

84
00:06:10,039 --> 00:06:15,199
for example is yellowed and starting to

85
00:06:13,199 --> 00:06:19,800
fade compared to the other sticker that's on the back of the card there's

86
00:06:16,880 --> 00:06:26,199
also wear to the PCIe pins and most importantly it looks exactly like the

87
00:06:22,400 --> 00:06:28,759
iafe GTX 1660 from colorful a brand that

88
00:06:26,199 --> 00:06:33,360
I have heard of before except they've what maybe pulled off the back plate and

89
00:06:31,280 --> 00:06:38,240
put new stickers on the fans man's second attempt isn't much better either

90
00:06:35,680 --> 00:06:45,759
this 1660 super shows similar signs of wear and wait a second this looks

91
00:06:41,960 --> 00:06:48,160
exactly like the Elsa these two cards

92
00:06:45,759 --> 00:06:53,120
are completely identical even down to the positioning of the QC sticker on the

93
00:06:51,000 --> 00:06:58,479
back of the cards they're even stamped for the same month these may have come

94
00:06:54,960 --> 00:07:00,360
out of the exact same badge and even

95
00:06:58,479 --> 00:07:05,240
more incriminating both of them happen to look an awful lot like a Galax GTX

96
00:07:03,160 --> 00:07:09,639
1660 super that we found over on Tech PowerUp though with a much lower cost

97
00:07:07,879 --> 00:07:15,599
heatsink design Under The Identical plastic shroud last up we've got this

98
00:07:12,360 --> 00:07:19,120
1660 super from 51 risk it too shows

99
00:07:15,599 --> 00:07:21,479
signs of lightwear but 51 risk unlike

100
00:07:19,120 --> 00:07:25,240
the others actually acknowledges this with a note about it in their manual

101
00:07:23,440 --> 00:07:30,599
stating that the wear on the card is due to their extensive testing processes

102
00:07:28,160 --> 00:07:35,919
seems a little sus especially given that the cooling setup is almost identical to

103
00:07:33,360 --> 00:07:40,280
the base model manui 1660 that we looked at a moment ago just with different

104
00:07:38,240 --> 00:07:46,240
coloring on the Shroud or rather a lack of it so I'm willing to bet that like

105
00:07:43,440 --> 00:07:50,680
the manie colorful is the actual manufacturer of this card in summary

106
00:07:48,400 --> 00:07:55,400
then three of our eight cards are almost certainly brand new one is probably

107
00:07:52,800 --> 00:08:00,319
brand new and of the other four the sellers appear to have simply slapped

108
00:07:57,280 --> 00:08:02,560
their logos onto a possibly refurbished

109
00:08:00,319 --> 00:08:06,520
cart made by someone else and in some cases didn't even manage to get the

110
00:08:04,520 --> 00:08:10,520
brand on the site consistent with the sticker on the cardart now it's time to

111
00:08:08,560 --> 00:08:16,400
pop these into a machine and see what software can tell us about their

112
00:08:12,960 --> 00:08:18,520
origin and yeah not a lot of these

113
00:08:16,400 --> 00:08:25,120
fields change as we scroll through our 1660 supers they're all clocked at 1530

114
00:08:22,199 --> 00:08:29,319
MHz with 1785 MHz boost and other than the BIOS version the variables seem to

115
00:08:26,960 --> 00:08:33,800
be pretty much the brand of the RAM and the subvendor where we get MSI which

116
00:08:31,879 --> 00:08:39,800
makes sense that's our control card along with a mix of colorful Galaxy and

117
00:08:37,240 --> 00:08:44,039
generic NVIDIA this combined with our mismatched BIOS versions seems to

118
00:08:41,680 --> 00:08:49,760
confirm our refurb and Rebrand hypothesis for the cards from manaji 51

119
00:08:46,880 --> 00:08:54,040
risk and kinology now to be clear a generic identifier isn't suspicious on

120
00:08:52,040 --> 00:08:57,839
its own Elsa for example has been around for a long time and appears to contract

121
00:08:56,320 --> 00:09:02,000
their manufacturing out to Other Board makers which might explain their

122
00:08:59,839 --> 00:09:05,480
non-unique designs of course for many buyers anyway this is all irrelevant and

123
00:09:04,440 --> 00:09:12,880
the only thing that matters is performance I mean who cares what stickers on it right so we tested every

124
00:09:10,240 --> 00:09:20,240
card again except the Elsa along with two controls an MSI 1660 super and this

125
00:09:16,440 --> 00:09:22,760
ASUS 1650 with a careful eye for any

126
00:09:20,240 --> 00:09:26,880
observed anomalies surprisingly in gaming the 1660 supers were very

127
00:09:25,240 --> 00:09:31,440
consistent across the board with only the milc lagging slightly behind the

128
00:09:28,920 --> 00:09:35,600
rest of the pack the manui 1660 meanwhile remember that's the rebadged

129
00:09:33,160 --> 00:09:41,079
colorful was also just about able to keep up with the supers a pretty good

130
00:09:37,399 --> 00:09:42,959
showing but the sorria 1650 ended up

131
00:09:41,079 --> 00:09:46,839
performing considerably worse than our ASUS control though that kind of makes

132
00:09:45,320 --> 00:09:51,440
sense if you consider how much less power it drew power consumption for our

133
00:09:49,079 --> 00:09:55,240
1660 supers meanwhile was quite consistent which makes sense given their

134
00:09:53,320 --> 00:09:59,720
performance moving on the spc.org productivity tests we ran showed similar

135
00:09:57,279 --> 00:10:04,519
results with the 1660 supers being very consistent with each other and our

136
00:10:01,440 --> 00:10:07,279
control and only the mulc lagging behind

137
00:10:04,519 --> 00:10:13,600
while our manui 1660 does an admirable job trying to keep up with the supers

138
00:10:09,279 --> 00:10:15,120
and once again the sir HRA 1650 is

139
00:10:13,600 --> 00:10:19,720
underwhelming in the test that it managed to complete and just straight up

140
00:10:17,480 --> 00:10:24,200
fails to complete the solid works test so wherever they're getting these gpus

141
00:10:22,040 --> 00:10:28,880
from they're legit they're made by NVIDIA you can't counterfeit that which

142
00:10:26,360 --> 00:10:33,839
raises the question well why not buy them then to answer we devised a final

143
00:10:31,600 --> 00:10:38,600
test where we reached out to each seller with a simple support request and we

144
00:10:36,279 --> 00:10:42,680
were pleasantly surprised to get fast helpful responses from each and every

145
00:10:41,000 --> 00:10:48,160
one of them that's obviously not a guarantee that you're going to get proper warranty support 11 and a half

146
00:10:46,040 --> 00:10:53,160
months down the road but I'm a bit too honest to just pry capacitors off these

147
00:10:51,120 --> 00:10:56,600
things with my noct to Edition LTT screw driver to see how they deal with a dead

148
00:10:54,600 --> 00:11:00,519
card by the way sign up for a stock notification on this thing now at LTT

149
00:10:58,519 --> 00:11:05,279
store.com with the average price of a 1660 super from Brands like ASUS MSI and

150
00:11:03,519 --> 00:11:11,079
Gigabyte hovering around $250 then are these lesser known Brands

151
00:11:08,480 --> 00:11:15,160
worth the risk I mean I'm not going to give an outright thumbs up to any of the

152
00:11:13,079 --> 00:11:21,200
ones that appear to be just slapping a sticker on someone else's possibly used

153
00:11:17,560 --> 00:11:24,399
product so that rules out manaji 51 risk

154
00:11:21,200 --> 00:11:29,760
and corn SL kology whoever you are

155
00:11:24,399 --> 00:11:32,839
leaving us with yes Sira and milie S ya

156
00:11:29,760 --> 00:11:34,839
1650 underperforms and is priced higher

157
00:11:32,839 --> 00:11:39,320
than some of the 1660 supers so unless you absolutely need a half height single

158
00:11:36,600 --> 00:11:44,320
slot card that's a big fat nope and of the two that remained milc did

159
00:11:41,040 --> 00:11:47,200
underperform a bit but costs far less

160
00:11:44,320 --> 00:11:51,360
while yes performed as expected includes cute stickers but comes at a higher

161
00:11:49,760 --> 00:11:54,079
price these guys have also been around for at least 20 years though which gives

162
00:11:52,760 --> 00:11:58,639
us some confidence that they'll be around to support us if something goes

163
00:11:55,880 --> 00:12:02,360
wrong however all that said with more trans transparency around the

164
00:12:00,000 --> 00:12:06,399
refurbished status of these other cards almost everything we saw today was a

165
00:12:04,360 --> 00:12:10,360
pretty okay value at the prices that they're charging you just have to decide

166
00:12:08,240 --> 00:12:14,360
how much or how little you want to gamble but you know you're never

167
00:12:11,760 --> 00:12:18,120
gambling with our sponsor delete me whether you like it or not your personal

168
00:12:16,000 --> 00:12:21,680
information is out there and companies are constantly sharing it without your

169
00:12:19,880 --> 00:12:25,440
consent this info can be used by robocallers and email spammers and it

170
00:12:23,800 --> 00:12:29,760
could even lead to identity theft or fraud it's a complicated problem

171
00:12:27,680 --> 00:12:33,600
thankfully the solution is easier with delete me delete me will help you find

172
00:12:31,680 --> 00:12:37,600
hundreds of online profiles sharing your personal data and remove it this would

173
00:12:35,800 --> 00:12:41,760
be a huge task that could take many hours doing it yourself but with delete

174
00:12:39,600 --> 00:12:45,880
me's Advanced tools and team of experts it can be done in minutes on average

175
00:12:43,600 --> 00:12:49,360
delete me finds and removes over 2,000 pieces of data for a customer in their

176
00:12:47,480 --> 00:12:52,519
first two years so if you want to get your personal information removed from

177
00:12:50,839 --> 00:12:58,199
search results on the web go to join delet me.com Linus techtips and use code

178
00:12:55,360 --> 00:13:03,079
LT for 20% off if you guys like this video why not watch me and Alex try to

179
00:13:00,440 --> 00:13:08,480
take a 4090 to 1,000 wats of power consumption didn't work but we got it a

180
00:13:05,040 --> 00:13:08,480
lot higher than stock