Founding MENA Show
Post on TikTok, InstagramViews count from Saudi Arabia, UAE, KW, QA, BH and OM
- CPM
- $2.00 per 1K views
- LASTS ABOUTLASTS
- Budget lasts about 16 days
Interviews from theCUBE at Supercomputing 2026 on AI infrastructure. Dell Technologies CTO Ihab Tarazi covers liquid-cooled networking and distributed AI inference, and Preet Virk, Andy Mills and Kevin Cochrane, CMO of Vultr, also appear. Cut one idea into a clip for Instagram, TikTok, X and YouTube.
PER CLIP
WHY THIS SCORE
Beats 71% of clipping campaigns
≈$9 per clip at 10K views
$2,429 of $2,500 is left, spent at about $59 a day.
Content Rewards reports $409,451 paid out on 221 campaigns since 1 Oct 2026, with per view pay in rounds about every 10 days.
Brief B: solid, 4 red flags (missing essentials, no ad disclosure, no deadline and rules changed later), clip must deliver one clear takeaway. The brand shares two footage links and two other links, and sets one content rule on format and AI content.
More format rules
2 podcasts, 40 minutes of talk to cut into clips.
About 36 minutes a clip, finding and cutting.
THE SOURCES
Tarazi says GPUs now carry HPC workloads while researchers run AI workloads, with tens of thousands to hundreds of thousands of GPUs sitting on a single fabric. He argues the network cooling matches Dell's GPU liquid cooling so operators avoid building a second air-conditioned data center, and claims the right software stack adds 60 to 70 percent more inferencing performance.
He adds that new models run 8 to 3 trillion parameters and need 16 to 72 GPUs rather than one or eight.
Covers the first 15 of 18 minutes.
WHO
WHY WATCH
Tarazi gives theCUBE a mini scoop on an unannounced product, Dell's first liquid-cooled network system, and claims it cuts rack assembly from weeks to hours.
In the first 15 minutes, Praveen Vishwakarma pitches a Penguin Solutions and Marvell appliance that moves HBM memory tens of meters away from the GPU and shares it across multiple racks. Andy Mills shows the module up close, with fiber pigtails coming straight into the HBM chip and DDR5 memory catching the overflow.
Vishwakarma claims more than 80% of scale-up network traffic is memory transactions and that optics could cut data movement energy to a quarter of copper's, and explains that a shared pool lets a new model point at an existing KV cache instead of copying it.
Covers the first 15 of 22 minutes.
WHO
Ehab Tarazi reveals Dell's first liquid-cooled network system in a theCUBE interview at SC26
Praveen Vishwakarma pitches a photonics HBM memory appliance shared across racks
#1 X 6.8K views
#2 X 6.7K views
#3 TikTok 5.9K views
By @spolo26
#4 X 3.5K views
By @nathannwil
#5 X 2.8K views
By @adam79977
#6 X 2.8K views
#7 X 2.7K views
#8 X 2.3K views
By @greatkllip
#9 X 1.5K views
By @adam79977
EVERY CUT
Brand footage18:12HORIZONTAL
Dell built the first liquid-cooled network
Tens of thousands of GPUs on one fabric
Compute, network, storage: all liquid-cooled
Brand footage22:02HORIZONTAL
The hidden traffic eating your data center power
What if memory lived meters from the GPU?
One box holds 16 terabytes of GPU memory
Next round, rate rises, top-upsGet alerts
Join on Content Rewards and post on TikTok, Instagram, YouTube or X. On Content Rewards, per view pay comes in rounds about every 10 days, until 30 days after the clip is submitted.
$2,429 of a $2,500 budget is left at 23:46 UTC on 19 Nov 2026.
Views on 4 apps count: TikTok, Instagram, YouTube and X. The rate is $2.00 per 1,000 views on YouTube and X and $1.00 on TikTok and Instagram.
A clip is paid once it has earned at least $3.
Post on TikTok, InstagramViews count from Saudi Arabia, UAE, KW, QA, BH and OM
Post on TikTok, Instagram, YouTubeAt least 40% of viewers in US, Canada, UK and Australia

Post on TikTok, Instagram, YouTube
Post on TikTok, Instagram, YouTube, XAt least 50% of viewers in US