Unicorn's NAV Quality: How Much of the Portfolio Rests on Cost, How Much on Models, and How Much on Market Price
This follow-up isolates NAV quality rather than the discount itself. Roughly two thirds of Unicorn's portfolio rests on model-based Level 3 marks, another third on cost-like marks, and almost none of it on market price.
After the main article established the discount, the more important question is what actually sits inside Unicorn's NAV. Not every one of the NIS 30.566 million portfolio carries the same quality. Year-end 2025 shows almost no assets with an observable market price, a large block resting on valuation work, and another block resting on the conclusion that cost is still the best estimate of fair value.
Four non-obvious points matter here:
- Almost the entire NAV is Level 3. Note 13 shows financial assets at fair value of NIS 30.566 million as of December 31, 2025, essentially all of them in Level 3, while the Level 1 layer is below NIS 1 thousand.
- Two thirds of the book rests on models, not on market price. Wipflash, Flow-Lit, Strix, and LeO add up to NIS 20.238 million, about 66.2% of the portfolio, and each is measured through OPM, multiples, DCF, a statistical model, or a combination of those tools.
- Almost all of the positive gap above cost sits in Strix. The Strix holding is carried at NIS 13.286 million versus cost of NIS 6.783 million, a NIS 6.503 million uplift. Flow-Lit adds another NIS 609 thousand above cost. Beyond that, the rest of the book is around cost or below it.
- Cost is not the same as a market anchor. Note 5 says that for dollar-denominated holdings measured on a cost basis, cost in dollar terms is retranslated at each reporting date using the period-end exchange rate. So even the "cost" layer can move in shekel terms without any external price discovery event.
The Real NAV Quality Map
The right way to read the portfolio is not only by company name, but by the quality of the valuation anchor behind each mark.
| Valuation layer | Main holdings | Fair value at year-end 2025 | Share of portfolio | What is really behind the number |
|---|---|---|---|---|
| External models or valuation work | Strix, Flow-Lit, LeO, Wipflash | NIS 20.238 million | 66.2% | OPM, DCF, multiples, and statistical models |
| Cost as valuation anchor | Supersi, Bioplasmar, VibeZ, San Brands, ActualSignal, Engini, Inteil | NIS 9.590 million | 31.4% | The partnership says no material indicator justified moving away from cost |
| Management estimates under stress | Shiratech, Screenz, Bobile | NIS 738 thousand | 2.4% | Slowdown, liquidation, or residual-value assumptions |
| Market price | Cyberwan | Below NIS 1 thousand | Negligible | Active quoted market price |
This matters because it separates NAV that rests on an observable price from NAV that rests on judgment. Unicorn has almost no market layer. Even within Level 3, the book is not one homogeneous block: some holdings are marked through valuation models that attempt to estimate business value, while others simply remain at cost because there has not yet been enough evidence to move away from it.
Where The Gap Above Cost Really Sits
The most important picture is not just the mix of the portfolio, but the gap between cost and fair value. Portfolio cost at December 31, 2025 stands at NIS 41.611 million, versus fair value of NIS 30.566 million. In other words, the book as a whole sits at 73.5% of cost. But that aggregate number hides a sharper fact: almost all of the positive gap above cost is concentrated in Strix.
Strix alone accounts for 43.5% of the portfolio, and Flow-Lit adds another 11.9%. Together they already make up 55.4% of NAV. Add LeO and Supersi, and four holdings reach 73.2% of the book. The implication is straightforward: the debate about NAV quality is not spread across 15 tiny names. It is concentrated in a small number of positions, led by Strix.
This is also where the "model" layer needs to be read carefully. A model can capture business value that is not visible in historical cost. But in Unicorn's case it also concentrates most of the positive valuation gap in one asset. Outside Strix and Flow-Lit, the book does not show a broad set of holdings carried above cost. LeO and Wipflash, for example, sit in the modeled bucket yet are still marked below cost.
ActualSignal: Another Check On The Same Round, Not A New Market Anchor
The December 21, 2025 update matters mainly because of what it does not prove. The report says the partnership exercised warrants to purchase ActualSignal seed preferred shares for USD 106 thousand. Note 5 adds that the right was granted as part of the original December 2024 investment, on the same round terms, for a 12-month period. It also makes clear that the exercise took place in December 2025, that the cash was paid only in January 2026, and that by December 31, 2025 the commitment was already recorded inside the financial asset and against payables.
That is an important point for NAV quality. The ActualSignal warrant exercise is not a new market observation. It is not a fresh outside round at a new price, not an active market quote, and not an independent valuation. It is mainly a continuation of the same transaction price embedded in the original agreement.
That is why the year-end carrying value, NIS 1.296 million, still belongs in the cost-anchor layer. The accounting cost of the position stands at NIS 1.442 million, and fair value is NIS 146 thousand below that. Put differently, another USD 106 thousand went in, but valuation quality did not really step up. If anything, the update underlines how even a company with a current report can remain inside the same internal Level 3 layer.
What An Almost Entirely Level 3 Portfolio Means In Practice
Note 13 sharpens the point further: at year-end 2025 the book is almost entirely Level 3, and year-end 2024 looked almost the same. That is not a technical footnote. It means the core question at Unicorn is not only how far NAV sits above the market cap, but how much of that NAV rests on anchors the reader can actually test against an outside market.
In practice there are three very different layers of confidence here:
- The model layer: it can create value above cost, but it is sensitive to assumptions. That is especially true in Strix, which is carried off a DCF built on two scenarios, 50% weighting for each scenario, and a 24% discount rate.
- The cost layer: it may look conservative, but it is not the same as market price. It only says there was no identified reason to move away from cost. In dollar names, FX moves and follow-on checks on the same terms can still shift it.
- The stress or liquidation layer: it is small in the book today, but it shows what happens once a holding is no longer leaning on a round price or a growth model and instead leans on management's residual-value estimate.
That leads to an uncomfortable but clear conclusion. Unicorn's NAV quality is not a simple question of management being aggressive or conservative across the board. It is a composition question. Roughly two thirds of the book rest on models, another third rests on cost as a substitute for price, and almost none of it rests on market. Inside that mix, most of the value, and even more of the upside above cost, is concentrated in Strix.
Bottom Line
This follow-up does not contradict the main article. It sharpens it. Unicorn does not trade merely at a discount to NAV. It trades at a discount to an NAV whose quality is split between very little market price, a large model layer, a meaningful cost layer, and high concentration.
That also explains why the next important trigger is not simply another report about investing more money on the same terms. What can really improve NAV quality is either a new outside price point in the modeled holdings, or a fresh external round that moves the cost layer forward, or an event that shortens the distance between fair value and accessible value. Until that happens, Unicorn's NAV remains a legitimate accounting number, but not one where every shekel deserves the same weight.
Disclosure: Deep TASE analyses are general informational, research, and commentary content only. They do not constitute investment advice, investment marketing, a recommendation, or an offer to buy, sell, or hold any security, and are not tailored to any reader's personal circumstances.
The author, site owner, or related parties may hold, buy, sell, or otherwise trade securities or financial instruments related to the companies discussed, before or after publication, without prior notice and without any obligation to update the analysis. Publication of an analysis should not be read as a statement that any position does or does not exist.
The analysis may contain errors, omissions, or information that changes after publication. Readers should review official filings and primary sources before making decisions.