Evogene 2025: ChemPass AI Is Now the Core, but the Thesis Still Runs Through Cash and Dilution

Getting to Know the Company

Evogene at the end of 2025 is no longer a broad platform story with several engines running in parallel. Management has tried to close the narrative around one engine, ChemPass AI, and two target markets, pharma and ag-chemicals. That is a real shift, and a constructive one. The problem is that the numbers still do not support the story with the same force. In 2025, almost all revenue still came from legacy activities, mainly Casterra and AgPlenus, while the human health segment, the part that should benefit most directly from ChemPass AI, generated no revenue at all.

A quick read of the Lavie Bio sale, the Biomica-Lishan agreement, and the expanded Google Cloud collaboration could make it look as if Evogene has already crossed its financing valley. That is incomplete. At year-end 2025 the company had $12.96 million of cash, but operating cash flow was negative $13.5 million, and equity attributable to Evogene shareholders was negative $82 thousand while non-controlling interests stood at $12.06 million. There is cash in the group and there are still assets with value, but common shareholders do not have clean access to that value.

What is actually working now? The company is sharper, the expense base is smaller, and management has turned part of the old portfolio from open-ended cash burners into sold assets or licensed programs. What still blocks a cleaner thesis? ChemPass AI has not yet shown that it can turn scientific positioning and early collaborations into recurring revenue at a scale that can replace the legacy activities being sold, wound down, or narrowed.

This matters now because of the practical market screen as well. Local market capitalization is roughly NIS 22 million, daily trading turnover on the TASE snapshot was only about NIS 13 thousand, and short interest remains low. So this is not a stock where the main question is whether shorts are attacking it. It is a stock where the read is shaped first by financing, liquidity, and the path by which value created inside the group actually reaches the listed common shareholders.

The current economic map looks like this:

That chart captures the tension well. Operating loss has narrowed over three years, but it still sits far above revenue. This is not a company that has already commercialized its core platform and is now polishing margins. It is a company cutting scope, monetizing assets, and buying time until one remaining engine can become more than a strategic story.

Events and Triggers

The portfolio really did shrink

In 2025 Evogene did what small, financing-sensitive platform companies are often pushed to do when the old story becomes too diffuse: it cut. Lavie Bio and the MicroBoost AI agriculture engine were sold to ICL for total consideration of $18.714 million, Lavie stopped being an active operating story, and Biomica moved in practice from internal development to a licensing model. Strategically, this is cleaner. It reduces organizational complexity, lowers burn, and forces management to define what the real business is.

But every such move has a tradeoff. The portfolio is now less diversified. Evogene no longer sits on several independent operating options that might open up in parallel. It is increasingly concentrated around ChemPass AI and a smaller set of value pathways. If that engine does not commercialize, there are fewer other businesses left to hide the problem.

Biomica moved from development story to royalty option

The February 4, 2026 agreement with Lishan matters not because it solves everything, but because it changes the shape of the risk. Biomica granted an exclusive worldwide license for BMC128, a microbiome-based oncology therapeutic, and in return may receive development milestones, sales milestones, and royalties, subject to the agreement terms. At the same time, Biomica was down to only 1.5 full-time employees, and the company explicitly states that no further activity is expected after completion of Phase 1 and execution of the licensing path.

In other words, Biomica is no longer expected to be a meaningful internal investment platform. That helps the cash profile. It is less helpful for anyone who wanted Evogene to keep building deep biotech value inside the group. Biomica has shifted from a burn box to a contingent royalty and milestone option.

Google Cloud sharpens the narrative, but not yet the income statement

The CEO letter, the annual report, and the February 2026 Google Cloud announcement all tell the same story: Evogene wants to be a computational chemistry company centered on ChemPass AI, not a holding structure with several unrelated tech engines. The expansion of the Google Cloud collaboration, now focused on integrating AI agents into ChemPass AI, clearly strengthens that story. It also fits with the line of scientific collaborations signed in late 2025 and early 2026.

The important point is what is missing. There is still no disclosed revenue framework that translates this effort into a clearly visible earnings model. No disclosed payment scale, no disclosure separating scientific proof points from economically meaningful commercial engagements. For now, Google strengthens the narrative more than it strengthens the income statement.

The financing trigger did not disappear, it changed form

On February 10, 2026 Evogene closed an inducement deal in which an existing holder exercised 3.384 million warrants at a reduced $1.00 exercise price, bringing in about $3.385 million of gross proceeds. In exchange, the company issued 5.077 million new warrants at a $1.25 exercise price. This is the core financing event of early 2026. The company bought time, but it paid for that time with future overhang.

This is not a one-directional event. It improves immediate liquidity, which is critical for a company that still discloses conditions raising substantial doubt around going concern. But it also tells the market that even in early 2026 cash was valuable enough to justify a meaningful economic concession on exercise price and a new warrant layer above the stock.

The short event map looks like this:

Efficiency, Profitability and Competition

The central 2025 insight is that the P&L is leaner, but not stronger. Revenue fell 30.4% to $3.853 million, and gross profit moved from a positive $3.197 million to a gross loss of $241 thousand. That was not caused by a broad deterioration everywhere. It was driven mainly by a late bill from the Casterra reset: a $2.18 million inventory impairment after the decision to exit Kenya.

That matters because it exposes revenue quality. Casterra was the group's largest revenue engine in 2025 with $2.168 million of revenue, but it also posted a segment operating loss of $3.54 million. AgPlenus generated $1.374 million of revenue and a segment operating loss of $4.097 million. Human health generated zero revenue and lost $2.653 million at the segment level. So even after the sharper strategic focus, no operating layer is anywhere near an economically balanced model.

The chart makes the hidden point obvious. The future engine has not replaced the legacy engines yet. Human health revenue has disappeared, and agriculture revenue fell after one-time payments and older collaboration streams rolled off. This is not the profile of an AI platform that has already entered a real commercial curve. It is the profile of a company still funding its new story through older activities, not through the new engine itself.

Profit quality is weaker than the headlines

Research and development expense fell 36% to $7.994 million, sales and marketing expense fell 25% to $1.476 million, and general and administrative expense fell 38.6% to $4.286 million. That is a real improvement in cost discipline. But it needs to be framed correctly. This is a cost base improvement created mainly by shutting down activities, cutting headcount, and moving risk outward to buyers or partners. It is not a margin expansion story created by commercial scale.

From a competitive standpoint, that also means Evogene has chosen to compete where its counterparties are much larger. In ag-chemicals, the company itself lists BASF, Bayer, Syngenta, FMC, and Corteva as major players with internal R&D capabilities. In pharma, it is trying to win relevance against biotech companies and research organizations with deeper financial and scientific resources. So the moat, if one develops, cannot come from balance-sheet strength. It has to come from proving that ChemPass AI can produce better candidates, faster, and in a way that creates real economic value for partners.

Revenue concentration is high, and disclosure is only partial

The top three customers generated $3.262 million of 2025 revenue, roughly 85% of the total. The first customer, which is also a shareholder in a subsidiary, generated $350 thousand. The other two largest customers, at $880 thousand and $2.032 million, are not identified by name in the filing. That is not a technical footnote. When a company of this size depends on a narrow customer base, customer identity affects revenue quality, bargaining power, and the risk of sharp volume disappearance.

This also connects to the geographic mix. In 2025, 53% of revenue came from Africa, 26% from Europe, 12% from Israel, and 9% from the United States. But the same year that Africa generated most of the revenue, Casterra had already decided to exit Kenya and refocus on Brazil. So even the current revenue map is not necessarily a clean guide to the next one.

Cash Flow, Debt and Capital Structure

It is important to define the cash frame here. I am using an all-in cash flexibility view, meaning cash left after the year's actual cash uses, not a normalized business-only cash generation view. For a company like Evogene this is the right frame, because the main question is not what would have remained if we ignored monetizations, repayments, or capital transactions. The main question is how much cash was actually left after everything that happened.

On that basis, 2025 looks far less comfortable than the Lavie sale headline suggests. Opening cash was $15.301 million. During the year the company used $13.502 million in operating cash flow, received $17.738 million from investing activity mainly because of the ICL transaction, and used $6.602 million in financing activity, mainly because of the $10 million SAFE repayment, partly offset by $4.283 million of equity issuance. The result was year-end cash of only $12.956 million.

That is the key point to keep in view: the Lavie monetization did produce cash, but it did not turn Evogene into a clean surplus-cash company. The proceeds were largely absorbed by ongoing burn and by closing out a previous financing obligation. So anyone reading the ICL sale as a full solution to the funding question is reading only half the story.

The balance sheet reinforces that. At year-end 2025 the company had $12.956 million of cash and cash equivalents, $12.2 million of working capital, government grant liabilities of about $3.129 million long term plus $56 thousand current, and lease liabilities of $2.198 million in aggregate. This is not a classic heavy-debt problem. It is a high-burn problem sitting on top of a narrow revenue base.

But the most important layer sits between total equity and what actually belongs to Evogene common shareholders. Total equity at year-end was $11.975 million. Of that, $12.057 million was attributed to non-controlling interests, and equity attributable to Evogene shareholders was negative $82 thousand.

This is one of the most important numbers in the whole read. It means group-level cash and value do not automatically flow through to common shareholders of the listed parent. Some of the value gets trapped along the way in minority interests and subsidiary structures. So even if the market tries to screen Evogene as a cash-rich microcap, that can be a dangerously simplified read. Group cash is not the same thing as clean common equity.

The picture improved somewhat after the balance sheet date. In February 2026 the company received about $3.385 million of gross proceeds from the warrant inducement. In March 2026 it received about $2.928 million from a Lavie Bio dividend. In addition, Biomica approved a dividend of up to $2.7 million, still subject to court approval. So the immediate cash pressure eased. But it eased through capital-structure bridges and distributions from assets that are ending life as independent operating businesses, not through internally generated surplus cash from the core platform.

Guidance and What Comes Next

First finding: 2026 will be judged first on whether ChemPass AI becomes a revenue line, not whether it keeps generating announcements.
Second finding: the legacy engines that carried 2025, especially Casterra and AgPlenus, do not look like a clean enough or durable enough base for the next phase.
Third finding: the extra cash raised in early 2026 bought time, but it also expanded the warrant layer over the stock.
Fourth finding: Lavie Bio and Biomica are no longer growth platforms in their own right. They are now mainly sources of cash returns, milestones, and royalties.
Fifth finding: 2026 looks like a bridge year with a very explicit proof test.

What must happen over the next 2 to 4 quarters

First, Evogene needs to show that ChemPass AI has measurable commercial value. Not more collaboration headlines alone, but research payments, licensing payments, or at least a disclosure framework that makes it clear the platform is starting to generate recurring economics. As long as human health stays at zero revenue, the market will struggle to believe that the new focus has already moved from presentation to financial statement.

Second, AgPlenus needs to show that Bayer-related work and its internal pipeline can produce more than early-stage funding. Management itself says 2025 revenue was hurt in part because 2024 included one-time payments and because a Corteva-related revenue stream had already ended. So 2026 needs replacement sources. Without that, AgPlenus remains a technologically interesting business with insufficient economic proof.

Third, Casterra has to prove that the move from African dependence to Brazil is real and not just aspirational. In 2025 it still benefited from delivery of prior ENI orders, but by the date of the annual report there were no additional seed orders. If Brazil does not provide a new volume base, 2025 may look in hindsight like an inventory-clearing year rather than a transition into a better business.

Fourth, the cash returning from subsidiaries needs to keep moving upward. Lavie Bio has already distributed cash, Biomica may do so as well, but that is still only part of the answer. The market will want to see value not just created or monetized, but actually reaching the listed parent instead of remaining stuck inside the structure.

What could improve the market read

Three things could improve the market read over the short to medium term. The first is real commercial proof around ChemPass AI, even if still early. The second is additional cash returning to the parent through a Biomica distribution or similar upstreaming event, because that would strengthen the sense that value is accessible rather than only theoretical. The third is continued migration of Biomica and Lavie from old burn stories into milestone, royalty, or cash-return stories.

There are also three things that could hurt the read. The first is another weakly priced financing step. The second is continued share-price pressure around the $1 level and deeper listing anxiety. The third is disappointment at Casterra or AgPlenus that makes it clear even the legacy engines supporting today's reports are not stable enough to hold the bridge period.

Management is trying to frame 2026 as the year the company already looks different. That is true structurally, but not yet economically. The right interpretation is that 2026 is a bridge year with a proof test. A bridge year because part of the group has already been sold or shifted into licensing. A proof year because the one engine now left at the center still has to show that it can generate revenue, not just attention.

Risks

Financing risk is still the first risk

The company itself says there were conditions that raised substantial doubt about its ability to continue as a going concern for twelve months from the annual report date, even though management believes its plan alleviates that doubt. The numbers are hard to argue with: a $14.034 million operating loss, negative operating cash flow of $13.502 million, and a history of funding through equity, asset sales, and warrant structures. The financing risk has not left the story.

Dilution is no longer a theoretical threat

The February 2026 inducement deal shows that dilution is an active financing tool, not a distant possibility. The company received needed cash, but it also issued 5.077 million new warrants. After already conducting a reverse split in 2024 and still flagging minimum-bid-price and Nasdaq listing risk, any additional financing move could prove very expensive for existing holders.

Customer concentration and partial disclosure

About 85% of revenue came from only three customers, and two of them are not identified by name. That is a difficult combination. Concentration is high, and the partial disclosure makes it harder for the market to judge revenue quality, bargaining power, and the true risk if one customer steps back.

Strategic concentration around one engine

Focusing on ChemPass AI improves clarity, but it also raises the operating leverage of any setback. The filing explicitly warns that the reduction in active subsidiaries and business lines increases concentration risk. That is not boilerplate. If the engine does not commercialize, there are not many other businesses left to mask the result.

FX exposure and listing pressure

The company does not hedge currency exposure, and a 5% move in the U.S. dollar versus the NIS would affect profit or loss by about $301 thousand. In addition, Evogene continues to highlight Nasdaq listing risk. The annual report notes that the Nasdaq share price was $0.94 on March 12, 2026 and makes clear that prolonged weakness in the share price could complicate continued U.S. listing. In a stock this illiquid, that is a practical risk, not just a regulatory footnote.

Conclusions

Evogene is much easier to understand today than it was a year ago. That is the positive side. It sold activity outside the new core, reduced burn, put ChemPass AI at the center, and added cash after the balance sheet date. The less comfortable side is that the economics of the business have not moved quickly enough: revenue still depends on legacy activities, the cleaner bottom line is driven mainly by monetizations and discontinued operations, and common shareholders still do not see a clean equity layer at the parent level.

Current thesis in one line: Evogene is no longer a diffuse platform story, but a one-engine company that has bought itself more time without yet proving that the remaining engine can fund itself.

What changed versus the older read? It used to be easy to see Evogene as a broad option basket with several engines open at once. Today it is a narrower company, which is strategically better but economically more exposed. Focus improved, diversification fell.

The strongest counter-thesis is that the market still does not properly price the fact that the company now has a more focused computational engine, a deepening Google Cloud relationship, an active agriculture arm, a Biomica royalty option, and real cash returning from Lavie Bio. If the new collaborations begin to convert into disclosed economics, a market value of roughly NIS 22 million could look too low.

What could change the market's interpretation over the short to medium term? Not one more innovation headline, but a sequence of three things: more cash returning to the parent, a commercial deal that shows ChemPass AI can generate real revenue, and proof that Casterra or AgPlenus can hold the floor until the new engine matures.

Why does this matter? Because Evogene is at the point where the market has to decide whether it is becoming a commercial discovery platform or simply liquidating older assets to buy a few more quarters of time.

Over the next two to four quarters, what strengthens the thesis is measurable commercialization around ChemPass AI, continued upstream cash returns from subsidiaries, and evidence that the old revenue base can be replaced. What weakens it is more expensive bridge financing, further share-price erosion, or proof that even the legacy engines that carried 2025 are not stable enough to support the transition.