Aperture Labs

The research and IP behind it all.

Aperture Labs is where the methodology is developed, validated, and protected. A decade of work on capability-driven analysis (four issued US patents, an NSF research award, and collaborations with leading universities) underpins everything VertEx does.

Patents

Four issued US patents.

The core methods behind capability-driven analysis (automatically identifying, scoring, and surfacing unmet technical needs) are protected by issued US patents, granted to Artemis Intelligence LLC.

US 11,995,133
Systems & methods for automatically identifying unmet technical needs & problems
Analyzes technical documents to detect, score, and rank “problem kernels”. The most recent grant in the core detection family.
Granted
May 2024
US 11,762,916
User interface for identifying unmet technical needs & problems
The interface for searching, analyzing, and interpreting those needs, and for inferring a company’s technical capabilities from its documents.
Granted
Sep 2023
US 11,392,651
Systems & methods for automatically identifying unmet technical needs & problems
Methods for automatically surfacing and ranking technical problems across many areas of technology.
Granted
Jul 2022
US 11,379,538
Systems & methods for automatically identifying unmet technical needs & problems
Foundational methods for analyzing document text to identify and display technical problems.
Granted
Jul 2022

Three of the four (US 11,379,538 · 11,392,651 · 11,995,133) form a single patent family covering the detection method, with continuations broadening its scope over time; US 11,762,916 covers the interface and capability-inference layer.

National Science Foundation · APTO

Funding a new science of problems.

Backed by a multi-year award from the National Science Foundation's APTO program, Aperture Labs is building a new way to forecast where technology is headed: by reading the problems a field is wrestling with, not just the patents and papers it produces. It is a bet that problems are the earliest, most predictive signal of what comes next, and the science of reading them at scale is what the NSF is funding.

BibliometricsCitation networks, patent counts, publication volume
Outputs · what's been done
Performance curvesTechnology trajectories, S-curves, benchmark progress
Trajectories · extrapolated
Technology characteristicsEmbeddings: the “shape” of a technology
Shape · what it looks like
ProblemsWhat technologies are trying to solve, and where they struggle
Fundamental · leading · predictive

Most technology forecasting stops near the top. The deeper, harder-to-reach problem layer is where the leading signal lives, and it requires reading full documents, not just titles and abstracts.

01 · Massive scale

Millions of documents, distilled.

We scan millions of patents, papers, and publications and pull out hundreds of thousands of discrete problems, technologies, solutions, and tradeoffs, all extracted and normalized into a single structured dataset.

02 · Modeled in depth

Every level of the stack, connected.

Each problem is placed at its level in the technology-to-market stack, from fundamental science and component integration to system design, manufacturing scale-up, adoption, safety, and regulation, and linked to its causes, its proposed solution technologies, and the companies and labs pursuing them.

03 · New ways to forecast

Metrics that read the future.

We are inventing metrics that read the shape of a field: whether a problem is actually getting solved, whether a technology is attacking the critical bottleneck or just optimizing broadly, and how concentrated or diverse a domain is. Together these signals point to which technologies will win, and where companies, labs, and governments should focus money and attention to accelerate progress.

Inside one domain · EV batteries
739K
Documents mined: patents, papers, and news
1.9M
Distinct problems extracted from the text
47K
Normalized problem concepts
26 yr
Of history reconstructed, 2000–2025

A snapshot of just one of three technical domains mapped so far. The same pipeline runs on any technology area in a matter of weeks.

National Science Foundation · APTO

Verifiable, quantitative history from messy data.

The same NSF program drove a second line of work: pulling quantitative, verifiable, historical data out of messy, unstructured sources. From decades of patents, journals, filings, and news, it extracts hard performance figures, reconciles wildly inconsistent units and conventions, traces every value back to the document it came from, and reconstructs how a technology's real capabilities have advanced over time.

  • Reads full documents, not just titles and abstracts, recovering figures buried deep in technical prose.
  • Reconciles inconsistent units, metrics, and reporting conventions into one comparable scale, with every value traceable to its source.
  • Separates goal vs. achieved values and rebuilds decades-long trajectories, the raw material for performance forecasting.
Energy density, reconstructed from the literature
Li-ion cell: achieved vs. credible ceiling (Wh/L), 2012–2025
2000 1500 1000 500 0 2012 2016 2020 2024 GAP TO TARGET Credible ceiling · ~2000 +87% ’18–’22

Let's talk.

Whether you're a technology-driven company exploring where your capabilities can win next, an investor trying to understand our trajectory, or a research partner interested in working with Aperture Labs, we'd like to hear from you.

info@aperture.ai · +1 440.462.4509 · Cleveland, OH · LinkedIn