CVX Research

What CVX Research does

CVX Research, Inc. is a privately held software company headquartered in Austin, Texas, founded in the mid-2000s (formally incorporated in 2012) to develop and distribute tools for disciplined convex programming. Its core product, CVX, is a MATLAB-based modeling framework that transforms MATLAB into a domain-specific modeling language for specifying and solving convex optimization problems under the Disciplined Convex Programming (DCP) ruleset, supporting problem classes including LP, QP, SOCP, SDP, and GP via backends such as SDPT3, SeDuMi, Gurobi, and MOSEK. The companion product, TFOCS (Templates for First-Order Conic Solvers), provides MATLAB templates for constructing customized first-order solvers for large-scale convex cone problems and is jointly owned with the California Institute of Technology under a BSD 3-Clause license. The software was co-designed by Michael Grant and Stanford Professor Stephen Boyd, and targets academic researchers, graduate students, and commercial engineers working in signal processing, machine learning, statistics, and engineering design optimization.

The company's revenue model has shifted materially over time. CVX originally required a paid CVX Professional license for commercial use to access bundled Gurobi and MOSEK solvers, while academic users received free access through partnerships with Gurobi Optimization and MOSEK ApS. As of April 21, 2024, CVX Research transitioned to a fully open-source model, eliminating the Professional license requirement and removing its primary historical revenue stream. Remaining monetization consists of paid professional support contracts for CVX and TFOCS users, with pricing not publicly disclosed. The company operates a community-led go-to-market motion via a StackExchange-style support forum (ask.cvxr.com), academic citations, GitHub distribution, and curriculum integration rather than direct sales, and has not disclosed any external venture or private equity funding.

Name

CVX Research

Legal name

CVX Research, Inc.

Website

https://cvxr.com

Company type

Private

Founded year

2005

Operating status

Operating

Headcount range

11–50 employees

Short description

CVX Research develops MATLAB-based convex optimization software — primarily the CVX modeling framework and the jointly-owned TFOCS solver templates — serving academic researchers, students, and commercial engineers in optimization, signal processing, and machine learning applications.

Ownership category

akta.pro rank

Headquarters

HQ city

Austin

HQ country

United States

HQ region

North America

Offices1 record

Markets served

GTM type

B2B

Offering type

Software

Cost components

Technology or R&D, Personnel, Infrastructure, Operations

Revenue model

1. CVX Software Licensing: CVX is freely available for academic and commercial use as of April 21, 2024. Previously, CVX Professional licenses were sold for commercial use with access to commercial solvers (Gurobi, MOSEK). The transition to fully open-source represents a shift away from licensing revenue.
2. Commercial Solver Integration Support: CVX historically sold CVX Professional licenses that provided access to bundled commercial solvers (Gurobi and MOSEK) and professional support. Academic users received full commercial capabilities at no charge through agreements with Gurobi and MOSEK.
3. Paid Professional Support Contracts: CVX Research offers paid support contracts for more in-depth technical assistance. Contact [email protected] for TFOCS support contracts. Professional support tickets available for CVX Professional license holders.

Pricing tiers

Go-to-market motion1 record

Distribution channels3 records

Marketing channels4 records

Core offering

CVX Research develops and distributes software for disciplined convex programming, principally CVX—a MATLAB-based convex modeling framework that turns MATLAB into a modeling language for specifying constraints and objectives via standard MATLAB syntax—and TFOCS, a set of MATLAB templates for constructing customized first-order solvers for large-scale convex models. Both products ship bundled with free solvers (SDPT3, SeDuMi) and integrate with commercial solvers (Gurobi, MOSEK), and are used by academic researchers and commercial engineers for signal processing, machine learning, statistics, and control systems applications.

Product overview

CVX Research provides software for disciplined convex programming. The main product is CVX, a MATLAB-based convex modeling framework that turns MATLAB into a modeling language for specifying and solving convex optimization problems. The companion product TFOCS (Templates for First-Order Conic Solvers) provides building blocks for constructing efficient, customized solvers for larger-scale convex models. Both products target engineers, researchers, and scientists working in signal processing, machine learning, statistics, and optimization. CVX is free for academic and commercial use with bundled free solvers (SDPT3, SeDuMi) and optional commercial solver integration (Gurobi, MOSEK). TFOCS is released under BSD license and is jointly owned by CVX Research and Caltech. Community support is provided through the CVX Forum, and comprehensive documentation is available online.

Differentiator

Problem solved

Functional benefit

Brands

• CVX: A Matlab-based convex modeling framework for disciplined convex programming
• TFOCS
• CVX Forum

Products and services

• CVX A MATLAB-based convex modeling framework that implements disciplined convex programming (DCP). CVX turns MATLAB into a modeling language, allowing constraints and objectives to be specified using standard MATLAB expression syntax. Free for academic and commercial use as of April 21, 2024, with bundled free solvers (SDPT3, SeDuMi) and optional commercial solver integration with Gurobi and MOSEK.
• TFOCS (Templates for First-Order Conic Solvers) A set of MATLAB templates and building blocks for constructing efficient, customized solvers for convex models. TFOCS scales to larger model sizes than CVX but requires models to be manually converted to standard conic forms. Released under BSD 3-Clause license and jointly owned by CVX Research and Caltech.
• Paid Professional Support Contracts Paid support contracts offered for more in-depth technical assistance on CVX and TFOCS. Trouble-ticket support is available to CVX Professional license holders, and TFOCS paid support contracts are arranged by contacting [email protected]. Pricing is not publicly disclosed.

Named customers2 records

Segments3 records

Ideal customer profiles2 records

Technology focussed Yes

API detail

Has API

No

API docs

API detail

Core technology

AI maturity

App detail

Integration7 records

AI capability6 records

Feature6 records

Partnerships

Four partnerships are on record, tiered major and core.

• Stanford UniversitymajorStrategic or Co-development Partner · 1 October 2013Stanford University, specifically Professor Stephen Boyd and colleagues, co-developed CVX and TFOCS. The Stanford convex optimization course (CVX101) has used CVX for instruction, with over 10,000 students enrolled as of 2014. Stanford Technical Report documents the academic foundations of the software.
• Gurobi OptimizationcoreTechnology or Integration · 1 January 2013CVX maintains a contractual agreement with Gurobi Optimization to ship Gurobi solver binaries with CVX distributions. Academic users can obtain Gurobi licenses at no charge through the Gurobi Academic Program. Commercial use of Gurobi within CVX requires either a bundled CVX+Gurobi license or a bring-your-own-solver license from CVX Research.
• MOSEK ApScoreTechnology or Integration · 1 January 2013CVX maintains a contractual agreement with MOSEK ApS to ship MOSEK solver binaries with CVX distributions. Academic users can obtain MOSEK licenses at no charge through MOSEK's Academic Licensing page. Commercial use of MOSEK within CVX requires either a bundled CVX+MOSEK license or a bring-your-own-solver license from CVX Research.
• California Institute of Technology (Caltech)coreStrategic or Co-development Partner · 1 October 2010Caltech is a joint copyright holder for TFOCS alongside CVX Research. TFOCS was conceived and written by Stephen Becker (Caltech), Emmanuel J. Candès (Stanford), and Michael Grant (CVX Research). Both Caltech and CVX Research jointly own the TFOCS intellectual property, and the software is freely available under a BSD 3-Clause license.

Scale indicators3 records

Recent moves6 records

Expansion highlights3 records

Others

• Stanford Optimization Lab (Boyd group): Academic research group led by Stephen Boyd (CVX's co-designer) that continues to publish algorithms, examples, and reference implementations relevant to CVX. The lab is a key ecosystem partner and a source of ongoing theoretical contributions that CVX incorporates.

Broad incumbents

• Gurobi Optimization: Commercial high-performance solver (LP, QP, MIP) that CVX integrates with and historically bundled. While not a direct modeling language competitor, Gurobi's own modeling interfaces and broader commercial reach make it an adjacent incumbent in the same optimization stack.
• MOSEK: Commercial solver (LP, QP, SOCP, SDP, GP, MIP) bundled with CVX and used by both academic and commercial CVX users. MOSEK ships its own modeling tools and represents an incumbent in the same convex optimization ecosystem.
• IBM CPLEX: Long-established commercial optimization solver used in enterprise and academic settings. While focused on solver capabilities rather than a high-level modeling language, CPLEX competes for the same end users and the same commercial optimization spend as CVX-bundled solvers.

Emerging players

• Pyomo: Python-based optimization modeling library supporting LP, NLP, MIP, and convex problems. Pyomo competes with CVXPY in the Python ecosystem and represents another open-source alternative path for users leaving MATLAB-based tooling.
• ECOS (Embedded Conic Solver): Open-source conic solver for second-order cone programs, commonly used as a backend for CVXPY and other Python-based modeling tools. ECOS is a partial alternative in the same open-source solver stack that competes with CVX's bundled free solvers.
• Convex.jl: Julia-based disciplined convex programming package analogous to CVX and CVXPY. It targets the same use cases (LP, SOCP, SDP) in the Julia ecosystem and represents an emerging-language alternative for convex modeling.

Regional players

• BARON (global optimization solver): Commercial solver for global (non-convex) optimization problems, with academic licensing widely used alongside CVX in research. While it solves a broader class of problems, BARON competes for the same research and applied-optimization budgets as CVX's commercial solver integrations.

Direct peers

• CVXPY: Open-source Python library for disciplined convex programming, providing the same modeling abstraction as CVX in a more widely used language. It is the most direct functional replacement for CVX and is the primary competitive threat to CVX Research's user base.
• YALMIP: MATLAB toolbox for modeling and solving optimization problems (linear, quadratic, semidefinite, etc.) with automatic solver selection. YALMIP is the closest in-language alternative to CVX and is widely adopted in the same academic and engineering user base.

Market position

Strengths5 records

Weaknesses5 records

Competitive moat5 records

Key risks6 records

Key highlights6 records

Customer concentration

• Xx.com/cvxr
• Facebookfacebook.com/CVX-Research
• LinkedInlinkedin.com/company/cvx-research-inc.

Revenue estimate

Valuation estimate

Number of profiles

Profiles2 records

Funding overview

Funding rounds

Investors

Funding detail is available on the Subscription and Enterprise plan.Contact sales →

M&A

Investments

M&A and investment is available on the Subscription and Enterprise plan.Contact sales →

Frequently asked questions about CVX Research