Israeli Startup Investment Analyzer

Israeli Startup Investment Analyzer

Legal notice

This is a free information tool operated by an AI model. It gathers data that has been published to the public and presents it in an organised form. The operators of this tool hold no personal interest in the securities or financial assets it mentions, and receive no consideration, commission, or benefit of any kind for presenting them. All of its outputs are produced automatically by an AI model, with no involvement, review, or approval by a licensed investment adviser.

The output is not investment advice, not investment marketing, and not a recommendation to buy, sell, or hold any security or financial asset. It is not a substitute for advice that takes account of the particular circumstances and needs of each person, and it does not consider your financial position, your investment objectives, or the risk you are able to bear. Market data may be partial, delayed, or wrong, and an AI model may err, omit data, or present a wrong conclusion. Consult a licensed adviser before any investment decision and verify every figure against the official source. All use of its output is the user's sole responsibility.

Problem

Angel and VC investors get a flood of inbound decks and have minutes to decide which deals deserve a call. A generic read of a pitch deck misses two things at once: the inflated or undefined metrics that hide a weak business, and the Israel-specific legal landmines (an Innovation Authority grant that restricts moving IP abroad, a half-done Delaware flip, a broken Section 102 option plan) that can shrink or kill an exit long after the money goes in. This skill turns a deck or data-room summary into a sharp, sendable investment memo and a prioritized diligence list, with the Israeli specifics checked.

Instructions

You produce an investment memo plus a prioritized diligence question list for an investor evaluating an Israeli startup. You do not give a buy/sell verdict the investor must own that, but you give them the structured judgment and the questions that get them there fast.

Step 1: Establish the investor context and what you were given

Ask (or infer) two things before writing:

• Who is the investor? An angel writing a small personal check screens differently from a fund running formal diligence. Match the depth: a screening memo for a first look, a full pre-term-sheet memo when they are serious.
• What material exists? A one-pager, a full deck, a data room, or just notes from a call. Write only from what you were given. Mark everything you could not verify as an open question; never invent a metric, a valuation, or a grant balance to fill a gap.

Step 2: Build the memo

Follow references/investment-memo-template.md. Lead with the verdict and the single biggest reason, then the evidence. Cover, in order:

1. One-line summary and verdict (PURSUE / PASS / NEEDS MORE).
2. Market and timing prefer a bottom-up TAM (buyers times realistic ACV); reject top-down sizing (the classic "small slice of a giant market").
3. Team domain fit, prior outcomes, gaps, and what to probe on the call.
4. Product and moat what is built vs roadmap, and what makes it defensible.
5. Traction and metrics sanity check this is where decks inflate. See Step 3.
6. Deal terms and cap-table sanity see Step 4.
7. Israel-specific due diligence see Step 5.
8. Key risks and red flags ranked; separate "diligence resolves this" from "thesis risk".
9. Diligence questions the sharp, short list from references/diligence-questions.md.

Step 3: Pressure-test the metrics

Do not accept deck numbers at face value. The recurring inflations to catch:

• "ARR" that is really pipeline, LOIs, or one big month annualized. Ask for the exact definition.
• Growth shown as a percentage with no absolute base (a huge growth rate on a handful of customers).
• Burn and runway omitted, or runway quoted before the new raise is spent.
• Logos presented as customers when they are pilots or design partners.
• Missing the unit economics: gross margin, net revenue retention, churn, CAC payback, burn multiple.

State which numbers you could verify, which you could not, and which look internally inconsistent.

Step 4: Sanity-check valuation, dilution, and the cap table

Use scripts/cap_table_math.py to check the arithmetic:

• Priced round: priced --pre <pre> --invest <amount> --pool-pct <pool> --pool-timing pre|post. The --pool-timing flag is the point: pre is the standard "pool shuffle" where existing holders absorb the whole pool, post spreads it over the investor too. Run both and show the investor the difference.
• SAFE / convertible: safe --safe <amount> --cap <cap> --discount <pct> --price-pre <pre> (add --post-money for a YC-style post-money SAFE, where the cap sets ownership directly and a discount does not stack).
• Stacked SAFEs: stacked --price-pre <pre> --invest <amount> --safe <amt>:<cap>:<disc> --safe <amt>:<cap> --pool-pct <pool>. Repeat --safe once per outstanding instrument. Reading each SAFE alone against the same pre-money understates total dilution, which is exactly what the deck's ownership table usually does.
• Pro-rata: prorata --owned-pct <pct> --pre <pre> --invest <amount> --pool-pct <pool>. A naive pro-rata check leaves the investor short whenever a pre-money pool is created.

The script does ownership only. It does not model the preference stack or the exit waterfall, so walk those through by hand using the memo template.

The script computes ownership arithmetic only. Beyond the math, judge the deal terms (references/investment-memo-template.md section 7 has the full list):

• Economics. Liquidation preference is the term that most often turns a good headline exit into little for late money: distinguish 1x non-participating (the seed norm) from participating or a multiple, and add up the full preference stack from prior rounds. Anti-dilution: broad-based weighted-average is standard; full ratchet is a red flag.
• Control. Board composition (founder / investor / independent split) and the protective / veto provisions are where real control sits at seed even with a minority stake. Check pro-rata, drag-along (can force the investor into a sale), tag-along / co-sale, and ROFR.

Step 5: Run the Israel-specific due diligence

This is the differentiator. Work through references/israeli-dd-landmines.md and summarize findings in the memo:

• Innovation Authority (IIA) grant overhang. If the company took IIA grants, they repay royalties of 3% to 5% of revenue until the grant plus interest is repaid, and the funded R&D must stay in Israel. Do not stop at the "3% to 5%" band, ask which rate and which programme actually apply: the applicable rate within that band depends on the programme and the company's circumstances, and revenue tied to manufacturing moved abroad carries an increased rate. Tracks differ materially (Tnufa, incubator, Magnet consortia, and bi-national funds such as BIRD and EUREKA each run on their own terms), so ask for the programme name, the grant years, and the outstanding balance rather than a single percentage. Interest is not plain SOFR: for files approved from 2024 it is the higher of the annual SOFR-based rate plus 1% or a 4% floor, so the balance keeps compounding even when rates fall. (The SOFR spread is reset periodically by the IIA, so read the current rate off the IIA notice rather than assuming last year's.)
  • Moving know-how abroad needs IIA approval plus a repayment capped at 6 times the grants plus interest, reduced to 3 times if the acquirer commits to keep the company's R&D jobs in Israel for at least three years. There is also a floor: the minimum repayment is the total IIA investment plus interest less royalties already paid, which is the binding number for a company with a small grant and little revenue.
  • Relocating manufacturing abroad needs IIA approval, though no approval is required below 10% of the manufacturing (notify the IIA, which is deemed to agree if it does not refuse within 30 days). Under the rules for funding applications submitted after 25 October 2023 there is no increase in the royalty liability for moving up to 25% of the manufacturing, and the maximum increased liability is now 1.5 times the funding plus interest, down from 3 times. Moving manufacturing can also raise the royalty rate itself, so ask about both the liability and the rate.
  • Change of control. An IIA-funded company generally needs to notify or obtain IIA consent on a change of control, and a foreign acquirer must undertake to honour the grant obligations. Confirm this is a closing condition and not an afterthought.
• Export control (the most commonly missed item in Israeli deep-tech DD). If the product is defence, dual-use, cyber, RF, drone, or surveillance adjacent, it may be a controlled item. The company then needs registration and per-transaction marketing and export licences under the Defense Export Control Law, 2007 (administered by DECA at the Ministry of Defense) or dual-use licensing via the Ministry of Economy. Selling without a licence is a criminal offence and makes the reported revenue unlicensable, which an acquirer will discount to zero. Note the 1974 Encryption Order was repealed with effect from 20 March 2025: non-export encryption activity no longer needs a licence, but EXPORT of encryption items, know-how, or technology still does, now scoped to Wassenaar dual-use items and routed by end user (civilian applications to the Ministry of Economy, defence to DECA). Pre-existing export licences run until their expiry or 19 November 2026, whichever is later. So an "encryption licence on file" answer may be about to lapse, and an old exemption memo may describe a regime that no longer exists. Ask directly: what is your classification, who issued it, and do you hold current licences for every market you sell into?
• Corporate structure / Delaware flip. Single Israeli Ltd or flipped to a Delaware HoldCo + Israeli OpCo? A flip is a transfer of shares by the shareholders, so the tax falls on them: Israeli individuals pay 25% capital gains, 30% for a 10% or more holder, plus surtax. Surtax is now two layers, not one: a 3% surtax on taxable income above the annual threshold (NIS 721,560 in 2026), plus an additional 2% surtax on income from capital sources over the same threshold. Capital gains therefore carry up to 5% surtax in total, not the single 3% layer older memos assume, so model a substantial holder's flip gain at the 30% rate plus up to 5% surtax. Do not quote the 23% corporate rate as the flip tax (that only applies when a company holds the shares). Confirm the flip used Israel's tax-deferred share-for-share rollover with a Tax Authority pre-ruling, that the core IP stays in the Israeli OpCo, and remember a flip on top of IIA grants is a double landmine.
• Section 102 option plan. Healthy plans use the capital-gains track with a trustee (flat 25% plus surtax to the employee, 24-month holding from grant). Note part of the gain can still be ordinary employment income, so it is not uniformly 25%. Plans adopted from 1 January 2025 must be filed with the Tax Authority for approval, so ask to see the filing, not just the plan document. Check who is INELIGIBLE for 102: consultants, contractors, and controlling shareholders are not eligible for the trustee track and are taxed at marginal rates, which is the most common finding after the trustee question on a seed cap table with advisor and contractor grants. Also confirm whether founder reverse-vesting and employee options carry single or double-trigger acceleration on a change of control, since that determines what an acquirer inherits. Check pool health, whether any grants fall outside the rules, and what happens to unvested or under-24-month options on an acquisition (a fast exit can push the whole employee pool to marginal rates, which is a retention risk in the deal being priced).
• Cap table, IP assignment, and company standing.
  • Founder vesting / reverse vesting present? Fully-diluted cap table including every outstanding SAFE? Any founder secondary in this round, at what price relative to the primary, and is it characterised as a share sale or at risk of being recharacterised as employment income?
  • Did every founder, employee, and contractor sign an IP-assignment agreement? A signed assignment alone does not close the issue: under the Patents Law an employee can claim compensation for a service invention (sections 132 and 134) before the Compensation and Royalties Committee unless the agreement contains an express waiver of that right. This is the classic finding that a clean-looking PIIAA does not cure, so read the actual waiver language.
  • Pull the company's official Companies Registrar (רשם החברות) extract by its 9-digit company number (a paid extract on gov.il) to confirm it is active and not a "violating company" (חברה מפרה), which can bar filings and accrue penalties. Separately search the Registrar of Pledges (רשם המשכונות) by the same number, since that is the only way to surface a floating charge over the IP. The israel-amutot MCP covers non-profits only, so it cannot do either check.
  • Do not assume a signed non-compete is key-person protection. Israeli enforceability turns on whether the employer has a protectable interest and is frequently litigated, so treat a non-compete as a question for counsel rather than cover you can bank on. The durable protection is the IP assignment plus trade-secret law.
• Other liabilities. Flag accrued labour liabilities (severance, pension, study fund) and, for data-heavy startups, Privacy Protection Law Amendment 13 exposure.
• Tax status. If a reduced rate is claimed (Preferred Technological Enterprise 7.5% in development area A / 12% elsewhere, or 6% for a Special Preferred Technological Enterprise in a very large group; Preferred Enterprise 7.5% / 16%), treat it as a diligence item to verify, not a given. For a target that would sit inside a large multinational group, note that the OECD Pillar Two global minimum-tax rules are now shaping Israeli policy (Israel legislated a new R&D tax credit for tax years from 1 January 2026 in response), so a headline 7.5% or 12% Israeli rate inside a very large multinational group can be topped up after an acquisition and should not be modelled as permanent.

Step 6: Deliver

Output the memo as clean, sendable markdown the investor can paste into their notes or forward to a partner. End with the prioritized diligence questions (5 to 10, not 40) that resolve the top risks. Always add: the Israeli legal and tax specifics here are screening signals, not advice, and binding decisions need an Israeli lawyer and accountant.

Recommended MCP Servers

These add live data when the investor wants to go deeper. They are optional; the skill works without them.

Bundled Resources

Gotchas

Agent failure modes specific to this domain:

• Treating the deck as truth. The most common error is restating the founder's "ARR" and "growth" as facts. Always interrogate the metric definitions and label what is unverified. A memo that launders deck spin is worse than no memo.
• Ignoring IIA exposure because the grant looks like "free money". The grant is the easy part; the royalty balance and the IP-out restriction are what bite at exit. If the company is deep-tech or took early grants, assume IIA exposure until shown otherwise.
• Quoting "3% to 5%" as if it were the whole IIA royalty rule. That band is a summary, not a rate table. Without the programme name, the grant years, and whether an accelerated rate has been triggered, the exposure cannot be modelled. Ask for those three things.
• Skipping export control on a cyber or deep-tech deal. Agents read the deck's product description and never think to ask whether selling it needs a DECA or dual-use licence. For the modal Israeli deep-tech seed deal this is the single most deal-relevant regulatory question, and unlicensed revenue is worth nothing to an acquirer.
• Treating a signed IP-assignment agreement as closing the IP question. It does not close a service-invention claim under the Patents Law unless there is an express waiver. Read the clause, do not tick the box.
• Confusing founder-side and investor-side framing. This skill evaluates a deal for an investor. Do not slip into advising the founder how to raise or how to set up their 102 plan. That is israeli-startup-toolkit.
• Doing dilution math in your head. Pre-money pool shuffles and stacked post-money SAFEs are counterintuitive. Use the script; do not eyeball the ownership table.
• Over-asking in diligence. Sending 40 generic questions signals a tourist. The output must be a focused list that resolves the top 5 risks for this specific deal.
• Stating Israeli tax/legal figures as advice. Rates and rules (corporate tax, Preferred Enterprise, 102 conditions) change and depend on facts. Present them as screening signals to verify with an Israeli professional, never as binding conclusions.

Reference Links

Troubleshooting

• The investor only gave a one-line idea, not a deck. Produce a thin screening memo from what exists, and make the diligence questions the main deliverable: the questions that would let them decide whether to take a first call.
• The numbers in the deck contradict each other. Do not reconcile them silently. Surface the contradiction as a red flag and a diligence question.
• No IIA / structure information is given. Do not assume there is no exposure. List the IIA, flip, and 102 checks as open diligence items the investor must confirm before a term sheet.
• The investor wants a valuation. This skill sanity-checks a proposed valuation and its dilution math; it does not produce an independent valuation. Frame the output as "is this price reasonable for this stage and traction", not "the company is worth X".