Financing an AI project in Israel: a roadmap across grants, credit and equity

A question we get in almost every first meeting: "is there a grant for this?" The answer is nearly always "probably, but that is not the first question to ask". A development program is measured by cash flow over time, and a grant is one component inside a structure - not a substitute for one.

Note: tracks, support rates and eligibility conditions change from time to time. What follows is a roadmap in principle; check the current terms of the relevant track before deciding.

The three sources

Equity. The most flexible and the most expensive in terms of dilution or private cash. Suited to the stage where it is not yet clear what is being built.

Credit. Bank or non-bank, sometimes state-guaranteed. Advantage: arrives relatively quickly and can be planned. Disadvantage: requires demonstrated repayment capacity and sometimes collateral.

Grants and support programs. The cheapest money and also the slowest. Submission, committee, approval and reporting run on a timeline of months. Grants typically reimburse part of recognized expenditure after the fact, so they improve the economics without solving present-day cash flow.

The most common cash-flow mistake

Companies plan the project assuming the grant will arrive, and start developing. The grant is approved - next quarter, and paid against reports of actual spend. Meanwhile salaries are paid every 30 days. That gap between spending and receipt is what breaks good projects.

The standard fix: a bridge credit line planned in advance against the grant's reporting schedule. You plan it before submission, not after the gap opens.

A four-stage roadmap

1. Define recognized expenditure. Before looking at tracks, break the budget into R&D salaries, subcontractors, equipment and infrastructure. Some tracks recognize only some categories, and that changes the economics.

2. Match the track. There are tracks for technological development, for early-stage companies, for internationally collaborative projects, for exporters and for priority regions. The match depends on company stage, type of development and location of activity.

3. Build the dossier. A technical description explaining the innovation and the technological risk, alongside an economic file: budget, forecast, ability to complete. The two parts are assessed separately and must tell the same story.

4. Plan the receipts. The reporting schedule, bridge financing, and alignment of project milestones with reporting milestones.

Worth knowing up front

• A grant is money with obligations: reporting, audit, sometimes limits on transferring know-how abroad
• Preparing a serious dossier takes weeks, not days
• A rejection is not the end - usually you can improve and resubmit, sometimes to a better-suited track
• AI project financing is now assessed on know-how and data too, not only code: who owns the data, and what remains with the company at the end

How we work on this

Here, the financial side and the technology side sit in the same meeting. The team that defines the architecture defines the financing structure, so the technical schedule and the receipts schedule are planned together. It is the only approach we have seen that prevents the gap that sinks projects in month seven.