Thank You for Powering Epilog’s Largest Raise Yet

We’re incredibly grateful for the support — Epilog has raised over $4.2M in this round, making it our largest raise to date, with 2,289+ investors participating. In total, we’ve now raised $9.4M+ to date and have over 7,070+ investors.

This milestone gives us the ability to bring on new talent and expand adoption of our vision platform across multiple verticals, and we couldn’t be more excited about what’s ahead.

Thank you to everyone who invested, shared our story, and supported us along the way. This momentum is just the beginning — we’ll continue sharing regular updates here and through our newsletter. All new investors will be subscribed, and you can also request updates directly at blog.epilog.com.

Best,

Michael

Follow us: © 2025 Epilog

Just Launched: The New epilog.com Website

We just launched a completely new website epilog.com, and you still have about 10 hours to review it before this round closes. 

The site reflects where we stand today, with clearer messaging, improved navigation, and a direct path to our vision and progress. It spotlights our flagship products: SideCar (AI-powered retrofit driver assistance), Vortex (edge AI expanding video storage up to 10×), CloseView (integrated 50MP AI vision platform with patented optical processing), and Wellcome (contactless AI health scanner for remote vital signs).

The new “Where to Buy” section highlights active commercialization and real deployments—hardware is shipping now.

EPILOG’S CAMPAIGN IS CLOSING TODAY!

Our Current StartEngine Will Conclude Today at 11:59PM PST, in 10 hours!
Join a community of 7,000+ investors backing our Vision AI technology.

📢 Epilog’s Raising on StartEngine Is Concluding Today

We’re currently conducting our 4th Regulation Crowdfunding (Reg CF) round on StartEngine, giving our community the opportunity to own a piece of Epilog as we scale our AI vision technology. Learn more and invest here: StartEngine.com/epilog

This Reg CF offering is made available through StartEngine Primary, LLC. This investment is speculative, illiquid, and involves a high degree of risk, including the possible loss of your entire investment.

Follow us: © 2025 Epilog

 $4M Milestone Achieved This Round

We have now exceeded $4M raised in the current offering. Across all Epilog offerings to date, total capital raised now exceeds $9.1 million.

In recent days, investor participation has accelerated, including participation from larger investors. While we are encouraged by this activity, the most important point for anyone still evaluating the offering is that this round has a fixed closing date.

This offering is scheduled to close on January 29, 2026. No extensions are currently planned.

As previously announced, Epilog has begun  entering a commercialization and execution phase. This reflects our ongoing transition from development toward real-world deployment of Physical AI systems.

Since 2015, Epilog has been focused on building AI systems designed to operate in the physical world — systems intended to support vehicles, security environments, and infrastructure through real-world sensing and edge intelligence.

Investors considering participation should be aware that the only way to participate under the current terms is prior to the scheduled closing date. As always, participation is subject to the terms, conditions, and risks described in the offering materials.

- Michael

Follow us: © 2025 Epilog

Imagine AI Vision At Walmart

It's hard to imagine, but Walmart faces lawsuits in the U.S. nearly 20 times per day—stemming from issues like slip-and-fall accidents, injuries from falling merchandise, false imprisonment, wrongful termination, copyright infringement, and pharmacy malpractice.^1​

For an operation of this scale—over 10,000 stores² estimated to generate petabytes of recorded video data daily—the challenge is centralizing and archiving massive volumes of raw vision into retained, actionable evidence while managing skyrocketing costs.

Many retailers limit retention to a few weeks—leaving them defenseless when claims emerge months later without visual proof—or rely heavily on manual processes. Neither scales, and both amplify liability risk.

This is where Epilog's Vortex delivers transformative corporate value through dramatically extended, cost-effective vision retention.

Vortex applies advanced AI compression, is designed to enable organizations to preserve visual records for months or even a full year on existing infrastructure. This extended timeline is game-changing: it ensures critical evidence is available for delayed incidents, claims, litigation, or regulatory reviews—providing superior auditability, stronger compliance, and material liability protection while directly slashing storage and bandwidth expenses.

Epilog's AI-driven vision platform is also designed to enhance efficiency with intelligent analytics that prioritize meaningful events, reducing manual overhead and enabling faster, more accurate investigations.

Together, Epilog technology has the potential to:

• Lower operating costs through more efficient storage and reduced waste

• Improved risk management via extended evidence retention and AI-assisted insights

• Scalable deployment across large distributed vision networks without linear cost growth

This isn’t theoretical: we demonstrated these capabilities live in New York City last November, have since received inquiries from supermarket operators through our dealer network, and are planning a formal product launch on March 23 at ISC West in Las Vegas.

For investors, Epilog is real-world AI vision infrastructure—technology intended to help large enterprises save substantially, operate efficiently, and protect themselves by ensuring visual evidence exists when it's needed most.

Our funding round closes January 29, 2026.

Follow us: © 2025 Epilog

Amazon Invested in Epilog; Could Google Be Next?

In the second quarter of 2025, Amazon Web Services invested in Epilog as part of the Company’s efforts to scale its AI-compression and vision-infrastructure platforms, Vortex and CloseView.

Epilog is advancing technologies focused on reducing the cost and complexity of storing, processing, and analyzing large volumes of visual data. The AWS investment supported this direction by enabling Epilog to expand backend processing capacity, accelerate development cycles, and harden its infrastructure for real-world deployment.

Impact on Our Platform

Following the investment, Epilog was able to:

• Expand development and testing environments supporting Vortex AI compression

• Improve scalability and performance across backend vision workflows

• Accelerate feature development and iteration velocity

• Reduce infrastructure-related cash expenditures

• Advance integrations between edge capture systems (Vortex/CloseView) and backend processing (EXO)

These improvements contributed to greater operational efficiency and supported Epilog’s broader objective of building scalable, energy-efficient vision infrastructure.

Strategic Significance

Epilog views Amazon’s investment as validation of its technical approach to AI-driven video compression and large-scale vision processing. The collaboration reinforced Epilog’s hybrid infrastructure strategy and supported continued progress across multiple vision markets.

Potential for Continued Investment 

Amazon Web Services has shown interest in potentially doubling its investment, subject to further evaluation and conditions. In parallel, Epilog engaged in similar exploratory discussions with Google regarding potential infrastructure support aligned with the Company’s AI-compression and vision-processing roadmap. Any additional investment or commercial arrangement remains subject to further evaluation, technical milestones, and customary approvals, and there can be no assurance that such discussions will result in a definitive agreement.

Michael 

Follow us: © 2025 Epilog

The Big Picture for AI Vision

In 2026, Epilog intends to bring our patented technology out of the lab and factory and into the real world—deploying AI vision at scale across everyday vehicles, parking facilities, airports, logistics centers, city corridors, clinics, and more. True vision isn’t just about seeing; it’s about making intelligent visual understanding effortless, intuitive, and instantly accessible at scale—even as data volumes grow massive, costly, and difficult to search.

Epilog is building a full-stack approach to make visual data more durable, searchable, and useful over time:

• SideCar is designed as an aftermarket, vehicle-based vision platform, bringing AI-enabled assisted driving, recording, and analysis to existing vehicles.

• Vortex is designed to reduce visual data volumes using scene-aware AI compression (targeting 2–10x depending on deployment) with indexed retrieval.

• CloseView is designed for wide-area, high-resolution environments, pairing capture with the same Vortex software layer to support AI-enabled visual forensics.

• Wellcome is under development as a next-generation digital health scanner, with plans for a mid-2026 debut at a health-focused trade show, subject to technical progress and compliance requirements.

• EXO,  in the cloud we’re also developing an agentic layer and a modular compute framework  intended to help automate time-consuming forensic review workflows 

That is the Epilog mission: delivering AI-powered vision for the real world—through intelligent devices that automate tedious tasks and create a safer environment for all.

Reminder: our current Reg CF round is scheduled to close January 29, 2026.

- Michael

Follow us: © 2025 Epilog

Extending Visual Storage With AI

As vision systems increase in resolution and deployment scale, storage costs and compute requirements scale accordingly. Many small and mid-sized organizations manage these constraints by discarding recorded data after relatively short retention periods—often within 7–30 days—which may limit the availability of historical records when incidents, claims, or disputes arise later.

These constraints can have practical implications across commercial environments. In certain circumstances, limited data may complicate post-incident review related to safety claims, asset loss, or operational disputes. According to third-party industry estimates, the broader visual intelligence market, primarily encompassing video surveillance, is projected to range from approximately $82 billion in 2025 to $163 billion by 2030. This reflects a projected compound annual growth rate of approximately 12%.¹

AI-Based Compression

Epilog’s Vortex Vision AI OS is designed to address these challenges by enabling more efficient visual data storage and management. The system operates across a range of hardware configurations, including AMD and NVIDIA-based platforms (Vortex, Closeview, Wellcome and SideCar) and incorporates proprietary, scene-aware AI compression techniques designed to reduce data volumes by 2–10x. In initial pilot deployments, these techniques have shown the potential to extend data retention from weeks to multiple months or longer, though results may vary depending on specific deployment circumstances and configurations.

Vortex can also support on-premises or hybrid environments and includes automated indexing features designed to improve the accessibility of archived data, such as:

• Object recognition and classification

• Scene categorization

• Automated event summarization

• Metadata enrichment to support faster search and retrieval

The overall objective of the platform is to support more durable and searchable visual archives, reduce reliance on continuous on-site processing, and improve the efficiency of post-event review workflows. Initial test deployments have shown approximately 4x increases in average data persistence under certain conditions, though results may vary based on deployment configuration, environment, and usage patterns.

Energy Efficiency

Beyond performance and cost advantages, Vortex—typically requiring only ~25 W of on-premises power—is inherently energy-efficient. As illustrated in the diagram below, real-time sensing and control remain at the edge, while high-intensity AI compression and analytics workloads are shifted upstream through a globally connected cloud AWS layer to centralized Epilog compute infrastructure. 

This architecture is intended to minimize on-site power consumption by shifting intensive AI workloads to power-optimized, increasingly renewable-supported data centers. The result is potentially lower carbon impact while maintaining forensic quality and scalability, with the goal of making Vortex smarter, safer, and more environmentally responsible, subject to ongoing evaluation and verification.

Michael

Follow us: © 2025 Epilog

Agentic AI for Actionable Intelligence

Agentic AI is intended to be a class of AI software designed to reason over time, retain contextual information, and coordinate actions across workflows. In visual intelligence applications, agentic AI is intended to operate above individual perception models by aggregating signals, applying rules or learned policies, and supporting higher-level decision processes. Rather than reacting solely to isolated events, agentic architectures are designed to incorporate context, history, and predefined objectives, potentially enabling more structured analysis and operational coordination.

Automating Labor-Intensive Forensic Event Reconstruction

Investigating incidents from archived visual data is traditionally an extremely labor-intensive process for humans. Analysts must manually review hours or days of footage across multiple cameras, identify relevant clips, correlate events across time and locations, reconstruct accurate timelines, and cross-reference external data—all while combating eye strain, human error, and escalating costs. In large-scale deployments (e.g., parking facilities, transportation hubs, or urban corridors), a single incident can require tens to hundreds of person-hours, delaying resolutions and increasing operational overhead.

Epilog aims to address this challenge with an Agentic AI layer, which is intended to be primarily integrated into Vortex Software. This layer is expected to automate complex event reconstruction and potentially reduce the manual effort historically required.

Business Model

A typical video-based incident investigation can be costly due to the labor-intensive review of large volumes of archived records across multiple locations and time periods. Professional forensic analysis commonly involves hourly rates of approximately $200–$450, with a single incident often requiring 15–30 hours or more, resulting in total costs that can reach several thousand dollars per case, in some instances exceeding the cost of Vortex equipment and associated software subscriptions.¹  

In the Epilog model, customers invoke agentic AI services on demand for specific investigations. The software performs reasoning, cross-location and time-based searches, event correlation, and timeline reconstruction, with optional integration into external tools. Processing is scheduled through Epilog’s EXO compute framework, with usage-based pricing tied to actual compute, data access, and processing, aligning costs with investigative needs.

Infrastructure and Roadmap

EXO, Epilog’s modular, open-standards-compatible framework, offloads intensive workloads to efficient cloud, private, or hybrid resources (currently in development). Internal prototypes include enhanced indexing, scalable search backends, and early LLM-driven agents. Epilog intends to plan initial demonstrations of agentic capabilities and EXO integration around 2026, with the aim to phase the software and services into Vortex and CloseView as primary platforms.

Market Dynamics in Visual Forensics and Intelligence

The market for visual forensic technologies is expanding as visual data plays an increasingly important role in security, compliance, and operational analysis across public and private infrastructure. The broader visual intelligence market—primarily encompassing video surveillance systems—is estimated to range from approximately $82 billion in 2025 to $163 billion by 2030, reflecting projected compound annual growth rates of approximately 12%, based on third-party industry estimates. 

Michael

Follow us: © 2025 Epilog

Visual Forensics for Critical Infrastructure

Critical infrastructure—including commercial parking facilities, transportation systems, large venues, utilities, logistics centers, and urban corridors—relies on continuous visual data to support safety, accountability, and operational awareness. As deployments scale, legacy systems have faced cost-performance trade-offs that can lead to reduced coverage and the loss of valuable data.

Vortex and CloseView are designed to address these challenges by enabling high-quality, AI-enabled vision across both fixed-location and wide-area environments. The platforms aim to preserve visual information in a manner that may support forensic review, operational analysis, and compliance-related use cases over extended spatial and temporal horizons, depending on deployment conditions and system configuration.

From Visual Data to Forensic Infrastructure

Vortex and CloseView support a transition from passive visual collection to active visual forensics, with the objective of making visual data more searchable, durable, and cost-effective as an evidentiary resource for critical infrastructure environments.

AI-based processing is designed to enhance the potential utility of visual data by improving clarity, retention efficiency, and analytical accessibility. Techniques such as noise reduction and resolution optimization are intended to support improved identification performance and reduce false positives, subject to data quality and environmental conditions.

Intelligent compression is designed to increase effective retention capacity on existing storage infrastructure at comparable cost. Continuous indexing and improved image quality are intended to support faster post-incident review and reconstruction.

Automation is intended to reduce manual review requirements and support scalable deployment across distributed environments. The platforms are designed to support tamper-resistant recordkeeping and configurable privacy controls for legal, regulatory, and insurance-related review, depending on customer implementation. Vortex was recently showcased at the NYC Cooperator Expo, and an industry debut is planned at ISC West in Las Vegas in March 2026.

Market Dynamics in Visual Forensics and Intelligence

The market for visual forensics is expanding as visual data plays an increasingly important role in security, compliance, and operational analysis across public and private infrastructure. The broader visual intelligence market—primarily encompassing video surveillance systems—is estimated to range from approximately $82 billion in 2025 to $163 billion by 2030, reflecting projected compound annual growth rates of approximately 12%, based on third-party industry estimates.

Michael

Follow us: © 2025 Epilog

Happy New Year From All Of Us At Epilog

A heartfelt thank you to every investor who joined us, amplified the round, or challenged us with great questions—by investing, you affirmed our bold bet on Vision AI. Our primary goal for 2026 are aiming to scale our commercialization efforts in a measured and disciplined manner, aiming to increase our fundraising activities, and planning to enhance our AI-vision infrastructure.

• Scale commercialization efforts (revenues)

• Scale fundraising activities (broaden investor base)

• Scale AI-vision infrastructure

Fundraising is deeply influenced by belief, and we are grateful for the strong support and confidence our investors have shown. As we step into the new year, we’re clearer-eyed, deeply grateful, and ready for the road ahead, with the future we’re building standing firmly on the foundation you helped solidify.

Happy new year,

The Epilog team

Follow us: © 2025 Epilog

🌱🧬  Scalable AI Infrastructure 

Throughout 2025, Epilog evaluated low-cost, highly efficient cloud computing architectures for scalable AI-vision processing. We assessed the cost, performance, and power-consumption tradeoffs between deploying heavy compute directly on edge devices versus shifting those workloads to cloud-based infrastructure tested in the context of Vortex AI compression.

In the second quarter, we secured a $100K strategic investment from Amazon, which supported early development of cloud-based AI video infrastructure. In parallel, we tested open source large language model (LLM) execution and  agentic-AI workflows on both AWS and local servers to better understand the underlying cost structure and scaling requirements.

By the fourth quarter of 2025, these efforts culminated in the development of a liquid-cooled compute module, which is currently in production. Built to work with Exogenous Optimization (EXO) software, our system is designed to scale both horizontally and vertically, potentially enabling a compute architecture that externalizes heavy AI workloads from edge devices into our own servers at significantly lower cost than relying exclusively on AWS resources.

Based on this work, we are now building a hybrid architecture in which AWS provides global connectivity and stream buffering backbone, while intensive AI computation is performed on Epilog-owned servers. This model preserves network and power redundancy while improving long-term control over compute economics.

As a secondary outcome of these efforts, we identified an opportunity to further improve efficiency by making its compute infrastructure more environmentally sustainable. Central to this initiative is a liquid-cooled, high-density AI compute module designed for continuous operation under demanding workloads, including video compression, analytics, and large-model inference. We are targeting estimated efficiency improvements of approximately 10–40% compared to conventional air-cooled systems, which continue to account for an estimated 85–90% of data-center deployments.[1]

Epilog plans to demonstrate elements of this strategy in 2026, pending development progress with Vortex and Closeview and eventually across all our end products.

Michael

Follow us: © 2025 Epilog

During the fourth quarter of 2025, SideCar entered production at Epilog’s Silicon Valley production facility, and a new SideCar website was launched. The Company also released an updated SideCar mobile application on both the Apple App Store and Google Play, introducing expanded functionality, including new Dashcam features.

SideCar made its public debut at the SEMA Show in November 2025, the world’s largest automotive aftermarket trade show. As of year-end 2025, SideCar supports 79 vehicle models, including platforms from Honda, Toyota, Subaru, Hyundai, and Lexus. Epilog is developing a next-generation SideCar OEM concept, which remains in development and is intended for potential future commercialization, subject to successful completion and market conditions.

In parallel, Epilog advanced execution across its forensic AI platform under the Aperis branding. During the third quarter of 2025, the Company launched a new website for its Aperis subsidiary (aperis.ai), and initial shipments of Vortex were delivered. Vortex, the Company’s AI-based video processing software, debuted at a New York City trade show in November 2025.

During the fourth quarter, Epilog initiated development of a major Vortex software update intended to support customer deployments in environments with constrained uplink bandwidth, while continuing to advance its Agentic AI capabilities. Development of a new CloseView hardware platform integrating Vortex software also progressed significantly during the fourth quarter. The current CloseView design is intended to support up to 100-megapixel resolution and incorporates the Company’s patented architecture. Both Vortex and CloseView are planned to debut at ISC West in March 2026.

In the fourth quarter of 2025, Epilog also advanced development of a next-generation Wellcome digital health scanner. The Company elected to postpone the Wellcome debut at CES and instead plans to introduce Wellcome at a health-focused trade show in mid-2026. These initiatives remain subject to various technical challenges, commercial viability assessments, and regulatory compliance requirements.

Michael

Follow us: © 2025 Epilog

2025 marked a year of significant momentum for Epilog across capital formation, operations, and investor engagement. 

During the year, the Company surpassed $3 million raised, bringing total capital raised to more than $8 million from nearly 7,000 investors. Mid-campaign, Epilog raised over $1 million in approximately four weeks and ranked as one of the top StartEngine Regulation Crowdfunding (Reg CF) campaign in the fourth quarter of 2025.

In the fourth quarter of 2025, Epilog further strengthened its operating foundation with the appointment of a Chief Financial Officer, the establishment of a Silicon Valley production facility, and the opening of international offices in Eindhoven, the Netherlands.

To enhance transparency and investor engagement, the Company launched its Investor Update Hub in the first quarter of 2025 (blog.epilog.com) and initiated quarterly investor update webinars.

In parallel, Epilog expanded its market presence and community engagement. During 2025, the Company exhibited at industry trade shows, conducted product demonstrations in Las Vegas and New York City, and expanded its international footprint.

The Company also launched the Spectrum Vision newsletter and rolled out new websites for both Epilog and Aperis. These initiatives are intended to support brand awareness, community engagement, and ongoing investor communications as Epilog continues executing its long-term strategy.

Michael

Follow us: © 2025 Epilog

AI vision systems are advancing toward higher resolution, longer data retention, and increasingly sophisticated analytics. As these trends accelerate, the cost and energy required to store and process visual data are expected to become increasingly significant.

Epilog is developing integrated hardware and software solutions intended to improve the efficiency of backend vision processing. Whether data originates from sensors in vehicles, buildings, or drones, our architecture is designed to support secure, scalable data operations across on-premises environments and the cloud.

At the core of this approach are hybrid processing technologies spanning the edge and compute infrastructure. This includes a new liquid-cooled compute module and newly designed operating software focused on improving vision processing efficiency. Our integrated solutions are intended to support storage and post-processing simultaneously across all our products, including SideCar, Vortex, CloseView, and Wellcome.

We believe these efficiencies may translate into lower energy consumption and a more environmentally efficient approach to AI vision processing.

AWS connectivity plays a key role in this strategy. By integrating with AWS services, our systems are designed to balance edge processing, private compute, and cloud resources, with the objective of minimizing unnecessary data transfer while maintaining elastic scalability.

We believe vision infrastructure should be scalable, intelligent, and environmentally efficient by design. This principle underpins Epilog’s development efforts and reflects our view that real-world AI vision may represent an important long-term technology shift.

Michael

Follow us: © 2025 Epilog

Video is becoming the largest, messiest data problem in the real world. A Billion cameras¹ , higher resolution, longer retention. That means more storage, more bandwidth, and more compute.

This is where Vortex changes the equation. 

With Vortex, we aim to compress 2-10 days of recordings to about 1 day in storage terms, potentially reducing energy consumption in a similar ratio, based on our current testing. What’s more this process also reduces energy consumption in the same ratio.

And because we pair compression with AI vision, it becomes smart footage, not dumb footage. You can search across it for what matters. If you need to find “the person in a red coat,” you don’t have to scrub through endless timelines. You query it.

Our technology aims to provide less storage use, reduced waste, potentially faster investigations, and improved outcomes, based on our current development and testing. That’s the kind of infrastructure we think the world needs more of, and it’s a core piece of what we’re building at Epilog.

Michael

[1] https://www.videoexpertsgroup.com/glossary/how-many-surveillance-cameras-in-the-world

Follow us: © 2025 Epilog

Reaching new heights, 2025 has been a milestone year for Epilog, marked by significant progress across the board. This progress reflects strong execution across technology development, product validation, and capital formation:

• Product validation achieved across two major platforms: SideCar and Vortex

• $3.2M+ raised this round, $8.3M+ raised to date

• 6,900+ investors supporting Epilog’s long-term vision

• Pursuing plans to begin generating revenue

We believe these milestones position Epilog for its next phase of growth as we continue building a unified Vision AI OS across multiple markets. We have established a fixed closing date for this offering, and don't anticipate extending.

Our Reg CF round closes January 29, 2026.

If you’ve been following along, considering joining, or thinking about increasing your investment, this is the stretch where it begins to matter—not because we’re trying to create hype, but because this is the point where momentum turns into execution.

Michael

Follow us: © 2025 Epilog

Holiday stockings are small moments of surprise that complement larger celebrations and accomplishments. In that tradition this year, we have prepared a batch of ten new technical innovations ready for patent application filings. 

Epilog’s Patent Process

We don’t use AI to dream up patent ideas; our emphasis is on quality, not quantity. Our innovations reflect the pace of organic technical discovery directly emerging from our labs. They are a byproduct of a rigorous process of building, testing, and validating new technologies we believe are suitable for intellectual property protection. Each innovation represents what we believe to be a tangible advancement in hardware, systems, or deployable AI architectures. We curate patent filings in portfolio-style batches to maximize efficiency and consistency.

Our ten inventions this year span imaging technology, radar, autonomous vehicles, AI compression, agentic systems, supercomputing, and AI vision platforms. We are committed to exploring new architectures, integrations, sensor designs, and system-level approaches through hands-on engineering, prototyping, and iterative development each year.

Some Noteworthy Patents

Many of the best modern technology platforms originated from practical, hands-on laboratory work focused on solving real-world problems—much the same way Epilog’s patents are developed. Some examples include:

• NVIDIA’s CUDA and GPU compute architecture emerged from internal graphics laboratories and evolved into general-purpose AI compute platforms.

• Google's TPU was developed in-house to accelerate machine-learning inference and is now a core cloud offering.

• ARM’s low-power CPU architectures grew from iterative laboratory optimization and are licensed across billions of chips.

• Boston Dynamics’ balance and locomotion technologies trace back to prolonged lab testing and real-world experimentation.

• Sony’s stacked CMOS image sensors were developed for the iPhone.

• Apple’s LiDAR depth mapping and ProRes/HEVC technologies were developed for the iPhone.

•  Microsoft’s Kinect depth sensing emerged from product internal R&D efforts.

• Netflix’s adaptive video streaming, along with Google’s GFS and Colossus file systems, began as internal solutions to operational challenges before becoming industry-defining technologies.

We expect that some of Epilog’s advancements this year may be incorporated into future products. The timing, implementation, and outcomes of these developments remain subject to ongoing research, testing, regulatory considerations, and other factors.

Warmest wishes,

Michael

Follow us: © 2025 Epilog

Why AI Compute Demand Is Exploding

By the time you read this sentence, OpenAI has probably processed another million chats worldwide.¹ Despite rapid adoption, only an estimated 7% of U.S. adults currently use ChatGPT regularly.² The proposed Stargate AI data-center project, if realized, could require up to 7 GW—comparable to the Dallas–Fort Worth metropolitan region’s peak load—to meet the increasing demand for AI.³⁴

At the same time, hundreds of competing open-source Large Language Models (LLMs) have entered the market, including LLaMA, Mistral, DeepSeek, and Falcon. Major semiconductor companies such as AMD, Intel, and Qualcomm continue developing inference-focused chips, while Google, Microsoft, and Amazon Web Services expand AI-optimized compute infrastructure. But what is driving demand for AI computation power?

The Next Frontier in AI

In computing, imaging has historically been one of the most resource-intensive modalities. A typical page of text is only a few kilobytes, while the average smartphone photo is hundreds of kilobytes—roughly 100× larger.⁵ Vision-based AI systems, which analyze real-world imagery rather than text, require substantially more computation. These demands are one reason large-scale AI data-center capacity is expected to expand significantly.

Since 2017, Epilog has centered its work on Vision AI—deploying intelligence at the sensor level, at the edge, and now in the cloud. We are developing compute capability to support expected future workloads. We believe Vision AI will serve as an important application layer for emerging technologies such as intelligent software systems, assisted-driving applications, next-generation robotics, health diagnostics, and perception-driven AI agents.

Spectrum Vision AI

It is our view that the foundational role of vision in both human experience and AI systems remains largely under-recognized. To help bridge that gap, we created Spectrum, a newsletter focused on AI Vision and the multispectral sensing capabilities already present in SideCar, Wellcome, and CloseView. Our goal is to build a community that extends beyond our existing investor base—one that stays informed on key developments, emerging applications, and early adoption of AI Vision technology as the world transitions toward perception-based AI and as Epilog advances its mission.

How to subscribe

If you haven’t subscribed yet, you can join the Spectrum newsletter and receive weekly insights on AI Vision by visiting:

Enter your email and click Subscribe. The first issue is scheduled to publish on Dec 15th, 2025.

Sources:

1. OpenAI system metrics referenced publicly by OpenAI during 2023–2024 developer events.
   

2. Pew Research Center: “AI and ChatGPT Adoption in the U.S.” (2024).
   

3. Reuters: “OpenAI, Oracle, SoftBank plan $500 billion Stargate data-center project” (2025).
   

4. U.S. Energy Information Administration (EIA): Texas ERCOT Region Peak Load Data showing Dallas–Fort Worth peak load in the ~7 GW range.
   

5. Apple iOS Technical Guidelines: typical iPhone JPEG/HEIC image sizes ranging from 200–800 KB.

Follow us: © 2025 Epilog

Dear Investors and Supporters,

We are pleased to announce that Epilog has surpassed $3 million+ raised in our current Regulation CF campaign, bringing our total funding to more than $8.2 million+ to date. 

This milestone reflects continued support of our mission. Nearly 1,490+ investors have participated in this round so far, expanding our community to 6,700+ investors who are supporting the advancement of Epilog’s AI-powered vision devices to market.

We appreciate the confidence so many of you have placed in our team. Your support enables us to continue executing on our plans and pursuing our long-term goals. We look forward to sharing additional updates as they become available.

With gratitude,

Michael

Follow us: © 2025 Epilog

In case you missed it, as part of Epilog’s planned quarterly updates, this webinar covered our Q3 2025 progress, including product milestones, sales and marketing initiatives, upcoming trade shows, and our revenue objectives. We also addressed investor questions throughout the session.

We highlighted ongoing commercial pilot programs, expansion of our reseller and distribution partnerships, and emerging market opportunities. In addition, we shared key updates on the development of our core products:

• SideCar -  Product Status

• Vortex/Closeview Live

• Wellcome - New Design Reveal for CESCheck Here

Best,

Michael 

Follow us: © 2025 Epilog