What is pre-seed funding, and how much engineering can it actually buy?

Pre seed funding is the first institutional money a startup raises, usually before it has the traction investors expect at the seed stage. Its purpose is to fund the work needed to prove the business and build enough evidence to raise the next round.

The important question is not how large the funding round looks. It is how much engineering time that money actually buys.

At the 2025 US median software developer wage of $135,980, every $100,000 of engineering budget covers about 8.8 months of one developer's wages. Once employer costs are included, that falls to about 6.2 months. At the 75th percentile, the same budget buys only about 4.9 months.

That makes engineer months the real constraint. A pre seed budget can look substantial on paper while covering surprisingly little development time.

The role also matters. "Engineer" covers several US occupation codes with a 47% pay spread, so your budget needs to reflect the type of developer you actually need. Offshore hiring can stretch the same budget further, but the difference reflects labour markets rather than a simple discount.

Technical debt also matters. Research shows that startups can accumulate technical debt deliberately when speed matters. The challenge is knowing which shortcuts support the next milestone and which ones will slow the company down later.

AI tools add another variable. They have become a regular engineering expense, but they do not reliably eliminate the need for developers.

The real question for any pre seed startup is therefore simple: how many useful engineer months can your funding buy, and what evidence will those months produce?

In this article

• What is pre seed funding?

• How Do You Convert a Pre Seed Round Into Engineering Time?

• Use the median, not the mean.

• The word “engineer” covers four different prices

• The salary is not the cost

• The full conversion table

• Why is the same engineer month priced so differently around the world?

• What does the money get spent on now?

• What does a pre seed engineering plan usually get wrong?

• How many engineers should the money buy?

• How do you turn the round into an engineering plan?

• What decision does each stage force?

• How do you know you are ready to raise pre-seed

• Where does vetted hiring fit in the arithmentic

• Conclusion

• FAQ

What Is Pre Seed Funding?

Pre seed is the earliest institutional funding a startup raises. It usually comes before the company has the revenue or user traction that seed investors expect.

The purpose is straightforward: fund a small team for a defined period and produce the evidence needed to raise the next round.

Pre seed rounds vary widely by geography, sector and founder experience. That makes a single "average" round size less useful than it might appear. Published figures often come from commercial datasets that do not disclose enough about their methodology to justify treating the number as precise.

The venture market itself also affects these figures. Josh Lerner and Ramana Nanda noted in the Journal of Economic Perspectives that a relatively small group of venture capital investors controls and influences a substantial share of capital for financing radical technological change. When capital is concentrated among a smaller number of firms, funding averages can reflect the behaviour of those investors rather than what a typical startup should expect.

For funding ranges and stage specific investor lists, see typical VC funding amounts at seed level, pre seed VC firms, top seed VC firms, early stage VC firms, and Series A VC firms.

The more useful question for this article is what that funding can actually buy, particularly when you translate the budget into engineering time.

How Do You Convert a Pre Seed Round Into Engineering Time?

Start with the budget. Divide it by the engineer's annual wage. Then account for the costs founders often leave out.

The US Bureau of Labor Statistics publishes wage data across the full software developer distribution. In 2025, the occupation included 1,687,890 workers. The figures below use annual wages, with the conversion calculated as $100,000 divided by the annual wage, multiplied by 12.

The row you plan against matters more than the size of the funding round. The same $100,000 buys about 63% more engineer months at the 25th percentile than at the 75th percentile.

That difference becomes important quickly when a pre seed runway depends on getting a product built before the next funding milestone.

Use the Median, Not the Mean

The BLS mean annual wage for software developers in 2025 was $148,100, about 9% higher than the median of $135,980. If you plan around the mean, $100,000 buys about 8.1 engineer months instead of 8.8.

That difference costs you almost a month of engineering time for every $100,000. The mean is higher because very highly paid senior developers pull the average upward.

For an early stage startup, the median is therefore a more useful planning figure than the mean. It gives you a better estimate of what the typical developer costs and how much engineering time your budget can actually buy.

The Word Engineer Covers Different Prices

Founders often put "engineer" in a budget without specifying the role. That can make the forecast look more precise than it really is.

The BLS prices software occupations separately, and the 2025 median wages vary significantly:

The difference between the highest and lowest median is about 47%. The same $100,000 can therefore buy very different amounts of engineering time depending on the role.

Define the occupation before setting the budget. Otherwise, the wrong assumption can quietly shorten or extend your expected runway.

The Salary Is Not the Cost

The wage figures above show what an engineer earns. They do not show what the employer actually pays.

The US Bureau of Labor Statistics measures total compensation separately. In the first quarter of 2026, private industry employers paid an average of $46.60 per hour in total compensation. Wages and salaries accounted for $32.60, while benefits added another $14.01. Wages therefore represented 69.9% of total compensation.

The same pattern appears in Europe. Eurostat found that information and communication businesses across the EU27 had an average hourly labour cost of €48.20 in 2025, compared with €37.00 in wages and salaries. That means a meaningful share of the employer's cost never appears in the employee's salary.

These figures are not directly comparable. The BLS figure covers US private industry, while Eurostat's figure covers the wider information and communication sector across the EU27. They are useful here because both show the same underlying point: salary alone understates the cost of employing someone.

Using the BLS ratio, the $135,980 median software developer wage becomes about $194,500 in employer cost. That reduces the amount $100,000 buys from about 8.8 engineer months to roughly 6.2.

At the 75th percentile, the $171,980 wage becomes about $246,000 in employer cost. The same $100,000 then buys roughly 4.9 engineer months.

Those are the more realistic figures to use when planning the cost of a traditional employee.

The Full Conversion Table

Once you account for employer costs, the same $100,000 can buy very different amounts of engineering time depending on how you hire.

The offshore figures use RocketDevs' published rates. These are company reported figures and have not been independently audited. The calculation is straightforward: divide $100,000 by the hourly rate, then divide by 160 hours per month.

At $21.99 per hour, that works out to $3,518.40 per month and about 28.4 engineer months. At $9.99 per hour, it is $1,598.40 per month and about 62.6 engineer months.

The point is not that cheaper engineering is always better. It is that pre seed funding is ultimately a time budget expressed in dollars.

The difference between roughly 6.2 engineer months for a grossed up US median employee and 62.6 months at the lowest listed contract rate is substantial. Whether that extra time is useful depends on the work, the level of supervision required and how well the developer fits the role.

Hiring time also belongs in the calculation. Arc.dev says it can find and hire pre vetted candidates in as little as 14 days. That is a company claim rather than an independent measurement, but even two weeks can reduce your effective runway before development begins.

Why Is the Same Engineer Month Priced So Differently Around the World?

The difference comes from the labour market you are hiring in. You are not simply negotiating a discount on the same engineer.

Several official and industry datasets show how much engineering costs can vary by country. They measure different things, so the figures should not be treated as directly interchangeable.

The World Bank figures show a much wider economic gap than developer compensation does. GDP per capita differs by roughly 33 times between the US and India. Stack Overflow's occupation level figures show a much smaller gap in developer compensation.

That distinction matters. Software skills have a more global labour market than the economies surrounding them. This makes international hiring a genuine way to extend an engineering budget, but the saving is usually nowhere near the difference suggested by GDP comparisons.

Using the Stack Overflow figures, $100,000 buys about 6.9 engineer months for a US back end developer. The same amount buys about 13.8 months at the German median and about 54 months at the Indian median. These figures are before employer costs, agency fees and management overhead.

Geography also matters within a single country. The BLS 2025 median wages for software developers ranged from $104,490 in Arkansas and $105,600 in North Dakota to $132,150 in Texas, $166,180 in New York and $174,410 in California. California alone had 284,390 software developers.

That is a 1.67 times difference without crossing an international border. Geography is therefore a labour market variable even when you hire domestically.

What Does the Money Get Spent On Now?

Increasingly, part of the engineering budget goes toward AI tools and the infrastructure behind them.

This is still new enough that many pre seed budgets do not include a dedicated line for AI usage. The evidence suggests that the cost is real. Stack Overflow's 2025 Developer Survey collected 49,009 responses from 177 countries and found that 84% of respondents were using or planning to use AI tools in their development process. It also found that 51% of professional developers use them daily.

GitHub's 2025 platform data points in the same direction. It recorded 582,196 new Python AI repositories during the year, an increase of 50.7% from the previous year. That represented close to half of all new AI repositories among 121 million new repositories overall. GitHub both publishes and operates the platform behind this data, so repository counts are useful as a measure of activity rather than proof that AI improves engineering outcomes.

Package registry activity shows that AI development tools are also being used at significant scale.In the month to 19 August 2026 the OpenAI npm package recorded 136,820,040 downloads and the Anthropic SDK 122,205,639, both measured directly from the npm registry. Anthropic's Python package on PyPI recorded about 201 million downloads.

The cost of this usage has created a market for tools designed to reduce it. One example is rtk, a CLI proxy that aims to reduce language model token consumption during routine development commands. The project claims reductions of 60% to 90%, although that figure has not been independently verified. Its adoption is harder evidence: the project had 76,448 stars and 4,799 forks within months of its January 2026 launch.

The important point is not the exact saving. It is that AI usage has become significant enough for teams to build dedicated tooling around its cost.

There is another assumption worth challenging: that AI automatically reduces the number of engineers you need.

Stack Overflow's survey found that 52% of developers either do not use agents or use simpler AI tools. It also found that 38% had no plans to adopt agents. Developers were particularly reluctant to use AI for higher consequence work. The survey found that 76% did not plan to use AI for deployment and monitoring, while 69% did not plan to use it for project planning.

These findings come from a single survey, so they should not be treated as settled industry consensus. They do support a practical planning rule, though. If your runway model assumes AI agents will take over operational work, keep human engineering capacity in the budget.

AI can change how much work an engineer completes. It does not make the engineer disappear from the cost model.

What Does a Pre Seed Engineering Plan Usually Get Wrong?

It assumes the speed it buys is free. Peer reviewed software engineering research suggests otherwise.

Giardino and colleagues, in their Greenfield Startup Model published in IEEE Transactions on Software Engineering, studied how startups develop software under pressure. They found that the need to reach the market quickly through lower precision engineering is eventually balanced by the need to restructure the product before pursuing further growth.

In other words, speed at pre seed comes with a cost. You may move faster now, but some of that work will need to be revisited when the product and company start to scale.

A separate study by Besker, Martini, Lokuge and Blincoe reached a similar conclusion. Their 2018 IEEE International Conference on Software Maintenance and Evolution study interviewed 16 professionals across seven software startups. They found that startups deliberately accumulate technical debt based on factors such as the stage of the company, developer experience, founders' software knowledge and employee growth. The teams aimed for a "good enough" level of software quality that allowed them to move forward without resolving every technical issue immediately.

Neither study examined AI assisted development, so their findings should not be treated as evidence about AI specifically. The underlying mechanism is broader. Startups make engineering trade offs because they are operating with limited time, money and information.

The lesson for a pre seed plan is not to eliminate technical debt. Some of it may be exactly the right trade off. The important thing is to make those decisions deliberately, know where the debt sits and include the cost of paying it back in the seed plan.

Otherwise, the engineering speed that helped you reach the next round can become part of the work that slows you down once you get there.

How many engineers should the money buy?

As few as produce the evidence, and none you cannot review.

There is more behind that than folklore. Mao, Mason, Suri and Watts ran a controlled experiment with 47 teams ranging from one to 32 people, published in PLOS ONE in 2016. They found that individuals in teams exerted less overall effort than independent workers, partly because they shifted effort toward less demanding and less productive tasks. Collaboration increased with team size, but so did the coordination cost. Adding people is a trade, not a straight gain. That trade gets worse when the founder cannot evaluate the work.

At pre seed, the practical version is simpler. Every additional engineer shortens the runway and increases the amount of work the founder must review. If one person is responsible for understanding what the team is building, more engineers can quickly create more output than that person can properly assess.

Size the team to the evidence you need. Then check the review path before checking the budget. If you are sizing for the next round rather than this one, see how many engineers you need to raise a seed round.

Pre seed does not buy you a finished product. It buys you time to find out whether you were right. The real question is how much evidence your budget can buy before the runway runs out.

How do you turn the round into an engineering plan?

Seven steps, in order. Each one should produce a number you can write down, not a feeling.

• Name the evidence. Write the single sentence that must be true when you next raise. “Thirty paying teams retaining above 60%” is evidence. “The product will be better” is not.

• Set the engineering allocation. Decide what share of the round goes to engineering rather than founder salary, tooling, legal or go to market. Everything that follows is calculated from this amount, not from the total round.

• Pick the occupation, not the label. Decide whether you need a software developer, web developer, QA specialist or database administrator. The 2025 BLS medians for these four roles differ by 47%.

• Convert the allocation into months. Divide the engineering budget by annual cost, then multiply by 12. Use the employer's total cost when hiring employees, not just the wage. For contractors, use the quoted hourly rate multiplied by 160.

• Subtract hiring latency. Two weeks may be the fastest advertised hiring timeline. Treat those weeks as lost capacity and subtract them from each engineer's available months before planning.

• Reserve for tooling and restructuring. Give model and development tooling a real budget line and revisit it quarterly. If research suggests the team will need restructuring, account for it in the plan rather than letting it become a surprise during diligence.

• Name the reviewer. Write down who approves the code and how many hours they have each week. If that box is empty, no hiring rate makes the arithmetic work.

What decision does each stage force?

How do you know you are ready to raise pre-seed?

You are ready when you can name the evidence the next round needs and show how this money will produce it.

Ask yourself three questions before you start the raise:

• What specific thing will be true in nine months that is not true now? If the answer is “the product will be better,” you are not ready. If it is “we will have 30 paying teams retaining above 60%,” you have a measurable target.

• How many engineer months will that take, and how confident are you? Convert the target into engineering capacity first. Then check whether the budget covers it with enough margin for the estimate to be wrong. Account for hiring latency too.

• Who reviews the code? This is the question founders often skip. Engineering capacity you cannot evaluate is not really capacity. It is risk with a payroll attached.

That third question is where the cost calculation can quietly break. A cheaper engineer month only helps if the output is usable. If you are a non technical founder without a reliable review process, you may discover problems only after the money has been spent.

Before you spend the first month, it is worth reading how to evaluate a developer when you cannot read the code.

Where does vetted hiring fit in the arithmetic?

It can change the denominator without changing the risk profile, provided the vetting filter is real.

RocketDevs rates start at $9.99 per hour for Associate developers, $21.99 for Mid senior developers and $30.99 for Senior developers. The company says each developer completes 6 to 8 hours of assessment, with a 98%+ rejection rate and a Top 2% acceptance bar. These are published RocketDevs figures and are not independently audited.

For a pre seed team, the most important feature may not be the headline rate. It is the 14-day money back trial. At this stage, a bad hire can cost months of runway that you cannot recover. Testing the match on real work for two weeks gives you an earlier signal before committing more of the budget.

The principle is simple: reduce the cost of a wrong decision before trying to reduce the cost of an engineer.

The longer term version of this approach matters once you move beyond the first hires. At that point, the question becomes how to scale the team without letting headcount grow faster than the evidence.

Conclusion

A pre seed round is not an engineering budget. It is a limited amount of time to prove that the company deserves more capital.

That changes how you should think about hiring. The goal is not to build the biggest engineering team you can afford. It is to buy enough capacity to produce the evidence your next round requires while keeping enough runway to respond when your assumptions are wrong.

Start with the milestone. Work backwards to the engineering capacity it requires. Price the actual cost of that capacity, account for hiring time and tooling, then make sure someone can review the work.

The cheapest engineer is not always the cheapest option. A lower hourly rate means very little if the work takes longer, requires expensive rework or cannot be evaluated until months later.

Good pre seed planning is therefore less about predicting the perfect team. It is about making the assumptions visible, testing them early and protecting the runway when they are wrong.

Spend enough to learn. Hire only as fast as the evidence demands. And never confuse more engineers with more progress.

Frequently asked questions

• What is pre seed funding used for?

Pre seed funding is used to produce the evidence a seed investor will want. That could be usage, retention, revenue or a technical result that reduces the risk around your thesis. The goal is to fund a small team for a defined period, not to build a finished product.

• How much is a typical pre seed round?

Round sizes vary widely by geography, sector and founder history. Published averages also come from commercial datasets whose methodology is not always disclosed. Rather than quote a number that suggests more precision than the data supports, convert your raise into engineer months and plan from there.

• How many engineer months does $100,000 buy?

At the 2025 US median software developer wage of $135,980, $100,000 buys about 8.8 engineer months before employer costs. Using the officially measured employer cost ratio of 69.9%, that falls to about 6.2 months.

At published vetted offshore rates, the same $100,000 converts to about 28 months at $21.99 per hour or about 63 months at $9.99 per hour, assuming 160 hours per month.

• How do you get pre seed funding?

Start by naming the evidence your next round will require. Then show how your plan will produce that evidence within the runway you are asking investors to fund. Focus your outreach on investors who regularly fund companies at the pre seed stage.

• How many engineers should a pre seed startup hire?

As few as produce the evidence, and none you cannot review. A controlled experiment across 47 teams found that people in teams exerted less individual effort than independent workers. Engineering capacity without a reliable review path can therefore become risk rather than progress.

• Is offshore hiring sensible at pre seed?

It can be. At pre seed, your constraint is often engineer months rather than dollars, so the same budget can buy substantially more capacity through offshore hiring.

That only works if someone on your side can evaluate the output. If you cannot review the work, the first problem to solve is not the hiring rate. It is the review process.

• Does AI reduce the number of engineers I need to fund?

Not reliably. Stack Overflow's 2025 survey found that 52% of developers were either not using AI agents or were using simpler AI assistant tools. Developers also showed stronger resistance to using AI for tasks such as deployment, monitoring and project planning.

Budget for AI tooling where it improves productivity. Do not assume it automatically means you can fund fewer engineers.

Start with a vetted developer.

Sources

• US Bureau of Labor Statistics, Occupational Employment and Wage Statistics, Software Developers (occupation 15-1252), 2025: the annual median of $135,980, the annual mean of $148,100, the 10th, 25th, 75th and 90th percentile series, and total employment of 1,687,890. Adjacent occupations 15-1251, 15-1253, 15-1242 and 15-1254 web developers, and the state series including California, New York, Texas, Arkansas and North Dakota, from the same source. Retrieved via the BLS public API, accessed 18 August 2026.
• US Bureau of Labor Statistics, Employer Costs for Employee Compensation, private industry, 2026 Q1: total compensation $46.60 an hour, wages and salaries $32.60 at 69.9%, benefits $14.01 at 30.1%. Accessed 18 August 2026.
• Eurostat, labour cost levels by NACE Rev. 2 activity, information and communication sector, 2025. Sector-level, not occupation-level. Accessed 18 August 2026.
• OECD, average annual wages, 2024, USD at purchasing power parity, via the SDMX public API. Economy-wide across all employees, not occupation-level. Accessed 18 August 2026.
• World Bank, GDP per capita, current US dollars, indicator NY.GDP.PCAP.CD, 2024. Accessed 18 August 2026.
• Stack Overflow, 2025 Developer Survey, 49,009 responses from 177 countries fielded 29 May to 23 June 2025: the AI section and the work and salary section. Accessed 18 August 2026.
• GitHub, Octoverse 2025, platform telemetry. GitHub sells AI coding tools and publishes this report. Accessed 18 August 2026.
• npm registry downloads API and the pypistats.org recent-downloads API, accessed 18 August 2026.
• rtk-ai/rtk, repository metrics via the GitHub API, accessed 18 August 2026. The token-reduction figure is the project's own claim.
• Giardino, Paternoster, Unterkalmsteiner, Gorschek and Abrahamsson, Software Development in Startup Companies: The Greenfield Startup Model, IEEE Transactions on Software Engineering 42(6):585-604.
• Besker, Martini, Lokuge and Blincoe, Embracing Technical Debt, from a Startup Company Perspective, IEEE ICSME 2018.
• Mao, Mason, Suri and Watts, An Experimental Study of Team Size and Performance on a Complex Task, PLOS ONE 11(4):e0153048, 2016.
• Lerner and Nanda, Venture Capital's Role in Financing Innovation, Journal of Economic Perspectives 34(3):237-261, 2020.
• Arc.dev hiring page, accessed 18 August 2026. The 14-day figure is Arc.dev's own claim.
• RocketDevs rates, vetting depth, rejection rate and trial terms are our own published figures and have not been third-party audited.

James Hitch, COO at RocketDevs