Insights:
Our competitors mainly seem to be operating in a monthly subscription basis, with some also offering pay-as-you-go options or the ability to purchase additional credits/hours. The main differences seem to be related to customer base and usage, with DALL-E being ideal for platform integration with its public API whereas Midjourney excels for professional artists/visual creators who want more control and ability to finetune images. Common complaints with the AI options include bias due to the training datasets and number of iterations required to reach the desired image. People seem to prefer them over stock images due to the efficiency and the ability to customize the images to very specific use cases.
- Enterprise demand = legally safe AI
- Buyers will adopt AI imagery when it comes bundled with indemnification and a clear way to compensate contributors. That combo is now set by Shutterstock and Getty, which launched indemnified, commercially safe gen-AI tools trained on licensed libraries and backed by contributor funds/revenue sharing.
- UX + trust gap in today’s tools
- Designers report stock sites are being flooded with (often unlabeled) AI content, degrading search quality; they want hard filters/labels and higher curation. Meanwhile, AI generators like Midjourney have been powerful but historically frictiony (Discord-first, now adding web), signaling unmet needs for a simple, unified search + generate experience with strong provenance signals.
- Workflow bifurcation
- Most creatives now use gen-AI for speed and ideation, but still look to authentic/real-world assets where brand trust, realism, or editorial standards matter, while legal/ethical worries hinder broader adoption. A winning product should support both modes (AI for generic/illustrative needs + authenticity-certified pools) with transparent rights and guardrails.
Users currently prefer real photos but are bullish on AI-generated outputs in the long run. In terms of our advantage, traditional stock’s credibility still holds strong. The main bottleneck is finding specific and stylistically aligned images. Current platforms like Google Images, Pinterest, and Freepik involve too many clicks and barriers, so the discovery experience is still missing. As for AI, it’s helpful for ideation but not yet reliable or accurate enough.
Competitive Analysis:
DALL-E
Around $0.04 per 1024×1024 image generated using the standard engine and ranging from $0.01 to $0.17 based on image quality, but subscribers to ChatGPT receive a certain amount of credits for image generation as well. The business model is similar to a subscription model since most users pay for a monthly subscription to receive a fixed amount of credits but it is also possible to buy additional credits. The customer base for DALL-E varies broadly, as it is arguably one of the most popular AI image generators. On the business side, it mainly consists of marketers, advertisers, small businesses, large enterprises, game developers and designers, artists, and e-commerce brands. You also get commercial rights to images you generate, meaning they are yours to sell/merchandise etc.
Midjourney
Midjourney is priced similarly at a monthly subscription model as well, but instead factors in the number of hours the GPU takes to generate the images. The tiers also differ in terms of GPU speed and the number of hours on GPU. For companies with rev > $1m, they need to get pro/mega tier to have commercial use. There is also no official public API unlike DALL-E. Mostly geared towards professional/aspiring visual creators and artists, who value being able to fine tune images more and a distinct aesthetic. Content generated is also public by default unless specified otherwise.
Secondary Research
- Shutterstock: AI generator + indemnification + Contributor Fund (official blog/press)
- Launched an AI generator with enterprise indemnification and a Contributor Fund tied to licensed training data and AI usage → positioned as “safe to use at scale.”
- Getty Images + NVIDIA: licensed, indemnified generative AI (official press release)
- Licensed-library-trained generative model with commercial indemnification and a creator revenue-share scheme.
- Adobe: 2024 Creative/GenAI usage survey (official report)
- Broad adoption of gen-AI among creatives for speed/experimentation; AI is now a routine part of many workflows.
- PetaPixel/CreativeBloq: designers say AI is making stock search unusable (labeling/quality)
- Designers report stock sites are increasingly “unusable” without strong AI labels/filters; search quality is degraded by large volumes of AI assets.
- MicrostockGroup forum: practitioner thread on AI content flooding stock sites
- Practitioners describe oversupply of AI uploads and weakened quality control; repeated calls for better labeling, filtering, and stricter review.
- US Copyright Office guidance on human authorship (AI works & registrability)
- Reaffirms human authorship requirement; purely AI-generated outputs aren’t copyrightable → driving enterprises toward indemnified, licensed platforms.
Link: https://docs.google.com/document/d/15yHmVGn0-kkwayYndIpqfmYwYqWhjDNRVuV7AKp2Y44/edit?tab=t.0
