Tenax ($TENX): Why Is the Cash Balance Larger Than the Market Value?

01 Scenarios: prospective evidence and capital allocation

These are conditional frames, not forecasts or trading recommendations. Each needs observable evidence. A historical market discount cannot substitute for that evidence.

Constructive

Regulatory feedback supports a clearly defined enriched population, with criteria and analysis specified before the relevant outcomes are known. LEVEL-2 enrolls on a workable schedule and demonstrates benefit with acceptable tolerability. Spending leaves enough resources for follow-up and submission work. This would turn the subgroup hypothesis into a basis for prospective confirmation; it would not retrospectively make LEVEL’s primary endpoint successful.

The reserve gives Tenax options about development pace. Slower spending is favorable only if it preserves the ability to answer the clinical question. A cheaper plan that undermines enrollment, follow-up or statistical interpretation would not necessarily improve the investment case.

Intermediate

The company publishes further analyses and continues development while the final regulatory package remains unsettled. Cash declines through operations, and investors must distinguish information about the existing dataset from independent confirmation. An unchanged headline runway can coexist with substantial clinical uncertainty.

A delay with ongoing site costs differs from a planned pause with reduced commitments. Neither can be diagnosed from a month of stock performance or an insider purchase. Compare what activity was completed with the cash consumed and the remaining work.

Adverse

The enriched approach fails to reproduce the subgroup benefit, adverse events impair adherence, or regulators require more work than the current reserve funds. Higher costs, longer follow-up or another study can bring financing forward. New equity near a depressed price changes per-share exposure even when the company had no debt at the previous reporting date.

A licensing agreement, merger, asset sale or distribution could change this case, but none is assumed. The June employment arrangements specify contingent benefits, not an announced strategic process. Any later transaction needs its own analysis of cash consideration, retained liabilities, rights and dilution.

Evidence that could change the assessment

The assessment combines an unsuccessful confirmatory trial, a testable hypothesis and a substantial but dated reserve. Evidence can improve or weaken each component independently; silence over an arbitrary interval is not automatically failure.

• Regulatory clarity: a disclosed meeting outcome and precise protocol could establish a feasible route; additional studies could extend costs and timing.
• Prospective efficacy: a new randomized result can test reproducibility. More analyses of the same LEVEL patients are not independent confirmation.
• Benefit and tolerability: exercise capacity, symptoms, discontinuations and serious outcomes must be judged together. Low baseline walking distance alone does not establish worse drug tolerability.
• Cash linked to activity: persistent consumption beyond the stated plan, after working-capital and one-off effects, could challenge funding assumptions. A declining balance alone is expected during development.
• Capital: updated warrants, financing terms and strategic agreements can change economic ownership. Legal shares, equivalents and beneficial ownership have different denominators.

A successful meeting cannot substitute for efficacy, positive biology cannot substitute for funding, and a balance-sheet surplus is not a promised shareholder payment. Tracking these distinctions allows the assessment to change as evidence arrives.

02 LEVEL: endpoints and tolerability

LEVEL randomized 241 participants at 41 sites:120 to drug and 121 to placebo. Its prespecified repeated-measures model gives an adjusted 3.5m difference, distinct from the 8.0m unadjusted observed difference. Observed week 12 tests were available for 116 and 120 patients. Rounded individual means do not reproduce the precise model contrast. The post-hoc quartile gradient motivates a hypothesis, not proof of causal interaction. Source

Serious-event totals and tolerability

Serious events were 10.8% versus 10.7% and adjudicated worsening 2.5% in each group. Balanced totals do not establish absence of risk: the deck lists two deaths on drug and none on placebo in the worsening breakdown, without establishing causality from small counts. Treatment-related events, discontinuations and reductions were more frequent on drug. Source

• Any adverse event: 86.7 per cent against 71.9
• Treatment-related adverse event: 38.3 per cent against 17.4
• Discontinuation for an adverse event: 8.3 per cent against 1.7
• Dose reduction: 15.0 per cent against 4.1

Headache, palpitations and hypotension contributed to tolerability differences. September 9 reports treatment-emergent events in 87.1% of drug patients below median walking distance and 86.2% at or above it; placebo rates 78.9% and 65.6%. These small comparisons do not prove equivalence but do not support assuming that the lower-capacity group necessarily tolerates treatment worse. Future evaluation needs actual subgroup discontinuation and safety evidence. Source

What the earlier trial did and did not establish

HELP used intravenous levosimendan and an open-label responder stage before randomization. The primary abstract reports 37 randomized responders and 29.3m placebo-adjusted walk improvement at six weeks; exercise pulmonary-capillary-wedge-pressure primary endpoint was not significantly reduced. Route, selection, sample size and duration differ from oral LEVEL. The trials cannot be pooled casually as the same treatment estimate. Source

Tenax has deprioritized significant imatinib development activity. Oral levosimendan is its prioritized clinical program; this does not mean no other asset rights exist. Source

Tolerability in absolute numbers

The disclosed adverse-event counts help make the percentages concrete. Treatment-related events affected 46 of 120 participants on drug and 21 of 121 on placebo. Treatment discontinuation due to an adverse event affected ten versus two, and dose reduction eighteen versus five. These are absolute differences of 25, eight and thirteen participants respectively, against groups of almost equal size. They are clinically relevant observations even though the company describes the overall safety profile as supporting development. They do not identify which particular future patient will tolerate therapy or establish that every event was caused by treatment.

Balanced serious-event totals of thirteen patients in each arm must also be interpreted with the individual event categories and follow-up duration. An absence of statistical separation on a small safety comparison would not prove that the risks are identical. Conversely, a numerical imbalance in an individual category does not by itself demonstrate causality. A future enriched study can clarify the combined effect of selection and longer exposure, but its safety findings should be reported with actual denominators, treatment duration and discontinuation patterns. Preserving those details is more informative than choosing either an unqualified safe label or an unqualified dangerous label. Source

What would count as replication?

The subgroup comprised 119 of 241 randomized patients, approximately 49.4% of the total. Its 26.3-metre estimate is larger than the overall 3.5-metre contrast, but those estimates come from overlapping patients. They are not two independent experiments. The favorable subgroup confidence interval, 6.0 to 46.7 metres, also leaves considerable uncertainty about the magnitude. A second study could produce a smaller effect without necessarily proving no activity; the interpretation would depend on its prespecified endpoint, uncertainty and clinical relevance. Conversely, reproducing a favorable biomarker change without reproducing functional or symptomatic benefit would answer only part of the development question.

The analysis hierarchy matters because examining multiple populations and outcomes creates more opportunities for favorable findings. Prespecification reduces the flexibility to choose groups after seeing results, but it does not automatically assign confirmatory statistical error control to every subgroup. Tenax explicitly labels the p-values nominal. The primary failure and the exploratory findings therefore belong side by side. A reader should ask what hypothesis the next protocol will test, how success is defined and whether that definition is fixed before outcomes are available. Reusing the same dataset to choose a threshold and then presenting its resulting effect as a validated prediction would overstate the evidence.

Endpoint measurement also matters. Walking performance can be affected by multiple aspects of health and testing conditions. A more selective population may increase the ability to detect a drug effect, but can also narrow the population to which results apply. Neither a biomarker response nor an age gradient establishes the best eligibility rule on its own. Prospective eligibility, adherence, background treatment and missing-outcome handling should be evaluated together. These are analytical questions raised by the disclosed data, not a claim that the company has already adopted a particular answer. Source Source

03 Regulatory strategy and licence

The strategy is described by Tenax, not a published agency decision or accepted application. No later SEC or issuer announcement confirming a Type C outcome was found through October 4.

What the company says it will do

August 10 describes intended FDA discussion and European consultation, including enrichment. No confirmed meeting date or endorsement was identified. Private interactions may precede public filings, so no public date is not proof that no request was submitted. Source

Tenax reports prior FDA agreement that a single Phase 3 at p=0.01 could support an NDA. This is a company account of a conditional filing route, not guaranteed approval. LEVEL failed that endpoint condition. Future evidence requirements need clarification; every regulatory discussion need not necessarily restart from zero.

LEVEL-2, by its registry record

Patent protection by formulation and territory

The October presentation lists U.S. patent 12,616,694 for subcutaneous use, issued May 5, 2026, with an October 15, 2041 expiry, and European patent EP 4076398 B1, issued February 18, 2026, with a December 15, 2040 expiry. These are company-reported patent dates, not a blanket 2041 exclusivity claim for oral TNX-103. October presentation, SEC Exhibit 99.1.

The licence, and where the rights actually came from

The licence dates to 2013. February 2024 expanded oral/subcutaneous PH-HFpEF rights, raised the FDA milestone to $10M, added $5M Japan approval and up to $45M commercialization milestones plus royalties. September 2025 broadened worldwide oral and manufacturing rights. Contingent payments were not June liabilities because triggers had not occurred. Source

June 29 extends to December 31, 2035 the U.S.-approval deadline after which either party can exercise a termination right; it also adds information/cybersecurity requirements. A separate supply agreement governs manufacturing, forecasting, quality, pricing and certain alternative-manufacturing/scale-up provisions. This deadline guarantees neither funding nor approval. Source

Commercial rights are distinct from commercial readiness

The Orion arrangements support development and potential commercialization, but possessing rights does not establish an approvable product or an economical launch. The June filing describes a supply agreement with an initial five-year term from first delivery and automatic three-year renewals unless the required non-renewal notice is given. It also describes forecasting, ordering and scale-up provisions. Those obligations connect scientific progress with operational commitments; they should be reconsidered when the expected program schedule changes.

The $10 million FDA-approval milestone is contingent and therefore different from current accounts payable. Ignoring it in a future successful-launch scenario would understate cash requirements, while subtracting it from June cash as though already overdue would misstate the historical balance sheet. The same distinction applies to the Japan and commercialization milestones. No royalty rate or manufacturing margin can be modeled from partially disclosed contract pricing. A credible future commercial model would need the final indication, eligible population, price, reimbursement, distribution requirements and disclosed contractual economics, rather than extrapolating directly from a subgroup effect in a failed primary trial. Source

Reading an amended trial as a new decision framework

The registry describes a 540-participant study with a 2:1 allocation and a 26-week primary assessment. That implies approximately 360 participants assigned to active treatment and 180 to placebo if the planned allocation and enrollment are achieved. These are design calculations, not actual enrollment counts. A larger total sample does not by itself guarantee that an enriched subgroup analysis will have adequate power: the effect size, variability, missing data and final population definition all contribute. Nor should the 12-week LEVEL effect be assumed to scale proportionally to 26 weeks. Longer treatment may change adherence, adverse-event exposure and the measured response.

There are practical questions when a protocol is revised after enrollment has begun. Which participants qualify under the new rule? Will earlier participants remain in the primary analysis, a separate cohort or another prespecified analysis? Will a new eligibility threshold affect the rate at which sites can recruit? The public registry summary does not answer every operational question, and Merlintrader has found no public disclosure of the missing protocol details. A future company explanation can reduce this uncertainty even before clinical results are available. However, an explanation must be separated from confirmation that the agency regards the resulting package as sufficient for a submission.

The six-month primary assessment, longer open-label observation and subsequent regulatory preparation create different resource needs. Funding only through an estimated primary-completion month is not necessarily funding through an eventual application, review or launch. The licence preserves an avenue for development, but regulatory deadlines in the contract do not supply the people, trial capacity or cash needed to reach those milestones. A useful update should connect the scientific design, operational schedule and financial plan rather than announcing each in isolation. Source

The October 8, 2026 Form 8-K moves the strategy from intention to action. Tenax says it is “enriching” LEVEL-2, has temporarily paused enrollment while the criteria are amended, and has requested a Type C meeting with the FDA on the overall development of levosimendan. The investor presentation filed as Exhibit 99.1 proposes a new entry rule, a six-minute walk distance of 100–300 m, or 300–400 m with a right atrial volume index of at least 25 mL/m², and reports that LEVEL patients meeting it showed a difference of more than 34 meters versus placebo at week 12 (95% CI +15 to +53 meters; nominal p = 0.001); more than half of LEVEL participants met the proposed criteria. That is a post-hoc subgroup with a nominal p-value, not a confirmed result: LEVEL missed its primary endpoint in the overall population. The company estimates that fewer than 5% of patients in the LEVEL-2 primary efficacy analysis will have been enrolled without meeting the stricter criteria, and states that the pause is not related to safety and was not requested by a regulatory authority. A meeting request is not a meeting outcome; agency feedback has not been disclosed. Form 8-K · Exhibit 99.1

04 Cash flows and dated valuation

H1 operating use $23.459M less Q1 $9.309M gives Q2 $14.150M, or $4.717M monthly versus H1 $3.910M. Q2 R&D rose to $12.850M from $6.121M; SG&A $5.934M and loss $17.803M. Accruals, working capital and stock compensation separate loss from cash. Trial activity explains part of costs; payment timing shifts cash between quarters. Source

Current assets $121.663M less liabilities $8.073M give working capital $113.590M. Cash less recorded liabilities was $109.903M. Neither is guaranteed distributable cash: prepaid research is not a bank balance, while trials and commitments continue. Management’s at-least-twelve-month funding assessment was dated, not assurance against future financing.

Where the money came from, and what it is worth per share

August 2024 and March 2025 placements raised about $125M gross combined. March 2025 brought $25M gross/$23.2M net through 378,346 common and 3,760,726 pre-funded warrants. H1 financing $43.870M exceeded operating use, increasing cash $20.411M. The $300M shelf is registration capacity, not committed cash; any proceeds depend on terms, access and limits. Source

Dated calculation:$126.076M before burn divided by 37,423,917 July 28 shares is $3.37. Adding 16,190,603 pre-funded equivalents gives 53,614,520 and $2.35 each. September 30 price $1.89 implies $70.731M basic or $101.331M equivalent capitalization. These are alternative conventions, not simultaneous market values. The figures carry different dates; neither is current liquidation value.

Cash sensitivities with explicit assumptions

A simple sensitivity calculation illustrates why the spending rate deserves attention. Starting with the $126.076 million arithmetic subtotal before subsequent expenditure, division by the historical Q2 monthly operating use of $4.7167 million gives approximately 26.7 months. Using the lower H1 monthly average of $3.9098 million instead gives approximately 32.2 months. Neither is a forecast or a replacement for management guidance: the starting subtotal mixes a quarter-end balance with later proceeds, and future spending, interest and funding can differ. The six-month spread between the two results shows how strongly a seemingly precise runway number depends on the selected historical period.

At a hypothetical spending rate 25% above Q2, approximately $5.8958 million monthly, the same subtotal covers about 21.4 months. At 25% below Q2, approximately $3.5375 million monthly, it covers about 35.6 months. These examples deliberately hold other cash flows at zero and do not assume that lower expenditure preserves the same clinical schedule. A delayed trial with persistent fixed costs can consume more cash per completed milestone even when monthly spending falls. Conversely, a temporary increase can reflect payments for work that advances the program. The relevant question is the remaining cost of the evidence package, not merely which historical average produces the longest quotient.

Six months at the Q2 rate would consume $28.300 million, about 22.4% of the $126.076 million subtotal. Twelve months would consume $56.600 million, about 44.9%. A financing assessment should therefore consider when additional capital might need to be arranged, not only the theoretical date when a balance reaches zero. The company may choose to raise before that point, or change its program. These are possible decisions, not statements that a financing is imminent. The historical no-borrowing balance sheet gives flexibility while leaving future capital requirements unresolved. Source

Separate operating results from cash conversion

The H1 cash-flow bridge is internally consistent: $97.565 million beginning cash, less $23.459 million operating use, plus $43.870 million financing inflows, produces $117.976 million ending cash. Operating cash use is smaller than the $33.519 million loss. The reconciliation includes $6.168 million stock compensation, a $2.975 million benefit from prepaid expenses and other current assets, and increases of $0.754 million in payables and $0.163 million in accrued liabilities. These explain the $10.060 million difference between loss and operating consumption; they do not create recurring product revenue.

This matters for the next quarter. A reduction in prepayments can make cash consumption look better while previously paid services are being expensed. An increase in payables can defer cash use into another period. Neither is necessarily problematic, but both affect comparison with a simple monthly model. Review the expense trend alongside the reconciliation and the activity described by management. A single quarterly cash figure above the illustration is not enough to conclude that development became cheaper, just as a figure below it does not automatically establish financial deterioration. The direction and persistence of the underlying changes matter. Source

05 Warrants and share denominators

June instruments are a dated starting point. Subsequent exercises and expiry alter claims and potential proceeds. Observed transactions, terms-based conclusions and the yet-unreported quarter-end reconciliation are distinct.

The aggregate table reports 7,508,153 at average $6.91; two tranches total 7,491,629 at $4.50/$5.65, blended around $4.53. The 16,524-share difference and aggregate strike are unexplained by those descriptions. Report both rather than silently correcting the filing or assigning the difference to a particular class. Source

The share count, and why the basic number understates it

H1 common growth 22,635,655 comprises 13,241,197 pre-funded exercises, 9,386,958 ordinary warrants, 5,000 options and 2,500 RSUs. Ordinary warrants contributed materially. Pre-funded conversion largely changes legal form of paid-for exposure; ordinary exercise also brings meaningful cash. Both require reconciliation before judging dilution or financing efficiency.

Authorization is 400M common/10M preferred, not guaranteed access to capital or exemption from requirements. June 210 Series A preferred convert in aggregate into one common share.

Per-share exposure and hypothetical future funding

The 16,190,603 remaining pre-funded equivalents are about 43.3% of the July 28 common count. Adding them increases the denominator from about 37.4 million to 53.6 million, so the cash-per-equivalent comparison is materially lower. Exercise at one cent would bring only about $161,906 before any cashless exercise mechanics, far less than a new financing of comparable share count. This distinction prevents treating conversion of already-paid instruments as a fresh source of the development budget. Beneficial-ownership blockers can affect the timing of conversion without making the underlying economic exposure disappear.

For scale only, a hypothetical $20 million gross common-stock financing at the historical $1.89 reference price would issue approximately 10.58 million shares before fees or any accompanying instruments. Relative to the 53.61 million common-equivalent base, that would increase the denominator about 19.7%; existing holders would own about 83.5% of the enlarged total if they did not participate. This is arithmetic, not an announced offering, forecast price or view that such a transaction should occur. Discounts, warrants, fees and the actual share price could change the result substantially.

The calculation also illustrates why a shelf is not guaranteed financing. Registration capacity permits certain securities to be offered subject to its terms and applicable requirements; it says little by itself about investor demand or the cost of capital. A prospective transaction must be assessed on net proceeds, additional claims and the development work it funds. Raising cash can extend operating capacity while reducing each existing share’s participation in any successful outcome. Those two effects can occur at the same time. Source Source

06 Ownership and governance

The session, and the tape around it

Nasdaq confirms September 30 close $1.89 on 4,223,443 shares, and August 10 $1.375 after August 7 $13.44. This documents repricing, not that shares stopped reacting or must stay in a narrow band. October 2 closed $1.98; the comparison uses the September 30 close.

Short interest, with its settlement date

Short interest fell 40.7%, from 9,276,905 August 31 to 5,503,842 September 15. The latter is 14.71% of July 28 common shares, not necessarily public float. Days-to-cover 1.03 uses average volume 5,350,727. Turnover and competing orders mean this ratio does not guarantee covering in a day without moving price.

Dated ownership and transactions

Ownership reports are snapshots. Transactions may be more recent than quarterly screens but establish no later holdings. Aggregate percentages depend on scope and denominator; the figures come from identified SEC filings rather than a purported live register.

FMR 514,925/1.4% August 31 compares with 3,942,373 June 30, down 86.9%. The interval straddles August 10; snapshots do not date every sale after results. ING’s September 11 amendment reports zero September 3. These establish exposure changes, not motives or institutional consensus. Source Source

Giordano’s July 31 13D reports 1,690,606/4.32%, comprising 11,324 common and 1,679,282 underlying options. Beneficial-ownership rules add relevant options to the denominator. It is not 1.69M purchased common. His later Form 4 common line is 22,174 after September purchases; derivatives remain separate. Source

Four purchases involve three insiders: Giordano 6,000@$1.875 September 11 and 4,850@$1.8221 September 15; Staab 5,675@$1.935 September 11; Doogan 53,000@$1.92 September 16. Code P shows added exposure, not knowledge of a favorable undisclosed outcome. Source Source Source

Analyst and listing context

No current analyst consensus is adopted. A stale or mixed feed does not prove no analyst revised after August 10. A useful note identifies firm, date, development assumptions and capital structure; unexplained old targets establish no upside.

Price above $1 and a quote page do not establish all Nasdaq compliance. No new deficiency announcement was found in SEC submissions through October 4, a bounded check rather than exchange certification.

Management, and the structure put in place before the readout

Giordano is CEO, Staab became CFO May 11, 2026, and Rich is CMO. January 9 records Rich’s four-day average week from January 1 with prorated salary. Staab’s September Form 4 has 450,000 options at $11.95, not the inducement-plan weighted average. Employment schedules and out-of-the-money awards do not establish trial views. Source

June 26 amended executive agreements and separate employee severance/change-in-control plans were reported June 29. Executives have individual agreements; benefits require specified termination and release. Giordano receives 18 versus 12 months salary/medical benefits. These obligations affect economics but establish no contemplated takeover or distribution. Source

Compare ownership measures without double counting

The CEO example shows why apparent differences between forms need not be contradictions. A beneficial-ownership report can include shares underlying exercisable options under the applicable rules, whereas the non-derivative table in a Form 4 reports common stock separately. Adding the large beneficial total to the common line would count some exposure twice. It would also confuse options requiring an exercise payment with shares already held. The July 31 filing explicitly states its composition, making the distinction verifiable rather than speculative.

Similarly, ADAR1 and its controlling person report shared exposure to securities held through managed funds. Their reported totals are not independent positions that can be added together. Ownership limitations on particular warrants explain why those underlying shares can be excluded from a reported percentage even while the holder has economic exposure to them. When assessing a later filing, first compare event dates and instrument categories, then examine transactions. A changed percentage alone can result from a changed denominator, and a large purchase is not proof of a particular scientific or strategic thesis. Source Source

07 Development timetable and practical monitoring

Different events answer different questions. An estimated trial-completion date is not a fixed release, while an intended agency meeting is not already scheduled. Undated developments can still be material.

Next financial report

The Q3 report is normally expected in November; Merlintrader has found no public announcement of a specific earnings date. It should update cash, expenditure, share counts and subsequent events. A revised runway is useful but not guaranteed simply because a quarterly report is due. Compare cash flows with working capital and financing before interpreting the change.

Regulatory and protocol milestones

Watch for a disclosed Type C outcome, European advice and precise LEVEL-2 criteria. Important details include treatment of already-enrolled patients, analysis populations, sample size, power, multiplicity and missing data. A design paper documents a plan, not necessarily agency endorsement.

Scientific communications before 2028

The September 9 deck anticipates a primary paper, additional publications and conferences during 2026–H1 2027, including HFSA, AHA, CVCT and THT. These are issuer plans without a specific new-data date asserted here. Some may revisit LEVEL. Determine whether each communication adds follow-up, an independent analysis or another presentation of known observations.

Longer-term controlled evidence

The August 19 registry estimates LEVEL-2 primary completion June 30, 2028 and overall completion June 30, 2029. Original LEVEL follow-up is estimated October 30, 2028. Delay can increase cumulative costs; earlier completion still needs adequate follow-up and interpretable analysis. Speed is not evidence of success.

Capital decisions

Warrants reached their September 22 expiry, so June potential proceeds no longer carry forward. Keep pre-funded equivalents, ordinary warrants, options and issued common distinct. Financing or strategic agreements require analysis of cash, fees, restrictions and consideration; an old shelf or severance plan is not evidence that a transaction was predicted.

The reserve provides capacity to test a revised hypothesis, while the original confirmatory failure remains part of the scientific record. Funding, prospective design and acceptable benefit-risk evidence are all needed. A negative historical enterprise-value calculation removes none of these uncertainties.