Blog
 » 

Custom App Development

 » 
330+ App Development Cost Statistics 2026: Budgets, Rates, Timelines & ROI (Updated August)

330+ App Development Cost Statistics 2026: Budgets, Rates, Timelines & ROI (Updated August)

330+ app development cost statistics for 2026, covering budgets, developer and agency rates by region, build timelines, no-code platform pricing and ROI.

Jesus Vargas

By 

Jesus Vargas

Updated on

Aug 25, 2026

.

Reviewed by 

Why Trust Our Content

330+ App Development Cost Statistics 2026 | Budgets & Rates

Ask five people what it costs to build an app and you will get five ranges, all of them wide, none of them sourced. Search the question and the first page is agency pricing pages quoting "$50,000 to $250,000" with no sample size, no methodology and no year attached.

That is a strange gap, because the underlying data exists. Standards bodies publish function-point costs from repositories of thousands of completed projects. National audit offices publish line-item breakdowns of government software builds. Labor statistics agencies publish developer wages by country. Platform vendors publish their own price lists. The numbers are public. They are just scattered across PDFs, audit reports and pricing pages that nobody has put in one place.

Here at LOW / CODE, we build software for a living, roughly a hundred projects a year across low-code, high-code and AI. We assembled this page so founders, journalists, analysts and researchers can find a defensible number instead of a marketing range.

Every statistic below was opened at its primary source and confirmed to appear there. Claims that traced only to another agency's blog, a content-farm listicle, an undated page or a dead link were rejected. One hundred and thirteen of them were, including the most-cited low-code statistic in the industry.

Key takeaways

  • The most-cited statistic in low-code has no traceable source. The claim that Gartner predicted 70% of new applications would use low-code by 2025 appears on hundreds of vendor pages and agency blogs. It does not appear on any live Gartner press release, glossary page or public research abstract. Neither does "citizen developers will outnumber professional developers 4 to 1." Both were rejected here.
  • Bigger projects cost less per unit of functionality, not more. Across 898 completed projects sized with ISO-standard function points, cost per function point falls consistently as project size rises through all four ISBSG size bands. Fixed project overhead spreads across more delivered output.
  • The build is roughly a fifth of what the software costs you. About $83 billion of planned US federal IT spending for fiscal 2025, or 79% of the total, went to operations and maintenance rather than development or modernization.
  • Cost reduction is no longer the main reason companies outsource. It fell from 70% of respondents naming it the primary driver in 2020 to 34% in 2024. Access to talent is now first at 42%.
  • The same developer costs 2.6 times more depending on the country. Average hourly labour cost in the EU information and communication sector runs from €71.7 in Ireland and €58.5 in Germany down to €29.3 in Poland and €24.0 in Bulgaria.
  • AI coding assistants produce a measurable but smaller gain than the marketing suggests. Three randomized controlled trials covering 4,867 developers at Microsoft, Accenture and a Fortune 100 company found a 26.1% increase in completed tasks. Less experienced developers gained more than senior ones.
  • Most pre-seed startups are building on less than a quarter of a million dollars. 35% of pre-seed rounds on Carta were smaller than $250,000 in Q4 2025, a record high, while only 8% exceeded $5 million.
  • Government software timelines are measured in decades. ICT projects on the UK's Government Major Projects Portfolio average 10.7 years to deliver, and that makes ICT the fastest major project category the UK tracks.

Global software development spending and market size

Software is now the second-largest line in worldwide IT spending and one of the fastest-growing. This section covers what the world spends on software and IT services, how those forecasts have moved during 2026, how spending breaks down by region, how large the US custom software and systems-design industry actually is, and how many developers exist to build any of it. Every figure below comes from a forecast, government dataset, or original survey published by the organization that produced it.

Worldwide IT and software spending

  • Worldwide IT spending is forecast to reach $6.37 trillion in 2026, up 14.2% from $5.58 trillion in 2025. (Gartner, July, 2026)
  • Software spending is forecast at $1.468 trillion in 2026, up 15.5% from $1.271 trillion in 2025. (Gartner, July, 2026)
  • Software accounts for 23% of all worldwide IT spending in 2026, the second-largest of Gartner's six segments after IT services. (Gartner, July, 2026)
  • Services spending, which includes application implementation and managed services, is forecast at $1.570 trillion in 2026, up 5.3%. (Gartner, July, 2026)
  • Infrastructure as a service is forecast at $287 billion in 2026, up 29.3%, the second-fastest growth rate of any segment. (Gartner, July, 2026)
  • Software, services and IaaS combined reach $3.33 trillion in 2026, or 52% of all worldwide IT spending. (Gartner, July, 2026)
  • Data center systems is the fastest-growing segment at 62.5% growth to $822 billion in 2026, ahead of IaaS, software and devices. (Gartner, July, 2026)
  • Gartner raised its 2026 IT spending forecast three times in six months, from $6.15 trillion in February to $6.31 trillion in April to $6.37 trillion in July. (Gartner, February, 2026) (Gartner, April, 2026) (Gartner, July, 2026)
  • Gartner's April 2026 forecast put IT services alone at $1.870 trillion in 2026, up 9.0%, using a taxonomy that combined application and infrastructure services with IaaS. (Gartner, April, 2026)
  • Gartner builds these forecasts from analysis of sales by more than a thousand vendors across hardware, software, IT services and telecommunications. (Gartner, July, 2026)

Spending by region

  • ICT spending across Asia/Pacific excluding Japan and China is forecast to reach $647 billion in 2026, up 5.4%, and to surpass $758 billion by 2029. (IDC, March, 2026)
  • Software is the fastest-growing technology group in that region at 16.8% year over year, while capturing 24% of regional ICT spend. (IDC, March, 2026)
  • Services absorb more than 23% of the Asia/Pacific ICT budget, driven by outsourcing and the expansion of global capability centers in India and Southeast Asia. (IDC, March, 2026)
  • IDC projects a 2026 to 2029 compound annual growth rate above 7.4% for Asia/Pacific excluding Japan and China. (IDC, March, 2026)
  • In EMEA, 61% of organizations plan to increase IT budgets in 2026, compared with 52% in North America. (Spiceworks Ziff Davis and Aberdeen, December, 2025)

The custom software and systems design market

The global developer population

  • SlashData estimates 48.4 million software developers worldwide as of Q3 2025, based on 29 survey waves of more than 10,000 developers each, calibrated against GitHub, Stack Overflow and employment statistics. (SlashData, Q3, 2025)
  • More than 180 million developers now use GitHub, with over 36 million joining in the 12 months to August 2025, the platform's fastest absolute growth on record. (GitHub, October, 2025)
  • The United States has 28 million GitHub developers and India has 21.9 million, up from 4.5 million in India in 2020. (GitHub, October, 2025)
  • India added more than 5.2 million developers in 2025, over 14% of all new GitHub accounts, and GitHub projects India will reach 57.5 million developers by 2030. (GitHub, October, 2025)
  • Net new GitHub developers by region in 2025: APAC +13 million, Europe +6.3 million, Africa and the Middle East +3.4 million, LATAM +3.2 million. (GitHub, October, 2025)
  • GitHub hosted 630 million repositories at the end of 2025, including 121 million created that year, and recorded 986 million commits, up 25%. (GitHub, October, 2025)
  • The US employed 1,693,800 software developers in 2024, projected to reach 1,961,400 by 2034, a 16% increase. (U.S. Bureau of Labor Statistics, August, 2025)
  • Software developers, QA analysts and testers together number 1,895,500 and are projected to grow 15% through 2034, five times the 3% average for all US occupations. (U.S. Bureau of Labor Statistics, August, 2025)
  • About 129,200 US openings for software developers, QA analysts and testers are projected each year through 2034. (U.S. Bureau of Labor Statistics, August, 2025)
  • The US median annual wage for software developers was $133,080 in May 2024, against $105,990 for computer occupations overall and $49,500 for all occupations. (U.S. Bureau of Labor Statistics, August, 2025)
  • Median software developer pay was $149,990 at software publishers and $129,890 in computer systems design and related services. (U.S. Bureau of Labor Statistics, August, 2025)
  • The global cloud native developer population reached 19.9 million in Q1 2026, roughly 39% of all developers worldwide, up from 15.6 million in Q3 2025. (CNCF and SlashData, March, 2026)

App development cost by type and complexity

Most published "average app cost" ranges come from agency marketing pages with no stated sample and no methodology. This section only uses numbers that trace to a standards body's project repository, a national audit office, a commissioned study with a published financial model, or a vendor's own price list. Where no defensible primary data exists for a sub-topic, we say so rather than filling the gap.

How complexity and size change cost

The International Software Benchmarking Standards Group is a not-for-profit founded in 1997 by national software metrics associations. It maintains a repository of completed projects sized using ISO/IEC standardised function point methods, and publishes short analyses of that data.

  • ISBSG analysed 898 completed projects with both project cost and functional size recorded, drawn from projects delivered after 2015 and rated A or B for data quality. (ISBSG, 2023)
  • Roughly 90% of those 898 projects cost less than 1 million euros, and the median project cost was approximately half that amount. (ISBSG, 2023)
  • Cost per function point falls as project size rises, across all four size bands ISBSG measured: small (30 to 100 function points), medium (100 to 300), medium (300 to 1,000) and large (1,000 to 3,000). Larger builds are more cost efficient per unit of delivered functionality, because fixed project activities are spread across more output. (ISBSG, 2023)
  • ISBSG identifies programming language as a second major driver of cost per function point, alongside project size. (ISBSG, 2023)

Native iOS versus Android

  • ISBSG examined 79 agile mobile projects delivered after 2015: 40 with a primary programming language of Android and 39 of iOS. All were enhancements adding functionality to an existing app. (ISBSG, August, 2025)
  • Android projects showed a lower Project Delivery Rate than iOS projects, meaning fewer hours spent per function point delivered. ISBSG suggests stricter App Store acceptance rules as one possible cause. (ISBSG, August, 2025)
  • Android projects also showed faster delivery speed, measured in function points delivered per calendar month, than the iOS projects in the same dataset. (ISBSG, August, 2025)
  • Shipping the same product on iOS, Android and web multiplies release work rather than dividing it. Canada's ArriveCAN produced 177 releases between April 2020 and October 2022: 66 Android, 70 iOS and 41 web. (Office of the Auditor General of Canada, February, 2024)

A government-audited build, end to end

Canada's Auditor General performed a full performance audit of a single mobile application. It remains one of the few app builds with a line-item cost breakdown published by an independent audit body. Figures are Canadian dollars.

Enterprise internal tools

The study below was commissioned by Microsoft and conducted by Forrester Consulting, which published its interview base, composite organisation assumptions and full financial model. It is vendor-commissioned and should be read as such.

  • Forrester modelled a 50% reduction in professional developer app development time for internal workflow applications built on Power Apps, based on interviews with five representatives at four organisations. (Forrester Consulting for Microsoft, July, 2024)
  • The model uses an average fully burdened annual salary of $175,500 for a professional developer building internal applications. (Forrester Consulting for Microsoft, July, 2024)
  • The composite organisation of 30,000 employees incurs $4.32 million per year in Power Apps licensing alone, supporting 200 professional developers and 1,800 citizen developers, against $46.1 million in three-year risk-adjusted benefits and $15.1 million in three-year costs. (Forrester Consulting for Microsoft, July, 2024)

Platform floor costs by app type

These are published list prices, which set the recurring floor beneath a build. They are not build costs.

  • Bubble charges $59 per month on Starter and $549 per month on Team, both billed annually, and covers web and native mobile apps in a single project plan. A paid plan is required to publish to the App Store or Google Play. (Bubble, 2026)
  • FlutterFlow charges $39 per month on Basic, $80 for the first seat on Growth and $150 for the first seat on Business, with additional seats at $55 and $85. (FlutterFlow, 2026)
  • Webflow charges $15 per month for a basic site and $25 for a CMS site, both billed yearly, with ecommerce site plans at $29, $74 and $212 per month and a Platform tier at $2,500 per month on an annual contract. (Webflow, 2026)
  • Webflow prices web app usage separately at $2 per million requests and $2 per 5 hours of CPU, on top of the site plan. (Webflow, 2026)
  • Sharetribe prices a multi-sided marketplace at $39 per month to build, $99 to launch a live marketplace, $199 for a professional marketplace on a custom domain and $299 to run custom code against the live environment, with 50, 250 and 500 free transactions per month respectively and $0.19 or less per additional initiated transaction. (Sharetribe, 2026)

Developer salaries and agency rates by region

Labor is the single largest line item in almost every software budget, and it moves more by geography than by anything else. This section sets out what developers are paid in the United States, Europe, Latin America and Asia, what a developer costs an employer once benefits are added, and what software firms actually collect per delivery professional per year. Every figure below comes from a government statistics agency, a large-scale developer survey, or a company's own reported results.

What software developers are paid in the United States

  • US software developers had a mean annual wage of $148,100 and a median hourly wage of $65.38 in May 2025, across 1,687,890 jobs. (U.S. Bureau of Labor Statistics, May, 2026)
  • The BLS Occupational Outlook Handbook puts the median annual wage for software developers at $133,080 and for software quality assurance analysts and testers at $102,610, both as of May 2024. (U.S. Bureau of Labor Statistics, August, 2025)
  • US computer occupations averaged $120,100 a year in May 2025 against $69,770 for all occupations, a premium of 72%. (U.S. Bureau of Labor Statistics, May, 2026)
  • Specialization changes US pay materially: computer and information research scientists averaged $153,930, software developers $148,100, data scientists $126,800, software QA analysts and testers $111,490, computer programmers $105,170 and web developers $98,770 in May 2025. (U.S. Bureau of Labor Statistics, May, 2026)
  • US employment of software developers, QA analysts and testers is projected to grow 15% from 2024 to 2034, with about 129,200 openings a year. (U.S. Bureau of Labor Statistics, August, 2025)
  • Levels.fyi reported a 2025 US median total compensation of $226,000 for a mid-level software engineer, $312,000 at senior, $457,000 at staff and $155,000 at entry level, from more than 245,000 data points across 5,000-plus companies. These are self-reported and include stock, so they skew toward equity-paying technology employers rather than the whole US labor market. (Levels.fyi, December, 2025)
  • Levels.fyi measured US software engineer pay growth of 2.67% year over year in 2025, with the half-year median moving from $190,000 in H1 to $195,000 in H2. (Levels.fyi, December, 2025)

Developer pay in Europe

  • UK full-time programmers and software development professionals had a median gross annual pay of £56,914 in April 2025, up 4.2% year over year across roughly 343,000 full-time jobs. (Office for National Statistics, October, 2025)
  • UK IT quality and testing professionals had a median full-time gross annual pay of £47,118 in April 2025, 17% below the programming and software development median. (Office for National Statistics, October, 2025)
  • Germany's federal employment agency puts the median gross full-time pay for a software developer at €6,301 per month, with a lower quartile of €5,034 and an upper quartile of €7,723. The series is capped at the social-insurance contribution ceiling, so the top of the distribution is truncated. (Bundesagentur für Arbeit, Entgeltatlas, 2025)
  • Mean annual gross earnings for professionals in the information and communication sector ranged from €78,189 in Germany and €90,905 in Ireland down to €29,721 in Poland, €30,147 in Hungary and €30,903 in Bulgaria. Germany pays 2.6 times Poland for the same occupational band. Reference year is 2022, the most recent EU Structure of Earnings Survey. (Eurostat, Structure of Earnings Survey 2022)
  • Within the same 2022 survey, Switzerland was the highest at €126,024 and the EU-27 average was €57,522 for professionals in information and communication. (Eurostat, Structure of Earnings Survey 2022)
  • Average hourly labour cost in the EU information and communication sector was €48.2 in 2025 against €34.9 for the whole economy, with Ireland at €71.7, France at €59.2, Germany at €58.5, Spain at €34.8 and Poland at €29.3. (Eurostat, April, 2026)
  • Ireland's information and communication sector had average hourly earnings of €48.92 in Q1 2026 and an average hourly total labour cost of €62.65, the highest of the country's 13 sectors and well above the €38.92 all-sector average. (Central Statistics Office Ireland, May, 2026)
  • Poland's information and communication sector paid an average gross 14,689.70 zł a month in 2025, the highest of any sector and 65% above the national average of 8,903.56 zł. (Główny Urząd Statystyczny, June, 2026)
  • Stack Overflow's 2025 survey put median back-end developer pay at $108,913 in the United Kingdom, $87,011 in Germany and $71,929 in France, against $175,000 in the United States. (Stack Overflow, July, 2025)

Developer pay in Latin America and Asia

  • Stack Overflow's 2025 survey put the median Indian back-end developer at $22,086 and the median Indian full-stack developer at $13,949, against $175,000 and $138,000 respectively in the United States. That is a 7.9x gap at the back-end median. (Stack Overflow, July, 2025)
  • Global median compensation across all Stack Overflow respondents was $89,427 for AI and ML engineers, $79,742 for back-end developers, $72,509 for full-stack, $69,609 for mobile, $62,015 for front-end and $57,442.50 for QA and test, from 23,928 respondents who answered the compensation question. (Stack Overflow, July, 2025)
  • India's technology sector was expected to cross $315 billion in revenue in FY2026 with direct employment of about 6 million people, a net addition of 135,000 in the year. That implies roughly $52,500 of sector revenue per employee. (nasscom, February, 2026)
  • Economy-wide average annual wages in 2025 were $86,977 in the United States, $76,285 in Germany, $61,259 in Korea, $50,183 in Japan, $49,074 in Poland and $24,463 in Mexico, against an OECD average of $64,809, in constant US dollars at 2025 PPPs. These are all-occupation baselines rather than developer salaries, and India is not in the dataset. (OECD, 2026)

Total employer cost and what agencies actually collect per developer

  • Employer costs for private-industry workers averaged $46.60 per hour worked in March 2026, of which wages were $32.60 and benefits $14.01. Benefits were 30.1% of total compensation, or an additional 43% on top of cash wages. (U.S. Bureau of Labor Statistics, June, 2026)
  • For professional and related occupations in the US information industry, total employer compensation cost was $96.92 per hour worked in March 2026: $62.89 of wages and $34.03 of benefits. Benefits were 35.1% of the total, adding 54% on top of cash wages. At 2,080 hours that is about $201,600 a year per employee. (U.S. Bureau of Labor Statistics, June, 2026)
  • In the professional, scientific and technical services industry, which includes software consultancies, employer compensation cost averaged $75.15 per hour worked in March 2026, split $51.37 wages and $23.78 benefits, or about $156,300 a year at 2,080 hours. (U.S. Bureau of Labor Statistics, June, 2026)
  • EPAM Systems reported $5.457 billion of 2025 revenue against approximately 56,600 delivery professionals, which works out to about $96,400 of client revenue per delivery professional for the year. (EPAM Systems, February, 2026)
  • Globant reported $2,454.9 million of 2025 revenue and ended the year with 26,906 technology, design and innovation professionals, about $91,200 of revenue per professional. Latin America supplied 21.1% of Q4 revenue, with Argentina the largest country in that region. (Globant, February, 2026)
  • Globant's IFRS gross profit margin was 35.0% for full-year 2025, meaning direct delivery cost consumed roughly two-thirds of every dollar billed. (Globant, February, 2026)
  • Endava reported FY2025 revenue of £772.3 million with headcount of 11,479 at 30 June 2025 and an average of 10,255 operational employees in Q4, about £67,300 of revenue per head. (Endava, September, 2025)
  • Tata Consultancy Services reported FY2026 revenue of $30,017 million against 584,519 employees, about $51,400 of revenue per employee, with cost of revenue of $17,990 million, or about $30,800 per employee. (Tata Consultancy Services, April, 2026)
  • Infosys reported $20,158 million of FY2026 revenue (Infosys, April, 2026) against a total headcount of 328,594, of which 310,887 were software professionals, about $61,300 of revenue per employee. (Infosys, 2026)
  • The spread between the highest and lowest of these five listed firms is 1.9x on revenue per delivery head, from about $96,400 at EPAM to about $51,400 at TCS, which is far narrower than the 7.9x salary gap between US and Indian back-end developers. Delivery-side wage arbitrage does not pass through to the client price one for one. (EPAM Systems, February, 2026) (Tata Consultancy Services, April, 2026)

In-house teams versus outsourcing versus agencies

The build-versus-buy comparison is usually argued with hourly rates on both sides, which understates the in-house number and overstates the difference. This section uses national statistical agencies for what an employee actually costs an employer, a central bank survey for the size of the offshore delivery market, a large executive survey for why organizations outsource, and a national audit office for the one place a contractor premium has been measured and published. Where a widely quoted figure could not be opened at its stated source, it is listed in the rejected section rather than repeated.

The size of the market and where the work goes

  • Worldwide spending on IT services is forecast at $1.570 trillion in 2026, up 5.3% from $1.492 trillion in 2025. Gartner's category covers application implementation and managed services plus infrastructure implementation and managed services. (Gartner, July, 2026)
  • Services is the largest single line in worldwide IT spending in 2026 at $1,570 billion, ahead of communications services at $1,354 billion and software at $1,468 billion, and it is also the slowest-growing: 5.3%, against 15.5% for software, 29.3% for infrastructure as a service and 62.5% for data center systems. Total IT spending is forecast at $6.369 trillion, up 14.2%. (Gartner, July, 2026)
  • India exported $204.7 billion of software services in 2024-25, a 7.3% increase on the prior year, rising to $218.6 billion once services delivered by foreign affiliates of Indian companies are included. The Reserve Bank of India contacted 6,766 software export companies and received 2,206 responses covering around 90% of total software services exports. (Reserve Bank of India, November, 2025)
  • 90.7% of India's software services exports in 2024-25 were delivered off-site rather than on-site at the client. (Reserve Bank of India, November, 2025)
  • The United States took 52.9% of India's software exports in 2024-25, down from 54.1%, while Europe rose from 30.8% to 32.8%. Exports to the US grew 4.9% against 14.3% for Europe. (Reserve Bank of India, November, 2025)
  • IT services accounted for $131.3 billion of India's $204.7 billion in software services exports, with BPO services at $55.8 billion, engineering services at $10.8 billion and software product development at $6.8 billion. (Reserve Bank of India, November, 2025)

Why organizations outsource, and why they bring work back

Deloitte's Global Outsourcing Survey is the longest-running survey that asks executives to rank their own stated reasons for outsourcing. The 2024 edition surveyed more than 500 global business and technology leaders, including more than 150 C-suite executives across industries.

  • Access to talent is the number-one driver of outsourcing decisions, named by 42% of respondents, ahead of increasing customer demands at 35%, spend optimization at 34%, improved quality and performance at 33% and adoption of a global delivery model at 33%. (Deloitte, 2024)
  • The share of respondents naming cost reduction as a primary outsourcing driver fell from 70% in 2020 to 48% in 2022 to 34% in 2024. Among organizations using outcome-based delivery models it is lower still, at 30%. (Deloitte, 2024)
  • 72% of organizations outsource application and software development, against 77% for IT infrastructure services, 77% for cybersecurity, 75% for innovative technologies such as generative AI and 69% for application support. (Deloitte, 2024)
  • 96% of surveyed organizations use third-party outsourcing as a source of talent and 78% use global in-house centers. (Deloitte, 2024)
  • Investment in third-party outsourcing is expected to increase for 40% of respondents, with 20% reporting a planned reduction. (Deloitte, 2024)
  • 70% of executives report having insourced scope that was previously with a third party at some point in the past five years, but 65% of those did so very selectively, moving back less than 25% of scope. Only 13% insourced more than 50% of scope. (Deloitte, 2024)
  • Insourcing is not driven by dissatisfaction: 82% of the respondents exploring more insourcing said outsourced services meet or exceed expectations, and 80% are maintaining or increasing their outsourcing investment. The leading insourcing drivers are better control over service quality and performance (68%), growing strategic capabilities in-house (64%) and spend optimization (56%). (Deloitte, 2024)

What an in-house engineer actually costs

Salary is roughly two-thirds of the employer's cost. The US Bureau of Labor Statistics measures the rest directly, which makes the fully-loaded multiplier a measured number rather than a rule of thumb.

  • Employer costs for employee compensation averaged $49.32 per hour worked for civilian workers in March 2026: $33.72 in wages and salaries and $15.60 in benefits. (U.S. Bureau of Labor Statistics, June, 2026)
  • For private industry workers, wages and salaries accounted for 69.9% of the $46.60 hourly cost and benefits for the remaining 30.1%. (U.S. Bureau of Labor Statistics, June, 2026)
  • For private industry management, professional and related occupations, total compensation was $78.10 per hour worked against $53.50 in wages and salaries, with benefits at $24.61, or 31.5% of the total. That is a fully-loaded multiplier of about 1.46 times base pay before any recruiting, equipment, software, office or management overhead. (U.S. Bureau of Labor Statistics, June, 2026)
  • The benefits load for that group breaks down as paid leave $7.34 per hour, insurance $5.73, legally required benefits $5.03, supplemental pay $3.60 and retirement and savings $2.89. (U.S. Bureau of Labor Statistics, June, 2026)
  • 1,687,890 people were employed as software developers in the United States in May 2025, at a mean annual wage of $148,100, a mean hourly wage of $71.20 and a median hourly wage of $65.38. Software quality assurance analysts and testers numbered 186,740 at a mean annual wage of $111,490, and web developers 70,190 at $98,770. (U.S. Bureau of Labor Statistics, May, 2026)
  • Median pay for software developers, QA analysts and testers was $131,450 per year, or $63.20 per hour, in 2024, against 1,895,500 jobs. Employment is projected to grow 15% from 2024 to 2034, adding 287,900 jobs. (U.S. Bureau of Labor Statistics, 2025)
  • Hourly labour costs in the information and communication sector averaged €48.2 across the EU in 2025, against €24.0 in Bulgaria, €26.0 in Romania, €29.3 in Poland and €34.9 in Spain, and €58.5 in Germany, €59.2 in France, €63.5 in Denmark and €71.7 in Ireland. This is total labour cost, meaning compensation of employees plus taxes minus subsidies. (Eurostat, April, 2026)
  • Across the whole EU economy, average hourly labour costs were €34.9 in 2025 and ranged from €12.0 in Bulgaria to €56.8 in Luxembourg. Non-wage costs were 24.8% of total labour costs in the EU and 25.6% in the euro area. (Eurostat, March, 2026)

Hiring, turnover and the measured contractor premium

  • Median cost-per-hire in 2025 was $1,200 for nonexecutive roles and $10,625 for executive roles, a ratio of 8.9 to 1, up from 7.0 in 2022 and 3.1 in 2017. Executive cost-per-hire is up 113% since 2017; nonexecutive cost-per-hire is down 27%. (SHRM, 2025)
  • Median time-to-fill in 2025 was 44 calendar days for nonexecutive positions and 45 days for executive positions, measured from requisition opening to offer acceptance. In organizations with 5,000 or more employees it was 61 and 60 days. (SHRM, 2025)
  • Nearly 7 in 10 organizations (69%) report difficulty recruiting for full-time regular positions. SHRM reports this figure in its 2025 recruiting benchmarking brief, sourced from its own 2025 Talent Trends survey. (SHRM, 2025)
  • In the US information sector in June 2026 there were 90,000 hires, 31,000 quits and 56,000 layoffs and discharges, monthly rates of 3.2%, 1.1% and 2.0% respectively. Layoffs exceeded quits, which is unusual outside a contracting labour market. (U.S. Bureau of Labor Statistics, August, 2026)
  • In professional and business services in June 2026 there were 1,042,000 hires, 488,000 quits and 504,000 layoffs and discharges, monthly rates of 4.6%, 2.2% and 2.2%. A 2.2% monthly quits rate is roughly 26% of the workforce a year on voluntary departures alone. (U.S. Bureau of Labor Statistics, August, 2026)
  • Across the 36 months of ArriveCAN development from April 2020 to March 2023, external resource cost exceeded internal resource cost in all but two months: April 2020, the first month, at $49,248 internal against $0 external, and April 2021, at $158,368 internal against $5,658 external. The peak month was March 2022, at $2,269,627 external against $143,470 internal. (Office of the Auditor General of Canada, February, 2024)
  • Beyond the four largest suppliers, the audit attributed $3.2 million to TEKsystems, $3.0 million to Donna Cona, $2.9 million to BDO Canada, $2.4 million to MGIS, $1.1 million each to 49 Solutions and to Makwa Resourcing / TPG Technology Consulting, and $1.0 million to Advanced Chippewa Technologies. (Office of the Auditor General of Canada, February, 2024)
  • Under the federal task authorization model used for ArriveCAN, the contract sets per diem rates and each task authorization sets a maximum rate, but the contractor chooses which of its own or its subcontractors' resources fill the roles, and payment goes to the contractor rather than to the subcontractors. The audit found the agency had little documentation supporting how and why the initial contract was awarded non-competitively, and that the same supplier was later involved in developing the requirements for the competitive request for proposal. (Office of the Auditor General of Canada, February, 2024)

Project timelines, budget overruns, and failure rates

Software budgets are not normally distributed. The average overrun on an IT project is modest, the median project lands close to its estimate, and a small share of projects overrun by multiples that would not appear at all under a bell curve. The statistics below come from peer-reviewed research, government audit bodies, and the original McKinsey-Oxford dataset, and they are grouped so the average and the tail are never confused with each other.

How long software projects actually take

  • Government ICT projects on the UK's Government Major Projects Portfolio take 10.7 years on average to deliver. (NISTA, July, 2026)
  • The 22 ICT projects on that portfolio carry a combined whole-life cost of £42.5 billion against £51.3 billion of monetised benefits, making ICT the smallest GMPP category by both number and value. (NISTA, July, 2026)
  • ICT is the fastest of the UK's major project categories: infrastructure and construction averages 14.8 years and military capability averages 20.3 years. (NISTA, July, 2026)
  • Every additional year a project is scheduled to run increases its cost overrun by 15%. (McKinsey & Company, October, 2012)
  • Across 5,392 IT projects completed between 2002 and 2014, the median estimated cost was USD 0.7 million and the mean was USD 6.6 million, which shows how heavily the population skews toward small projects with a few very large ones. (Journal of Management Information Systems, August, 2022)

Budget overruns: the average and the tail

Why IT budget risk behaves differently from every other kind of project

Failure rates, scope creep, and wasted budget

  • 62% of projects were completed within their original budget and 55% on time; 73% met their original goals or business intent. (PMI, March, 2021)
  • 34% of projects experienced scope creep and 12% were deemed outright failures. (PMI, March, 2021)
  • 35% of project budget was lost to failed projects, and wasted investment from poor project performance ran at 9.4% of spend, down from 11.4% the year before. (PMI, March, 2021)
  • Shifting requirements and technical complexity account for 9 of the 45 percentage points of average cost overrun on large IT projects; unclear objectives and lack of business focus account for 13 points, execution issues 11, skill gaps 6, and 6 points are unexplained. (McKinsey & Company, October, 2012)
  • 34 of 189 UK government major projects (18%) were rated red at the end of March 2026, meaning successful delivery appears unachievable; 29 (15%) were green and 109 (58%) amber. (NISTA, July, 2026)
  • Of the 42 projects that left that portfolio during the year, 26 reported delivery against their objectives and 7 were brought to early closure. (NISTA, July, 2026)
  • The US federal government invests more than $100 billion a year in IT, roughly 80% of it on operating and maintaining existing systems, many of them legacy. (GAO, January, 2025)
  • GAO has made more than 1,800 recommendations on federal IT management since 2010; 463 were still unimplemented as of January 2025. Federal IT acquisition has been on GAO's High-Risk List since 2015. (GAO, January, 2025)
  • A peer-reviewed survey of 193 experienced IT project managers found 61.66% of projects rated in the top two levels of global success, and criticised the widely quoted low-success figures for non-disclosure of methodology and for judging projects only on scope, time, and cost compliance. (ACM Queue, September, 2024)

What it costs to fix a defect late

The no-code and low-code market, and what the platforms actually cost

Two questions get asked about no-code and low-code, and the published answers to both are usually wrong. The first is how big the market is, where most citations trace back to a resold summary rather than the analyst firm that produced the number. The second is what the tools cost, where most citations are stale by a year or more.

The market figures below come from Gartner, IDC, JetBrains and the vendors' own published statements. The pricing figures were captured by opening each vendor's public pricing page in August 2026 and recording what it displayed. All prices are US dollars as published.

Market size and forecast

  • Gartner projects the worldwide low-code development technologies market will reach $58.2 billion by 2029, at a compound annual growth rate of 14.1%. (Gartner, November, 2025)
  • Gartner forecast the worldwide low-code development technologies market at $26.9 billion in 2023, an increase of 19.6% over 2022. (Gartner, December, 2022)
  • Gartner's segment table put total low-code development technologies revenue at $18.497 billion in 2021, $22.462 billion in 2022, $26.869 billion in 2023 and $31.949 billion in 2024. (Gartner, December, 2022)
  • Low-code application platforms are the largest single component of that market, forecast to grow 25% to nearly $10 billion in 2023. Gartner states that its LCAP market definition includes no-code platforms, because it defines a no-code application platform as an LCAP that only requires text entry for formulae or simple expressions. (Gartner, December, 2022)
  • Citizen automation and development platforms were the fastest-growing segment in that forecast, at 30.2% growth for 2023, from a base of $732 million in 2022. (Gartner, December, 2022)
  • Gartner's earlier forecast put the same market at $13.824 billion in 2021, up 22.6% from $11.272 billion in 2020 and $9.153 billion in 2019. (Gartner, February, 2021)
  • IDC expects the low-code, no-code and intelligent developer technologies market to grow from $32.3 billion in 2024 to $225.4 billion in 2029, a 47.4% compound annual growth rate. IDC's market definition is broader than Gartner's and folds in generative AI developer technologies, which accounts for most of the gap between the two forecasts. (IDC, December, 2025)

Who is actually building with it

  • Gartner predicts that by 2026, developers outside formal IT departments will account for at least 80% of the user base for low-code development tools, up from 60% in 2021. (Gartner, December, 2022)
  • Gartner research found that on average 41% of employees outside of IT, whom Gartner calls business technologists, customize or build data or technology solutions. (Gartner, February, 2021)
  • Gartner predicted that half of all new low-code clients would come from business buyers outside the IT organization by year-end 2025. (Gartner, February, 2021)
  • Gartner surveyed 2,820 business technologists in March 2021 and found that 77% routinely use a combination of automation, integration, application development or data science and AI tools in their daily work, and that 76% assume enterprise risk ownership for that work. (Gartner, September, 2021)
  • In JetBrains' State of Developer Ecosystem 2025, based on 24,534 developers across 194 countries surveyed between April and June 2025, 17% of respondents globally reported using low-code or no-code tools for building websites and applications. (JetBrains, January, 2026)
  • JetBrains reports that business process automation, not app building, remains the top low-code and no-code use case worldwide. (JetBrains, January, 2026)
  • IDC's October 2025 survey spotlight, drawn from a worldwide survey of 1,000 full-time professional developers at organizations with 500 to 10,000 or more employees, concludes that low-code and no-code platforms have moved from tools that primarily supported knowledge workers to platforms widely adopted by professional developers. (IDC, October, 2025)

Platform scale, as the vendors state it

These are self-reported figures published by the vendors on their own sites. They are not independently audited.

  • Microsoft states that Power Platform has 56 million monthly active users, growing 27% year over year, and cites its FY2025 third quarter earnings call as the source. (Microsoft, May, 2025)
  • Bubble reports 7.2 million apps launched by builders worldwide and 28.6 billion workflows run across all Bubble apps in 2025, with apps built on Bubble transacting over $1 billion in the year. (Bubble, December, 2025)
  • Bubble reports that builders created more than 180,000 native mobile apps in the roughly six months after it launched native mobile development in June 2025. (Bubble, December, 2025)
  • FlutterFlow states on its pricing page that over 3.3 million users build with the platform. (FlutterFlow, August, 2026)
  • Softr states on its pricing page that it is trusted by over 1 million businesses; Glide states over 100,000 companies; Airtable states over 500,000 companies. (Softr, August, 2026) (Glide, August, 2026) (Airtable, August, 2026)

Live platform pricing, captured August 2026

Each figure below was read directly off the vendor's public pricing page on the date of capture. Where a tier's price did not render, it is omitted rather than estimated.

  • Bubble publishes a free plan at $0 per month, Starter at $59 per month billed annually and Team at $549 per month billed annually, with a single project plan covering web, iOS and Android. The page states that annual billing saves up to 20%, and that a paid plan is required to publish to the Apple App Store or Google Play. (Bubble, August, 2026)
  • Bubble's Starter plan includes 175,000 workload units per month, 5 mobile build submissions per month and 3 live app versions; Team includes 5 app editors, 20 build submissions per month and 8 live app versions. Additional file storage is $3 per 100 GB per month. (Bubble, August, 2026)
  • FlutterFlow publishes a free plan at $0, Basic at $39 per month, Growth at $80 for the first seat and $55 for the second, and Business at $150 for the first seat and $85 for seats 2 to 5. The page carries a monthly and annual billing toggle stating that annual billing saves roughly 25%. (FlutterFlow, August, 2026)
  • Glide publishes GlideOS Free at $0 per month for 1 project, 1 app, 100 MB of storage and 2 team members, and Solo at $25 per month for 5 projects, unlimited apps, 25 GB of storage and 100 starting AI credits. Glide states that GlideOS is a separate product from the original Glide platform and requires a separate subscription. (Glide, August, 2026)
  • Webflow publishes site plans at $15 per month for Basic and $25 per month for CMS, both billed yearly, and a Platform tier at $2,500 per month requiring an annual contract. (Webflow, August, 2026)
  • Webflow's ecommerce site plans are published at $29, $74 and $212 per month, all billed yearly, and Workspace seats at $39 per month for a full seat and $15 per month for a content seat, both billed yearly, with reviewer seats free. (Webflow, August, 2026)
  • Webflow Cloud web app usage is billed separately from the site plan at $2 per million requests and $2 per 5 hours of CPU usage, with bandwidth add-ons starting at $20 per month for 50 GB billed annually and running to $974 per month for 2.45 TB. (Webflow, August, 2026)
  • WeWeb publishes Free at $0 per month, a front-end only plan at $13 per month, Launch+ at $20 per month, Grow+ at $66 per month and Scale+ at $208 per month. Scale+ includes 200 GB of bandwidth then $0.25 per GB, 100 GB-hours of Lambda compute then $0.072 per GB-hour, and 150 GB of file storage then $0.03 per GB. (WeWeb, August, 2026)
  • Xano publishes Free at $0, Essential at $85 per month billed annually and Pro at $224 per month billed annually. The free plan is rate limited to 10 requests per 20 seconds and capped at 100,000 database records. (Xano, August, 2026)
  • Xano prices its featured add-ons at $180 per month for a CPU and autoscale boost, $5 to $10 per month for an additional 5 GB to 10 GB of database and media storage, and $500 per month for enhanced security including HIPAA and a BAA. (Xano, August, 2026)
  • Airtable publishes Team at $20 per seat per month billed annually or $24 billed monthly, and Business at $45 per seat per month billed annually or $54 billed monthly. The free plan is capped at 1,000 records per base, 5 editors and 100 automation runs per month. (Airtable, August, 2026)
  • Airtable's Business plan includes 125,000 records per base, 100,000 automation runs, 100 GB of attachments per base and 20,000 AI credits per paid user per month, with extra AI credits starting at $20 per month for 10,000 credits and the Portals add-on starting at $150 per month for 15 guests. (Airtable, August, 2026)
  • Retool publishes builder seats at $10 per month on Team and $50 per month on Business, with non-building internal users at $5 and $15 per month respectively. The free plan is capped at 5 users and 500 workflow runs per month. (Retool, August, 2026)
  • Retool's external user pricing on Business is free for the first 50 users, then $8 per month for 51 to 250, $6 for 251 to 500 and $4 above 500. Additional workflow runs are $75 per 5,000 runs per month. (Retool, August, 2026)
  • Softr publishes Basic at $19 per month, Pro at $99 per month and Business at $329 per month, with a free plan at $0. Additional builder seats are $5, $10 and $20 per month on those three tiers. The page offers a yearly billing option described as two months off. (Softr, August, 2026)
  • Softr's Business plan includes 3 builders, 30 team app users plus 100 client app users, 1 million database records, 25,000 workflow actions per month and 100 AI credits per month. Additional custom domains are $15 per month, or $13 per month billed annually. (Softr, August, 2026)
  • Adalo publishes Free at $0, Starter at $36 per month billed annually and Team at $160 per month billed annually, with no usage-based or token-based charges on any tier. Starter covers 1 published app, 1 app editor and 5 GB of team storage; Team covers 5 published apps, 10 app editors and 125 GB. The page carries a toggle stating annual billing saves 20%. (Adalo, August, 2026)
  • Zapier publishes a free plan at $0 per month including 100 tasks per month and two-step Zap workflows, and Professional starting from $19.99 per month. Zapier prices in task tiers running from 100 to 2 million tasks per month, and states that one Zapier MCP tool call consumes two tasks. (Zapier, August, 2026)
  • Make publishes a free plan at $0 per month with up to 1,000 credits per month and a 15-minute minimum interval between scenario runs, and Core at $12 per month for 10,000 credits per month. Make states that as of August 27 credits are its billing unit, and that usage allowance scales at 5 GB of data transfer and 10 MB of data storage per 10,000 monthly credits. (Make, August, 2026)

AI-assisted development: productivity evidence and tool pricing

AI coding tools are now close to universal among professional developers, and the published evidence on what they do to output is genuinely mixed. Controlled trials at Microsoft, Accenture and Google measured double-digit gains, while a randomized trial with experienced open-source maintainers measured a slowdown that the participants themselves failed to notice.

This section collects the controlled experiments, the large-scale telemetry, the code-quality counterevidence and the current published price of every major AI development tool.

Controlled experiments on developer productivity

  • Developers using GitHub Copilot completed an HTTP server task in 1 hour 11 minutes on average versus 2 hours 41 minutes without it, a 55% reduction, across 95 professional developers. The result was significant at P=.0017 with a 95% confidence interval of 21% to 89%. (GitHub, September, 2022)
  • Three randomized controlled trials covering 4,867 developers at Microsoft, Accenture and an anonymous Fortune 100 company found a 26.08% increase in completed tasks among developers given an AI coding assistant, with a standard error of 10.3%. (Management Science, February, 2026)
  • In the same study, less experienced developers showed both higher adoption rates and larger productivity gains than senior developers. (Management Science, February, 2026)
  • A randomized controlled trial with 96 full-time Google software engineers on a complex enterprise-grade task estimated AI shortened time on task by about 21%, with the authors noting the confidence interval is large. (Google, October, 2024)
  • A randomized trial with 16 experienced open-source developers across 246 real issues found that allowing AI tools increased completion time by 19%, with a confidence interval between +2% and +39%. (METR, July, 2025)
  • The same developers forecast a 24% speedup before the study and still believed they had been sped up by 20% afterward, a 39-point gap between perceived and measured effect. (METR, July, 2025)
  • Economics experts predicted a 39% reduction in completion time and machine-learning experts predicted 38% for that same trial. Both predictions had the wrong sign. (METR, July, 2025)
  • The developers in that trial averaged five years of experience on repositories with more than 22,000 GitHub stars and over one million lines of code, and were paid $150 per hour. They used Cursor Pro with Claude 3.5 and 3.7 Sonnet. (METR, July, 2025)
  • METR's follow-up experiment, run from August 2025 with 57 developers across 143 repositories and more than 800 tasks, reported "a speedup of -18% with a confidence interval between -38% and +9%" for returning participants and -4% (interval -15% to +9%) for newly recruited developers. Both intervals cross zero. (METR, February, 2026)
  • METR states its follow-up data "is only very weak evidence" because developers unwilling to work without AI increasingly declined to participate, and 30% to 50% of developers avoided submitting tasks where AI had high expected value. (METR, February, 2026)
  • In a 202-developer controlled study, code written with GitHub Copilot had a 53.2% greater likelihood of passing all 10 unit tests, and averaged 18.2 lines of code per error versus 16.0 without, a 13.6% improvement. (GitHub, November, 2024)
  • Blind reviewers rated Copilot-assisted code 3.62% more readable, 2.94% more reliable, 2.47% more maintainable and 4.16% more concise, and were 5% more likely to approve it. (GitHub, November, 2024)
  • In a peer-reviewed user study of 47 participants, those given an AI assistant wrote significantly less secure code than the control group and were more likely to believe their code was secure. On the encryption task, 67% of the AI group produced a correct solution versus 79% of the control group. (ACM CCS, November, 2023)

Large-scale telemetry and survey evidence

  • 90% of nearly 5,000 surveyed technology professionals use AI at work, and respondents report a median of two hours of their most recent workday interacting with AI, roughly one quarter of an eight-hour day. (DORA, September, 2025)
  • More than 80% of DORA respondents report AI increased their productivity, but more than 40% describe the increase as only "slight." Fewer than 10% perceive any decrease. (DORA, September, 2025)
  • Trust in AI-generated output is thin: 4% report "a great deal" of trust and 20% "a lot," while 46% say "somewhat," 23% "a little" and 7% "not at all." (DORA, September, 2025)
  • DORA's 2025 finding is that AI adoption now improves software delivery throughput but continues to increase delivery instability, which the report attributes to testing, review and deployment systems that have not scaled with code generation. (DORA, September, 2025)
  • In DORA's 2024 survey, a 25% increase in AI adoption was associated with an estimated 1.5% reduction in delivery throughput and a 7.2% reduction in delivery stability. (DORA, October, 2024)
  • The same 2024 analysis found a 25% increase in AI adoption associated with a 7.5% increase in documentation quality, a 3.4% increase in code quality, a 3.1% increase in code review speed and a 1.8% decrease in code complexity. Individual-level gains and system-level losses appeared in the same dataset. (DORA, October, 2024)
  • 39.2% of 2024 DORA respondents reported little (27.3%) or no (11.9%) trust in the quality of AI-generated code, while 75.9% relied on AI for at least part of their job. (DORA, October, 2024)
  • 84% of Stack Overflow's 49,000-plus respondents use or plan to use AI tools, up from 76% the prior year, and 51% of professional developers use them daily. (Stack Overflow, July, 2025)
  • 46% of developers actively distrust the accuracy of AI output versus 33% who trust it, and only 3% "highly trust" it. (Stack Overflow, July, 2025)
  • The top AI frustration, cited by 66% of developers, is "AI solutions that are almost right, but not quite," and 45.2% say debugging AI-generated code is more time-consuming than expected. (Stack Overflow, July, 2025)
  • 75.3% of developers say they would ask another person for help when they do not trust an AI answer. (Stack Overflow, July, 2025)
  • GitHub Copilot reached 50 million users, and one in three pull requests on GitHub now involves an agent. (Microsoft, July, 2026)
  • Google's agentic development platform Antigravity had more than 2.4 million weekly active users as of the second quarter of 2026. (Alphabet, July, 2026)
  • Adidas reported that teams working in loosely coupled architectures with fast feedback loops saw productivity gains of 20% to 30% from generative AI, while teams tightly coupled to ERP systems saw little or no benefit. (DORA, September, 2025)

Code quality, security and review burden

  • Analysis of 623 million code changes from 2023 to 2026 found block duplication rose from 40.3 to 73.0 per million lines, an 81% increase and the highest level on record. (GitClear, January, 2026)
  • Over the same dataset, moved code, the signal for refactoring, fell from 21% of changed lines in 2022 to 3.8% year to date in 2026, while within-commit copy/paste rose from 9.4% to 15.7%. (GitClear, January, 2026)
  • Cross-file function calls per thousand lines fell 35% between 2023 and 2026, error-masking constructs rose 47% and two-week code churn rose 15%. (GitClear, January, 2026)
  • Across 150-plus large language models tested on 80 coding tasks, AI-generated code passed security review only 55% of the time while syntax correctness exceeded 95%. (Veracode, March, 2026)
  • Security pass rates split sharply by language: Python 62%, C# 58%, JavaScript 57% and Java 29%. (Veracode, March, 2026)
  • Pass rates also split by vulnerability class: 86% for insecure cryptography and 82% for SQL injection, but 15% for cross-site scripting and 13% for log injection. (Veracode, March, 2026)
  • Reasoning-focused models achieved the highest observed security pass rates at 70% to 72%, still short of production reliability. (Veracode, March, 2026)

Current published pricing for AI development tools

  • GitHub Copilot Free includes 2,000 completions per month; Pro costs $10 per month with $15 in monthly AI credits, Pro+ costs $39 with $70 in credits and Max costs $100 with $200 in credits. (GitHub Copilot pricing page, August, 2026)
  • GitHub Copilot Business costs $19 per granted seat per month and Copilot Enterprise costs $39 per granted seat per month, each with a pooled monthly AI credit allowance. (GitHub Copilot plans documentation, August, 2026)
  • Cursor Hobby is free; Pro costs $20 per month, Pro Plus $60 and Ultra $200. (Cursor pricing documentation, August, 2026)
  • Cursor Teams Standard costs $40 per user per month and Teams Premium $120 per user per month, with Enterprise priced on request. (Cursor pricing page, August, 2026)
  • Cursor's included third-party model usage is denominated in dollars: $20 monthly on Pro, $70 on Pro Plus and $400 on Ultra, after which usage continues at API rates on pay-as-you-go billing. (Cursor pricing documentation, August, 2026)
  • Claude Code is included in Claude Pro at $17 per month on annual billing or $20 billed monthly, and in Claude Max starting at $100 per month for 5x or 20x Pro usage. (Claude pricing page, August, 2026)
  • Claude Team standard seats cost $25 per member monthly or $20 on annual billing, and premium seats cost $125 monthly or $100 annually, with a two-member minimum. Both seat types include Claude Code. (Claude Team plan page, August, 2026)
  • Devin has four self-serve plans: Free, Pro at $20 per month, Max at $200 per month and Teams at an $80 per month minimum. (Devin billing documentation, August, 2026)
  • Devin Teams charges $40 per month per full seat and $0 for flex seats, which draw entirely from a shared pool of prepaid on-demand credits; purchased credits roll over and never expire. (Devin billing documentation, August, 2026)
  • Windsurf now ships as Devin Desktop, and windsurf.com/pricing redirects to devin.ai/pricing. Devin Desktop access is bundled with Devin Pro, Max and Teams full seats. (Devin Desktop plans documentation, August, 2026)
  • In March 2026 Windsurf replaced prompt credits with daily and weekly token-based quotas, with usage past the quota billed at model API list prices. Existing Pro subscribers were grandfathered at $15 per month and Teams seats at $30 per month. (Windsurf quota documentation, August, 2026)
  • Replit Starter is free; Core costs $25 per month or $20 per month billed annually and includes $25 of monthly credits, and Pro costs $100 per month or $95 billed annually and includes $100 of monthly credits. (Replit pricing page, August, 2026)
  • Lovable Free costs $0; Pro costs $25 per month with 100 monthly credits and Business costs $50 per month with 100 monthly credits, with Enterprise on volume-based pricing. (Lovable pricing page, August, 2026)
  • Lovable charges credits per action rather than per seat: Plan mode costs 1 credit per message, while Default mode varies with task complexity, from 0.50 credits to restyle a button to 0.90 to remove a footer. Free, Pro and Business workspaces also receive 5 daily build credits, capped at 30 per month on Free. (Lovable subscription documentation, August, 2026)
  • v0 Free includes $5 of monthly credits and a 7-message daily limit; Plus costs $30 per user per month and Business costs $100 per user per month, each including $30 of monthly credits plus $2 of free daily credits on login. (v0 pricing page, August, 2026)
  • Bolt Free includes 1 million tokens per month with a 300,000 daily cap; Pro costs $25 per month for 10 million tokens and Teams costs $30 per member per month for 10 million tokens. (Bolt pricing page, August, 2026)
  • Every tool above except GitHub Copilot Free and Bolt Free meters usage in credits, tokens or quotas rather than seats alone, which means the monthly subscription price is a floor and not a total. (Cursor pricing documentation, August, 2026) (Devin billing documentation, August, 2026)

The costs nobody budgets for: maintenance, technical debt, platform fees, and infrastructure

The quote you sign covers building the thing. It does not cover keeping it alive. Once software ships, the bill continues as operations and maintenance, as developer hours lost to bad code, as store commissions on every transaction, as cloud spend that grows faster than usage, and as compliance deadlines set by regulators and platform owners.

The numbers below come from government spending data, published survey research, and the platform policy pages that actually set the rates.

What it costs to keep software running

  • About $83 billion, or 79%, of planned federal IT spending for fiscal year 2025 was for operations and maintenance rather than development or modernization, across the 24 Chief Financial Officers Act agencies. (U.S. Government Accountability Office, July, 2025)
  • The 11 most critical federal legacy systems GAO identified are 23 to 60 years old and cost roughly $754 million a year to operate and maintain. (U.S. Government Accountability Office, July, 2025)
  • Developers report an average 41.1-hour work week, of which 17.3 hours are spent on maintenance: debugging, refactoring, and modifying existing code. (Stripe, September, 2018)
  • Developers spend an average 13.5 hours a week specifically on technical debt, which is 33% of the working week. (Stripe, September, 2018)
  • Time spent on "bad code" averages 3.8 hours per developer per week, which Stripe valued at nearly $85 billion in annual worldwide opportunity cost. (Stripe, September, 2018)
  • Stripe put total developer efficiency loss at 31.6%, equal to roughly $300 billion of global GDP lost each year, from an aggregate developer GDP base of $918 billion. (Stripe, September, 2018)

Technical debt as a line item

  • CIOs estimate technical debt at 20 to 40 percent of the value of their entire technology estate, before depreciation. (McKinsey & Company, October, 2020)
  • CIOs reported that 10 to 20 percent of the technology budget earmarked for new products is diverted to resolving technical debt. (McKinsey & Company, October, 2020)
  • 30% of CIOs surveyed believe more than 20% of the technical budget ostensibly dedicated to new products is diverted to technical debt. (McKinsey & Company, June, 2022)
  • 60% of CIOs surveyed said their organization's technical debt had risen perceptibly over the previous three years. (McKinsey & Company, October, 2020)
  • Almost half of firms that completed modernization programs were unsuccessful in reducing technology debt, based on analysis of 220 companies across five geographies and seven sectors. (McKinsey & Company, June, 2022)
  • Companies in the bottom 20th percentile for technical debt invest 50 percent less than average on modernization and are 40 percent more likely to have incomplete or canceled modernization programs. (McKinsey & Company, June, 2022)
  • One multinational insurer found technical debt amounted to 15 to 60 percent of every dollar it spent on IT, a cost that had never appeared in its modernization business cases. (McKinsey & Company, June, 2022)
  • One large North American bank found its 1,000-plus systems and applications together generated over $2 billion in technical-debt costs. (McKinsey & Company, June, 2022)
  • The cost of poor software quality in the United States reached at least $2.41 trillion in 2022, of which accumulated technical debt principal was roughly $1.52 trillion, a figure the report notes is close to the total spent on the entire US IT labor base that year. (CISQ, November, 2022)

App store commissions, developer fees, and forced update cycles

  • The Apple Developer Program costs 99 USD a year and the Apple Developer Enterprise Program costs 299 USD a year. (Apple, August, 2026)
  • A Google Play developer account costs a US$25 one-time registration fee. (Google, August, 2026)
  • Apple's App Store Small Business Program charges 15% commission for developers with up to 1 million USD in proceeds in the prior calendar year; crossing that threshold moves future sales to the standard rate. (Apple, August, 2026)
  • On standard App Store subscription terms, developers receive 70% of the subscription price in the first year and 85% after a subscriber accumulates one year of paid service, with free trials excluded from that count. (Apple, August, 2026)
  • From October 1, 2026 Apple replaces the EU per-install Core Technology Fee with a 5% Core Technology Commission on digital transactions in apps distributed outside the App Store. (Apple, August, 2026)
  • Under Apple's unified EU terms effective October 1, 2026, App Store commission is 26% on Apple In-App Purchase, 20% on alternative in-app payment processing, and 15% on out-of-app offers with actionable links, with reduced rates of 15%, 10% and 10% for Small Business, Mini Apps and Video Partner Program participants and for subscriptions past their first year. (Apple, August, 2026)
  • Google Play's revised service fees, live in the US, UK and EEA since June 30, 2026, are 10% on a developer's first $1 million of annual earnings and 10% on auto-renewing subscriptions regardless of revenue. (Google, August, 2026)
  • Above the first $1 million, Google Play charges 20% on non-recurring transactions from new installs and 25% from existing installs, plus a 5% billing fee when Google Play's billing system is used. (Google, August, 2026)
  • Google Play's new fee structure reaches Australia and Japan on September 30, 2026, South Korea on December 31, 2026, and the rest of the world on September 30, 2027. (Google, August, 2026)
  • Apple reviewed more than 9.1 million app submissions in 2025 and rejected over 2 million of them, comprising 1.2 million new apps and nearly 800,000 updates, a rejection rate of roughly 22%. (Apple, May, 2026)
  • Apple rejected over 443,000 submissions for privacy violations and over 371,000 for copying other apps, spam, or misleading users in 2025. (Apple, May, 2026)
  • Apple terminated 193,000 developer accounts over fraud concerns and rejected more than 138,000 developer enrollments in 2025. (Apple, May, 2026)
  • Apple reviews 90% of submissions in less than 24 hours on average, and over 40% of unresolved review issues relate to a single guideline, 2.1 App Completeness, which covers crashes, placeholder content and incomplete information. (Apple, August, 2026)
  • Since April 28, 2026, apps uploaded to App Store Connect must be built with Xcode 26 or later using an iOS 26, iPadOS 26, tvOS 26, visionOS 26 or watchOS 26 SDK, a recurring rebuild obligation independent of whether the app changes. (Apple, August, 2026)
  • From August 31, 2026, new Google Play apps and updates must target Android 16 (API level 36), and existing apps must target at least API level 35 to stay available to new users on newer devices; extensions run to November 1, 2026. (Google, August, 2026)

Infrastructure, cloud, security, and accessibility

  • Estimated wasted cloud spend stands at 29%, rising for the first time in five years, in a survey of 753 cloud decision-makers. (Flexera, March, 2026)
  • 76% of large enterprises now spend more than $5 million a month on public cloud. (Flexera, March, 2026)
  • Supabase lists Pro at $25 per month and Team from $599 per month, with dedicated compute add-ons from $10 a month for a Micro instance to $3,730 a month for 16XL, plus $100 a month for point-in-time recovery and $10 a month for a custom domain. (Supabase, August, 2026)
  • Bubble lists Starter at $59 per month and Team at $549 per month on annual billing, with additional file storage at $3 per 100 GB per month, and charges server usage as metered workload units on top. (Bubble, August, 2026)
  • The global average cost of a data breach reached $4.99 million in 2026, based on breaches at 602 organizations between March 2025 and February 2026. (IBM, July, 2026)
  • AI-enabled breaches cost about $6 million on average, roughly $1 million more than the global average, while organizations using AI and automation in security operations cut breach costs by almost $2 million. (IBM, July, 2026)
  • The Justice Department's ADA Title II web and mobile app rule adopts WCAG 2.1 Level AA and carries average annualized costs of $3.33 billion at a 3 percent discount rate and $3.52 billion at 7 percent over its first ten years. (U.S. Department of Justice, April, 2024)
  • Those accessibility compliance dates were pushed back one year in April 2026, to April 26, 2027 for entities serving 50,000 or more people and April 26, 2028 for smaller entities and special district governments. (U.S. Department of Justice, April, 2026)

Startup and MVP economics

Most first products are built on far less money than the "average cost to build an app" articles suggest. The numbers below come from Carta's cap-table data, CB Insights' shutdown post-mortems, the BLS Business Employment Dynamics survival series, real startup payroll records, and a survey of 2,000 startup builders.

They describe what founders actually raise, what they actually spend it on, how long it lasts, and how often it does not work.

What a first round actually contains

  • U.S. startups on Carta raised $10.4 billion across 50,316 SAFEs and convertible notes in 2025. (Carta, February, 2026)
  • That was a 1% decline in total pre-seed cash invested versus 2024 but a 13% decline in the number of instruments issued. (Carta, February, 2026)
  • 35% of pre-seed rounds on Carta were smaller than $250,000 in Q4 2025, a new high, while only 8% were larger than $5 million. (Carta, March, 2026)
  • Among SAFEs that raised at least $1 million, the average deal size reached $1.4 million in 2025, up from $1.1 million in 2024. (Carta, March, 2026)
  • Median post-money SAFE valuation caps in 2025 hovered around $10 million for rounds of $250,000 to $1 million and $15 million for rounds of $1 million to $2.5 million. (Carta, February, 2026)
  • In 2025, 97% of rounds on Carta below $500,000 were pre-seed rounds, and 93% of those were structured as SAFEs. (Carta, March, 2026)
  • Priced equity rounds become as common as SAFEs and notes once a financing reaches $3 million, and are the most likely structure above $4 million. (Carta, March, 2026)
  • Startups on Carta raised $119.5 billion across 4,859 new rounds in 2025, a 16.9% increase in capital on the lowest round count in at least six years. (Carta, February, 2026)
  • Average round size across all stages reached $30.2 million in Q4 2025, up from $19.3 million a year earlier. (Carta, February, 2026)
  • Median dilution on rounds from seed through Series C fell from about 18% to 16% during 2025, down from 19% two years earlier. (Carta, February, 2026)
  • At Series A, the median AI startup valuation was 38% higher than the median non-AI valuation in 2025; at Series E and later the premium reached 193%. (Carta, February, 2026)

Who is building, and with what

  • 22% of startups are founded by non-technical founders, with the technical-founder share falling from 82% to 78% across two survey waves of roughly 2,000 startup builders. (Supabase, 2026)
  • 61% of surveyed startups have a single founder, up from 53% a year earlier; 29% have two founders and 7% have three. (Supabase, 2026)
  • 61% of startups now have more than half their codebase written by AI, 40% are at 76% to 100%, and only 2% are at zero. (Supabase, 2026)
  • Among non-technical founders specifically, the share with 76% to 100% AI-generated code rises to 54%. (Supabase, 2026)
  • "Technical complexity" as the biggest business challenge fell from 24% to 11% in a single year, the largest movement in the survey; customer acquisition is now first at 32%, product-market fit second at 14%. (Supabase, 2026)
  • 56% of startups got their first paying customers from a personal or professional network, and 67% have never tried paid acquisition. (Supabase, 2026)
  • 31% of the heaviest AI-code users are monetizing, against 56% of startups using no AI-generated code. (Supabase, 2026)

What technical talent costs

  • The median annual wage for U.S. software developers was $133,080 in May 2024; the bottom 10% earned under $79,850 and the top 10% over $211,450. (U.S. Bureau of Labor Statistics, August, 2025)
  • At seed stage, the average CTO cash salary is $155,000, above both the average seed CEO at $153,000 and the average seed COO at $144,000, based on anonymized payroll records from venture-backed startups. (Kruze Consulting, August, 2026)
  • Average CTO cash pay rises to $196,000 at Series A and $238,000 at Series B, with an overall average of $167,000 and a median of $166,000. (Kruze Consulting, August, 2026)
  • Median startup CEO salary rose from $125,000 in 2018 to $147,000 in 2024, while the average moved from $130,000 to $141,000 after peaking at $150,000 in 2022. (Kruze Consulting, March, 2026)

Burn, runway, and how long the money lasts

  • The average venture-backed startup held close to 22 months of runway heading into 2024, while the median held just under 12 months. (Kruze Consulting, October, 2024)
  • In 2019, 42% of startups had less than six months of runway. (Kruze Consulting, October, 2024)
  • The standard guidance is to begin fundraising at roughly 10 months of remaining runway, and for seed and Series A companies to plan for a minimum of 18 months. (Kruze Consulting, October, 2024)
  • Carta's benchmark for seed-to-Series A graduation is 5% low, 10% median and 20% high after year one; 15%, 25% and 35% after year two; and 20%, 35% and 45% after year three. (Carta, March, 2026)

Failure rates from real datasets

  • 77.9% of U.S. private-sector establishments that opened in the year ending March 2024 were still operating one year later, meaning 22.1% closed inside their first year. (U.S. Bureau of Labor Statistics, 2026)
  • 51.4% of establishments that opened in the year ending March 2020 were still operating five years later, in March 2025. (U.S. Bureau of Labor Statistics, 2026)
  • 34.7% of establishments that opened in the year ending March 2015 were still operating ten years later, in March 2025. (U.S. Bureau of Labor Statistics, 2026)
  • 980,333 new private-sector establishments opened in the United States in the year ending March 2025. (U.S. Bureau of Labor Statistics, 2026)
  • Across 431 VC-backed startups that shut down since 2023, "ran out of capital" was cited in 70% of cases, poor product-market fit in 43%, bad timing or macro conditions in 29%, and unsustainable unit economics in 19%. (CB Insights, March, 2026)
  • Those 431 companies had raised a combined $17.5 billion before dying; the median company raised $11 million and the average raised $48 million. (CB Insights, March, 2026)
  • The median time from last fundraise to shutdown is 22 months, and nearly a quarter of the companies had gone more than three years since their last raise. (CB Insights, March, 2026)
  • Two-thirds of failed startups were shrinking headcount in the six months before shutdown, nearly a third died with 10 or fewer employees, and about 15% died with more than 100. (CB Insights, March, 2026)
  • Failed fintech startups had a median of just $4 million in equity funding, against a dataset-wide median of $11 million and a healthcare median of $47 million. (CB Insights, March, 2026)

Return on investment: what custom software and automation actually pay back

Software budgets are easy to measure. Returns are harder. This section collects the figures that are actually published with a stated methodology: independently modelled ROI and payback periods for internal applications and automation platforms, the measured cost of manual work and disconnected systems, error and downtime benchmarks, and the share of organizations that can show a return at all.

Several of the strongest ROI numbers come from Forrester Total Economic Impact studies, which are commissioned and paid for by the vendor being measured. Those are labelled as vendor-commissioned throughout. Their methodology is published, their composite organizations are described, and they are citable when attributed honestly.

Measured ROI and payback on internal applications and automation

  • A Forrester Total Economic Impact study of Microsoft Power Platform modelled a 224% three-year ROI, with $118.2 million in benefits against $36.5 million in costs and payback in under six months. The composite organization has 30,000 employees and $10 billion in revenue. Vendor-commissioned by Microsoft. (Forrester Consulting for Microsoft, July, 2024)
  • The same Power Platform study attributes $44.4 million of three-year benefit to employee time savings and $43.6 million to development cost avoidance, with roughly 25% efficiency gains per impacted worker. Vendor-commissioned. (Forrester Consulting for Microsoft, July, 2024)
  • A separate Forrester study of Microsoft Power Apps alone modelled 206% ROI, $46.1 million in benefits against $15.1 million in costs, $31.0 million NPV and payback in under six months. The composite organization runs 200 professional developers alongside 1,800 citizen developers. Vendor-commissioned. (Forrester Consulting for Microsoft, July, 2024)
  • A Forrester study of Microsoft Power Automate modelled 248% ROI, $55.93 million in benefits against $16.08 million in costs, $39.85 million NPV and payback in under six months. Vendor-commissioned. (Forrester Consulting for Microsoft, July, 2024)
  • A Forrester study of the SS&C Blue Prism intelligent automation platform modelled 330% ROI and $53.40 million NPV, including $12.7 million of productivity benefit from roughly 500,000 hours saved annually by year three. Based on five practitioner interviews and a survey of 166 decision-makers. Vendor-commissioned. (Forrester Consulting for SS&C Blue Prism, April, 2024)
  • A Forrester study of Quickbase modelled 315% ROI over three years, $16.7 million in net present value, payback in under six months and a 15% reduction in wasteful operational spend. Vendor-commissioned. (Forrester Consulting for Quickbase, 2023)
  • A Forrester study of the OutSystems low-code platform modelled 363% three-year ROI, payback in under six months and 60% faster development cycles, based on interviews with six decision-makers. Vendor-commissioned. (Forrester Consulting for OutSystems, 2025)
  • Every one of these commissioned studies lands on the same payback figure: under six months. The consistency is partly methodological, since Forrester's TEI model front-loads labour savings, so treat sub-six-month payback as the vendor-modelled ceiling rather than a typical result. (Forrester Consulting for Microsoft, July, 2024)
  • Deloitte's 2025 Tech Value Survey of 548 decision-makers found 84% of organizations investing in AI reported gaining ROI, and 83% of those investing in data management reported the same. All respondent organizations had at least $500 million in annual revenue. (Deloitte, November, 2025)
  • 45% of Deloitte respondents expect near-term ROI, defined as under three years, from basic automation, while roughly 60% expect longer timeframes for more advanced AI capability levels. (Deloitte, November, 2025)
  • Deloitte found average digital budgets rose from 7.5% of revenue in 2024 to 13.7% in 2025, an average of $1.8 billion against $13.4 billion of average revenue. (Deloitte, November, 2025)

The cost of manual work, spreadsheets and disconnected systems

  • Workers toggled between applications roughly 1,200 times each day, spending just under four hours a week reorienting themselves afterwards, or about 9% of their time at work. The study tracked 20 teams totalling 137 users across three Fortune 500 companies for up to five weeks. (Harvard Business Review, August, 2022)
  • Knowledge workers spend 58% of the day on "work about work," meaning coordination rather than the skilled work they were hired to do, and estimate that improved processes would save them 4.9 hours per week. Survey of 9,615 knowledge workers across six countries. Vendor-commissioned by Asana. (Asana, March, 2023)
  • Knowledge workers used an average of 8.8 apps, rising to 10 for director level and above. 15% of workers using 6 to 15 apps report missing messages and actions, rising to 25% for those using 16 or more. Vendor-commissioned by Asana. (Asana, March, 2023)
  • Desk workers report spending 41% of their time on tasks that are low value, repetitive, or lack meaningful contribution to their core job functions. Survey of 10,281 desk workers across six countries, conducted January 2024. Vendor research by Salesforce-owned Slack. (Slack Workforce Lab, 2024)
  • A Forrester study of Boomi modelled 347% ROI, $9.8 million NPV and payback in under six months from replacing point-to-point integration work, with customers reporting 65% shorter integration timelines. Vendor-commissioned. (Forrester Consulting for Boomi, September, 2025)
  • A Forrester study of MuleSoft modelled 426% ROI and $10.0 million NPV, driven by a 60% reduction in API and integration delivery effort, a 45% asset reuse rate and a 70% reduction in ongoing management effort. Vendor-commissioned. (Forrester Consulting for MuleSoft, October, 2025)

Error rates: what happens when a spreadsheet is the system

  • Across 85 intensive inspection studies, most conducted by commercial spreadsheet auditing firms, errors were found in 94% of the spreadsheets examined. (Raymond Panko, EuSpRIG / arXiv, 2015)
  • Across 14 laboratory studies of spreadsheet development involving 967 participants, the average cell error rate was 3.9%. That is in line with error rates in program statements, measured at 1.9% to 3.7% across four industrial code-inspection corpuses totalling around 10,000 inspections. (Raymond Panko, EuSpRIG / arXiv, 2015)
  • Seven field audits of 367 real organizational spreadsheets found errors in 24% overall, but the audits using the strongest methodologies found errors in at least 86% of spreadsheets examined. Flagged as a pre-2023 source; this is the original field-audit dataset the later literature builds on. (Raymond Panko, EuSpRIG / arXiv, 2000)

Downtime, transformation failure, and the share that can prove value

  • A single hour of downtime costs more than $300,000 for over 90% of mid-size and large enterprises, and 41% of enterprises put hourly downtime cost between $1 million and over $5 million. Survey of more than 1,000 firms worldwide, November 2023 to March 2024. Figures exclude litigation and penalties. (ITIC, 2024)
  • 57% of operators say their most recent major outage cost more than $100,000, and for the second consecutive year one in five say it cost more than $1 million. (Uptime Institute, May, 2026)
  • 76% of logistics transformations never fully succeed, failing to meet budget, timeline or KPI targets. Survey of 306 logistics professionals at organizations with $500 million or more in annual revenue, conducted November to December 2023. (Gartner, August, 2024)
  • More than 80% of those respondents had attempted four transformations in fewer than five years, averaging almost one a year. (Gartner, August, 2024)
  • Leaders who treat internal resistance as a resource rather than an obstacle improved their odds of transformation success by 62%, while the directive "get with the program" approach cut the odds by 47%. Only 20% of surveyed leaders used the former approach. (Gartner, August, 2024)
  • 55% of chief supply chain officers are unclear on the ROI of their AI investments, even though 67% of supply chain digital investment is now allocated to AI. Based on a survey of 394 supply chain professionals conducted November 2025 to February 2026 and 135 senior leaders surveyed January to April 2026. (Gartner, August, 2026)

Methodology

These statistics are built for citation, comparison and budgeting. Here is how sources and numbers were handled.

This page was compiled and reviewed in August 2026. It prioritizes primary sources: standards bodies and project repositories such as ISBSG, national audit offices, government statistics agencies including BLS, Eurostat and OECD, official vendor pricing pages and help centers, SEC filings and earnings calls, and original survey research from Gartner, IDC, McKinsey, Deloitte, Forrester, Stack Overflow, JetBrains, GitHub and Stanford HAI.

Every statistic was verified by opening the source page and confirming the number appears there. A claim quoted in a secondary article but absent from the cited original was rejected. So was any claim whose only trail led to an agency marketing page, an undated listicle or a URL that no longer resolves. One hundred and thirteen claims were rejected on those grounds during compilation. The rejections are documented per section in our working files, and the most consequential is worth naming: the widely repeated Gartner attribution that 70% of new applications would use low-code by 2025 could not be confirmed on any live Gartner page.

Prices are in US dollars before tax unless the source states otherwise, and reflect published list prices at the time of review. Platform pricing changes often, and quote-only enterprise tiers are labeled as such. Where a source reports in another currency, the original currency is preserved rather than converted, because a conversion at an unstated exchange rate creates false precision.

We do not publish a single "average cost to build an app." The projects in these datasets differ by an order of magnitude in scope, and averaging them would produce a number that is confidently wrong. Cost is reported against the variable that drives it: functional size, complexity band, region, team composition or platform.

Where no defensible primary data existed for a sub-topic, that sub-topic was left out rather than filled with an estimate.

This page is reviewed and updated monthly. Platform pricing is re-checked against vendor pricing pages at each review.

The real question is not what an app costs

The range you get quoted says more about who is quoting than about what you are building.

A build has a defensible cost once three things are settled: what it has to do, who it has to serve, and what happens to it after launch. Most budgets go wrong because the third question never gets asked, and the maintenance, platform fees and technical debt in the sections above arrive as surprises.

At LOW / CODE we have delivered 450+ projects for companies including Sotheby's, American Express, GAF, Zapier, Medtronic and Coca-Cola. We pick the tool that fits the problem, which sometimes means Bubble or FlutterFlow and sometimes means React and Supabase, and we scope against what the thing has to do rather than against a template.

Internal business apps start at $15,000. SaaS products, marketplaces and native mobile apps start at $30,000. Websites start at $9,000.

If you want a number for your specific build rather than a range, talk to us.

Last updated on 

August 25, 2026

.

Jesus Vargas

Jesus Vargas

 - 

Founder

Jesus is a visionary entrepreneur and tech expert. After nearly a decade working in web development, he founded LOW/CODE Agency to help businesses optimize their operations through custom software solutions. 

Custom Automation Solutions

Save Hours Every Week

We automate your daily operations, save you 100+ hours a month, and position your business to scale effortlessly.

FAQs

What is the average cost to build an app in 2026?

Why do app development cost estimates vary so widely?

Does low-code development cost less than traditional development?

What percentage of software budget goes to maintenance?

Do AI coding tools actually reduce development costs?

What causes app development projects to go over budget?

Watch the full conversation between Jesus Vargas and Kristin Kenzie

Honest talk on no-code myths, AI realities, pricing mistakes, and what 330+ apps taught us.
We’re making this video available to our close network first! Drop your email and see it instantly.

Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.

Why customers trust us for no-code development

Expertise
We’ve built 330+ amazing projects with no-code.
Process
Our process-oriented approach ensures a stress-free experience.
Support
With a 30+ strong team, we’ll support your business growth.