Overview
Imprezia is an active technology startup currently participating in the Y Combinator Summer 2025 batch. The company positions itself as "The World's First AI Ad Network," aiming to leverage artificial intelligence to optimize digital advertising workflows [1][2]. Founded by Bishesh Khadka, a former Meta engineer and MIT alumnus, the venture operates in the highly competitive ad-tech sector, seeking to disrupt traditional programmatic advertising models through automated, AI-driven decision-making [4][5].
The premise that Imprezia is a defunct or failed startup is factually incorrect. As a member of the Summer 2025 cohort, the company is in its early operational phase, actively developing its product and seeking initial market fit. There is no public evidence of shutdown, acquisition, or pivot away from its core thesis. Consequently, a traditional post-mortem analysis of failure is not applicable, as the company has not yet reached a terminal state or demonstrated sufficient longitudinal data to assess long-term viability.
The outcome for founders and investors remains undetermined. Imprezia is currently in the build-and-launch phase typical of early-stage YC companies. The primary risk factors lie in future execution, market adoption, and the ability to differentiate from entrenched incumbents like Google and Meta, rather than past operational failures. This report serves as a foundational profile of the company’s current status, founding context, and strategic positioning, rather than a retrospective on closure.
Founding Story
Imprezia was founded by Bishesh Khadka, whose professional background provides the foundational insight for the company’s value proposition. Khadka is an alumnus of the Massachusetts Institute of Technology (MIT), where he developed the technical rigor necessary for building complex software systems [5]. Following his education, Khadka joined Meta, one of the world’s largest digital advertising platforms, where he worked as an engineer [4]. This experience inside the "walled garden" of a major ad-tech incumbent likely exposed him to the inefficiencies, opacity, and manual friction inherent in current digital advertising ecosystems.
The specific insight that led to Imprezia’s creation stems from the intersection of Khadka’s engineering expertise at Meta and the emerging capabilities of generative AI. While at Meta, Khadka would have observed firsthand how ad targeting, bidding, and creative optimization are handled at scale. However, the broader ad-tech landscape outside of Meta and Google remains fragmented, relying on legacy programmatic infrastructure that often lacks the real-time intelligence of newer AI models. The founding thesis appears to be that AI can not only optimize existing ad buys but fundamentally restructure how ad networks operate, moving from rule-based programmatic buying to predictive, autonomous AI agents.
Public records primarily highlight Bishesh Khadka as the primary founder. While Y Combinator companies typically have two or three co-founders, early directories and public filings for Imprezia predominantly feature Khadka [3]. It is possible that co-founders are either less public-facing or joined the venture after initial incorporation. The lack of widely publicized co-founder narratives suggests that Khadka is the central technical and visionary force behind the initial prototype and YC application.
The company’s formation likely occurred in late 2024 or early 2025, culminating in its acceptance into the Y Combinator Summer 2025 batch [1]. This timeline aligns with the rapid surge in AI-focused startups following the widespread availability of large language model APIs. Khadka’s transition from Meta employee to founder represents a classic "insider-outsider" trajectory: leveraging deep industry knowledge from a dominant player to build a challenger brand that promises greater efficiency or transparency through new technology.
There is no public record of a major pivot prior to the Summer 2025 batch. The company has consistently marketed itself as an AI ad network from its earliest public appearances. This focus suggests that the founders identified a specific, high-value problem in ad-tech—likely related to campaign management, creative generation, or bid optimization—and committed to solving it with AI from day one, rather than iterating from a different initial product.
Timeline
- Pre-2024: Bishesh Khadka attends MIT, building a foundation in computer science and engineering [5].
- 2024 (Early-Mid): Khadka works as an engineer at Meta, gaining direct experience with large-scale advertising infrastructure and algorithms [4].
- Late 2024: Imprezia is incorporated. The founding team begins developing the initial prototype for an AI-driven ad network.
- January 2025: Imprezia is officially listed as part of the Y Combinator Summer 2025 batch, marking its entry into the accelerator program [1].
- Spring/Summer 2025: The company actively participates in the YC program, refining its product and preparing for Demo Day. No public shutdown or pivot events occur during this period.
What They Built
Imprezia describes its core product as "The World's First AI Ad Network" [2]. While detailed technical documentation and public user interfaces are not yet widely available due to the company’s early stage, the value proposition can be inferred from its positioning and the founder’s background. Traditional ad networks rely on programmatic real-time bidding (RTB), where algorithms bid on ad impressions based on predefined rules and historical data. Imprezia aims to replace or augment this with AI agents that can autonomously manage campaigns, optimize creative assets, and predict performance with greater accuracy than legacy systems.
The product likely functions as a platform for advertisers to input their goals, budget, and creative materials. The AI system would then handle the complex task of distributing these ads across various publisher networks, adjusting bids in real-time, and potentially generating or modifying ad creatives to improve click-through rates. This contrasts with traditional demand-side platforms (DSPs) that require significant manual configuration and expertise to operate effectively. By automating these decisions, Imprezia seeks to lower the barrier to entry for smaller advertisers while improving returns for larger ones.
Key features likely include automated bid management, AI-generated ad variations, and predictive analytics for campaign performance. The technology architecture probably leverages large language models (LLMs) for creative tasks and reinforcement learning for bidding strategies. This dual approach allows the system to not only decide where to place an ad but also what the ad should look like to maximize engagement in that specific context.
The differentiation from alternatives lies in the depth of AI integration. While companies like Google and Meta use AI internally, their platforms are walled gardens that prioritize their own inventory. Imprezia positions itself as an independent network, potentially offering access to a broader range of publishers without the bias of a single ecosystem. This independence is a key selling point for advertisers seeking diversification away from the duopoly of Google and Meta.
As an early-stage YC company, the product is likely still in beta or early access. The focus during the Summer 2025 batch would be on validating the core AI models with a small group of initial users, gathering data to refine the algorithms, and demonstrating proof of concept to potential investors. The evolution of the product will depend heavily on feedback from these early adopters, particularly regarding the accuracy of AI predictions and the ease of use compared to established DSPs.
Market Position
Target Customers
Imprezia’s primary target customers are digital advertisers, ranging from small and medium-sized businesses (SMBs) to mid-market enterprises. SMBs are often underserved by complex programmatic platforms, lacking the resources to manage sophisticated campaigns. Imprezia’s AI-driven automation promises to make high-performance advertising accessible to this segment. Mid-market companies, meanwhile, may seek alternatives to Google and Meta to diversify their ad spend and reduce dependency on single platforms. Publishers may also be secondary customers, benefiting from higher-yield ad placements driven by AI optimization.
Market Size
The global digital advertising market is vast, valued at over $600 billion annually. Programmatic advertising, the segment Imprezia targets, accounts for a significant majority of this spend. However, the market is mature and dominated by a few key players. The opportunity for Imprezia lies not in the total market size, but in the subset of advertisers seeking more efficient, automated, and independent solutions. The rise of AI in marketing has created a new sub-segment of "AI-first" ad tech, which is growing rapidly but remains a small fraction of the overall market.
Competition
The competitive landscape for Imprezia is structurally challenging. The company operates in a market characterized by extreme consolidation and high barriers to entry.
Incumbent Advantage: Google and Meta control the majority of digital ad inventory and data. Their platforms benefit from massive scale, proprietary user data, and integrated ecosystems that are difficult for startups to replicate. Google’s Demand-Side Platform (DV360) and Meta’s Ads Manager are deeply entrenched, with features that continuously evolve to incorporate AI. Competing on distribution or data depth is nearly impossible for a new entrant.
Platform Dependency: Many ad-tech startups rely on access to inventory from major platforms. If Google or Meta change their API access rules or prioritize their own tools, independent networks can be squeezed out. Imprezia’s claim of being an "AI Ad Network" suggests it may aggregate inventory from various sources, but its ability to compete depends on securing reliable, high-quality supply.
Competitive Axes:
- Distribution Reach vs. Product Depth: Incumbents win on distribution. Imprezia must compete on product depth—specifically, the superiority of its AI algorithms. If its AI can demonstrably outperform Google’s automated bidding in specific niches, it can carve out a niche.
- User Trust vs. Feature Completeness: Advertisers trust Google and Meta due to track records. Imprezia must build trust through transparency and results. Feature completeness is less critical than efficacy; a simpler tool that delivers better ROI can win over a complex, underperforming incumbent.
- Independence vs. Integration: Imprezia’s key differentiator is independence from the Google/Meta duopoly. This appeals to advertisers worried about platform risk. However, this advantage is only viable if the independent inventory is of comparable quality.
The competitive landscape has shifted with the advent of generative AI. Previously, ad-tech competition was about data scale. Now, it is increasingly about algorithmic intelligence. This shift offers a window for startups like Imprezia to compete on the quality of their AI models rather than just data volume. However, incumbents are also integrating generative AI, meaning Imprezia must innovate faster than well-resourced giants.
Business Model
Imprezia’s revenue model is likely based on a combination of transaction fees and software subscriptions, typical of ad-tech platforms. As an ad network, it may take a percentage of ad spend managed through its platform (a take rate). Alternatively, it could charge advertisers a SaaS fee for access to its AI tools, regardless of spend volume. A hybrid model, with a lower take rate plus a subscription for advanced AI features, is also possible.
Given the company’s early stage, specific revenue figures are not public. However, we can infer directional unit economics. Ad-tech margins can be thin due to the cost of goods sold (COGS), which includes payments to publishers. If Imprezia acts as a pure intermediary, its gross margins will be constrained by the spread between what advertisers pay and what publishers receive. If it positions itself as a SaaS provider, margins could be higher, but customer acquisition costs (CAC) may be elevated due to the need to educate the market on AI benefits.
With Y Combinator backing, the company likely has a runway of 12-18 months. Assuming a typical YC seed funding range and a small team (2-5 people), the annual burn rate is likely between $500,000 and $1 million. To achieve sustainability, Imprezia will need to demonstrate strong retention and increasing ad spend per user. The absence of public revenue data is a signal that the company is still in the product-validation phase, prioritizing growth and technology development over immediate monetization.
Traction
No public traction data, such as active users, ad spend volume, or revenue metrics, is available for Imprezia at this time. As a member of the Summer 2025 YC batch, the company is likely still in the early stages of customer acquisition and product refinement. Publicly available information is limited to its founding team and value proposition.
Post-Mortem
Note: Imprezia is an active company and has not failed. The following section analyzes potential risks and structural challenges that could lead to future failure, based on the current market dynamics and the company’s positioning. This is a predictive risk assessment, not a historical post-mortem.
Structural Challenge: The Duopoly’s Data Moat
The primary risk for Imprezia is the insurmountable data advantage held by Google and Meta. AI models in advertising are only as good as the data they are trained on. Google and Meta possess decades of user behavior data, allowing their AI to predict conversion probabilities with high accuracy. Imprezia, as a new entrant, lacks this historical data. Without access to comparable data sets, its AI models may underperform incumbents, leading to lower ROI for advertisers and churn.
- Specific Event/Risk: If Imprezia’s initial AI models fail to match the performance of Google’s Performance Max or Meta’s Advantage+ campaigns, advertisers will revert to the incumbents.
- Attempted Remedy: Imprezia may attempt to niche down, focusing on specific verticals or publisher types where incumbent data is less dominant.
- Outcome Risk: If the niche is too small, the business cannot scale. If the niche is large, incumbents will likely replicate the feature.
Platform Dependency and Inventory Access
Imprezia’s ability to serve ads depends on access to publisher inventory. If it relies on open exchanges, it competes with thousands of other buyers, driving up costs. If it seeks direct publisher partnerships, the sales cycle is long and resource-intensive.
- Specific Event/Risk: Major publishers may prioritize direct deals with Google or Meta, leaving Imprezia with lower-quality remnant inventory.
- Attempted Remedy: Building a proprietary network of premium publishers.
- Outcome Risk: High customer acquisition costs for publishers and advertisers alike, leading to negative unit economics in the early stages.
The "Feature, Not a Company" Trap
AI-driven optimization is increasingly becoming a standard feature in existing ad platforms. If Imprezia’s core innovation is simply "better AI bidding," it risks being absorbed as a feature by larger DSPs or ad servers.
- Specific Event/Risk: A major DSP integrates a similar AI agent, offering it to existing customers at no additional cost.
- Attempted Remedy: Differentiating through user experience and independence.
- Outcome Risk: Advertisers may prefer the convenience of an integrated solution over a standalone tool, even if the standalone tool is slightly better.
Founder Concentration Risk
With Bishesh Khadka identified as the primary public founder, the company faces key-person risk. The complexity of building an ad network requires expertise in engineering, sales, and operations. If the founding team lacks depth in go-to-market strategy, the product may fail to reach critical mass.
- Specific Event/Risk: Difficulty in hiring senior sales talent to compete with established players.
- Attempted Remedy: Leveraging YC network for hires.
- Outcome Risk: Slow growth and inability to secure key publisher partnerships.
Key Lessons
- Imprezia’s reliance on AI differentiation faces a data barrier: Imprezia launched its AI ad network in 2025, a time when incumbents like Google and Meta already possess vast proprietary datasets. Unlike earlier ad-tech startups that competed on inventory access, Imprezia must prove its AI can outperform models trained on decades of user data, a structural disadvantage that requires niche targeting to overcome.
- Independence from walled gardens is a double-edged sword: By positioning itself as an alternative to Google and Meta, Imprezia appeals to advertisers seeking diversification. However, this independence means it lacks the integrated feedback loops of closed ecosystems. The company must build trust through transparency, as it cannot rely on the brand equity of a major platform to reassure advertisers.
- Timing in ad-tech requires more than just new technology: Imprezia entered the market when AI hype is high, but advertiser budgets are tight. The market rewards proven ROI over experimental technology. Imprezia’s success will depend on its ability to demonstrate immediate, measurable improvements in ad performance, rather than just the novelty of its AI approach.