Ep. 174 | AI Just Learned to Actually Watch Your Videos

TwelveLabs raised $100 million in a Series B round co-led by NEA and NAVER Ventures, with participation from Amazon and Red Bull Ventures. The company is building video superintelligence — AI that perceives, understands, and reasons about existing video footage rather than generating new video from text prompts.

Michael and Frank break down the distinction between video generation (Sora, Veo) and video understanding (TwelveLabs). Most businesses generate video footage they never analyze — security archives, training recordings, customer interactions, event footage. TwelveLabs aims to convert video from a storage cost into a searchable, analyzable strategic asset.

They deliver a three-part framework: inventory your video assets and assess their potential analytical value, distinguish between generation needs (marketing) and understanding needs (archives), and start with a pilot project on one high-value video archive before committing to platform-wide deployment.

Topics: TwelveLabs · Video Superintelligence · Video Understanding · Pegasus · AI Video Analysis · Small Business Video · Security Footage AI · Video Archives · AI Pilot Projects · Enterprise Video Intelligence · AWS Trainium

---

Frequently Asked Questions

What is TwelveLabs and what does it do?
TwelveLabs builds AI systems that understand and reason over video content. Its Pegasus model can ingest up to two hours of continuous video and answer complex questions about what is happening across pixels, audio, text, and motion. Unlike video generation tools (Sora, Veo), TwelveLabs focuses on making existing video footage searchable, analyzable, and actionable.

How can small businesses use video understanding AI?
Businesses can analyze security footage for operational patterns, search training video archives for specific content, extract insights from customer interaction recordings, and convert video storage costs into searchable knowledge bases. The key is matching the technology to businesses with meaningful video volume — occasional video may not justify the processing cost.

What is the difference between video generation and video understanding?
Video generation (Sora, Veo, Kling) creates new video from text or image prompts. Video understanding (TwelveLabs) analyzes existing video footage to extract insights, answer questions, and identify patterns. For marketing content needs, generation tools are appropriate. For extracting value from accumulated video archives, understanding tools are the right category.

---

About the Hosts

Michael is a small business owner and entrepreneur since 1983, founder of Cadenhead Services and 850 Media. He speaks from four decades of real operational experience — not whitepapers.

Frank is an AI — an OpenClaw-powered agent serving as Digital Media Director at 850 Media. An AI co-hosting a show about AI for business owners is not a gimmick. It is a live demo of exactly what the show is about.

Transcript

SPEAKER_00 0:20

I'm Michael, a small business owner and entrepreneur since 1983, founder of Cadenhead Services and 850 Media. I speak from four decades of real operational experience, not white papers. This is Control AI Profit. And this week, a company raised $100 million to teach AI to actually watch video.

SPEAKER_01 0:43

12 labs raised $100 million in a Series B round, co-led by NEA and NAEV Ventures, with participation from Amazon, Radical Ventures, and Red Bull Ventures. The company is building what it calls video superintelligence, AI systems that perceive, understand, and reason about video rather than simply generating it.

SPEAKER_00 1:06

This distinction matters. Most video AI news covers generation tools like Sora, Veo, or Cling, tools that create video from text prompts. 12 Labs does the opposite. It takes existing video footage and makes it searchable, analyzable, and actionable.

SPEAKER_01 1:23

The technology is called Pegasus. It can ingest up to two hours of continuous video, understand what is happening across pixels, audio, text, and motion, and answer questions about the content. Not just find the scene with the red car, but identify every instance where a customer interacts with product packaging and summarize the common behavioral patterns.

SPEAKER_00 1:45

For small businesses, the implications are significant because most businesses generate video footage they never analyze, security camera archives, training session recordings, customer service calls captured on video, marketing footage from events, product demonstration videos. All of this sits in storage as a cost rather than an asset.

SPEAKER_01 2:09

12 Labs is betting that businesses will pay to convert video from a storage liability into a searchable knowledge base. The AWS partnership is telling here 12 Labs optimized its inference for AWS trainium chips and will launch new models on AWS first. Amazon sees video understanding as an enterprise service worth investing in.

SPEAKER_00 2:31

Here is my framework for small business owners. First, inventory your video assets. Most businesses do not know how much video they have, where it is stored, and what it might be worth if it were searchable. Security footage alone, if analyzed for patterns, could reveal operational insights about customer flow, employee behavior, and safety risks. Second, second, distinguish between video generation and video understanding. If your business needs marketing content, Sora and its competitors are the right tools. If your business needs to extract value from existing video archives, 12 labs and similar understanding platforms are the right category.

SPEAKER_01 3:15

Third?

SPEAKER_00 3:16

Third, evaluate whether your video volume justifies the cost. Video understanding AI is compute intensive. Processing two hours of footage requires significant processing power. For a small business with occasional video manual review may still be cheaper. For a business with thousands of hours of accumulated footage, the automation value is clear.

SPEAKER_01 3:37

The Korea angle is notable. 12 Labs operates from San Francisco and Seoul, and NAV Ventures co-led the round. South Korea has invested massively in AI infrastructure. We covered the nearly $900 billion sovereign AI plan in an earlier episode. Korean investors backing video understanding suggest this technology is viewed as strategically important for media, entertainment, and corporate intelligence.

SPEAKER_00 4:04

Small businesses should also consider the competitive angle. If your competitors are analyzing their customer interaction videos to identify failure points in service delivery, and you are storing your videos in an archive that nobody watches, you are seeding an advantage. Video understanding levels the playing field between large enterprises with dedicated analytics teams and smaller businesses that can now buy the capability as a service.

SPEAKER_01 4:30

The Red Bull Ventures participation is also a signal. Red Bull generates enormous amounts of event and sports footage. They are not investors in AI for philosophical reasons. They are investors because they believe video understanding will extract more value from their content catalog than traditional editing and manual tagging can achieve.

SPEAKER_00 5:48

Years of unstructured data that is theoretically valuable but practically unusable.

SPEAKER_01 5:54

Because video is not just content, it is a record of everything your business does. And until now, nobody could read it at scale.

SPEAKER_00 6:02

That's it for this week. I'm Michael, and this is Control AI Profit.

SPEAKER_01 6:08

Frank is an AI, an open claw powered agent serving as digital media director at 850 Media. An AI co hosting a show about AI for business owners is not a gimmick. It is a live demo of exactly what the show is about. See you in the next one.