Decentralised News · 2027 planning series · Reviewed 21 November 2026
What Matters
Buying shares in an AI company, buying an agent-related token and paying for an AI tool are three different capital decisions. Equity starts with company ownership and the rights attached to the security. A token starts with its specific economic and technical terms. A software purchase starts with whether useful work exceeds the full cost of obtaining it. The AI label cannot make those rights, cash flows or measurement methods interchangeable.
The Signal: Evidence reviewed
Reviewed: 21 November 2026. The source base includes SEC investor education on stock ownership and costs, the ERC-20 interface and an example of workflow-software pricing. DN contributes a rights-first comparison and two transparent scenario ledgers. All numerical examples are fictional. No company, token or subscription is ranked, recommended or assigned a future return.
DN Alpha Thesis: choose the claim before the ticker
A successful AI business can produce several kinds of benefit: company profit, developer revenue, lower operating costs for customers and activity on a network. The buyer receives only the benefit connected to the thing they actually purchase. A productive product can coexist with an unattractive investment price. An active token can coexist with weak holder rights. A useful subscription can be valuable to a customer who has no interest in owning its provider.
DN’s decision rule is to write the return mechanism in one sentence before comparing options. “I own this class of shares.” “This token grants these documented rights.” “This subscription helps me complete this workflow at this total cost.” If the sentence remains vague, a spreadsheet will add precision to an unresolved premise.
Start with a comparison of rights
| Purchase | First evidence to inspect | Benefit to measure | Question that can invalidate the thesis |
|---|---|---|---|
| Equity | Issuer, security class, ownership rights, disclosures and purchase terms. | Changes in the investment’s realizable value and distributions actually received. | Does the security give the exposure and rights you think it does, at a price you can justify? |
| Agent-related token | Exact chain and contract, holder mechanism, supply, control and applicable terms. | Realizable token value and the specific benefits or payments the holder can actually claim. | Does product usage create a documented benefit for this token holder? |
| Software or agent service | Service scope, limits, fees, cancellation, data handling, support and accepted-output requirements. | Realized business benefit after implementation, usage, review and operating costs. | Can the user repeatedly obtain valuable accepted work, rather than merely generate activity? |
This is a commercial decision framework, not an accounting classification or a legal taxonomy. A particular product may combine several rights, and a token may represent a security or another regulated arrangement. The actual instrument and applicable law need review. The table should make that investigation more precise, not replace it.
Equity: inspect the company and the exact security
Investor.gov describes stock as an ownership position in a corporation; many stocks also provide voting rights. The practical research task is to identify the issuer and share class, then read the rights and disclosures that apply to the security being offered. A brand name alone is insufficient when a purchase uses an intermediary, fund, derivative or contractual wrapper.
For an AI business, DN proposes separating product evidence from investment evidence. Product evidence concerns whether customers can use the service and obtain accepted outcomes. Investment evidence concerns the company’s finances, financing needs, ownership structure, valuation and the rights of the particular investor. Strong evidence in the first category does not complete the second.
The same distinction applies to a business that uses several AI providers. Revenue generated by one agent workflow may flow to the operator, a software vendor and an infrastructure provider. Owning shares in one participant gives exposure to that entity, not automatic ownership of the entire workflow. Map the revenue recipient before attributing a growing market to a particular security.
Tokens: connect usage to a holder mechanism
The ERC-20 interface describes functions and events for fungible tokens. It does not create one universal entitlement to a company’s profit or an agent’s revenue. The research must continue beyond the interface: what does holding this particular token permit, and what evidence shows how that benefit reaches the holder?
Distinguish access, governance, service credits, distribution rights, redemption terms and a discretionary buyback policy. These may be economically meaningful, but they do not share one payout formula. A token required to pay for a service creates a different relationship from a token with a documented distribution mechanism. A governance vote is also different from an enforceable right to receive a payment.
For each claimed benefit, record its source, beneficiary, conditions, fees and delivery evidence. If a project describes “revenue,” ask whose revenue it is. A customer payment to an operating company is not automatically a transfer to token holders. A buyback can be documented without creating an individual holder’s cash receipt. DN’s payout-rights reconciliation develops that distinction in detail.
Then examine what can change the mechanism. Supply increases, eligibility conditions, administrator powers and amendments can alter the relationship between usage and holder benefit. A price chart will not answer those questions. The AI-agent token evidence checklist is a useful companion when reviewing the contract, product and market record.
Software: buy an operating result you can observe
A software subscription is usually evaluated by the work it enables under its service terms. The customer should define one accepted outcome before buying capacity. An accepted outcome might be a correctly reconciled invoice, a usable research brief or an accurately prepared support response. “One agent run” is an activity count; it is not automatically a business result.
Pricing units can also differ from the unit of value. For example, n8n’s published plans describe workflow executions and plan-specific capabilities. This is a reminder to read the actual billing unit and limits, not a recommendation of a particular plan. A workflow may additionally require model, data, hosting or human-review costs depending on its design.
DN suggests a pilot ledger with baseline volume, accepted output, review time, rework, recurring charges and actual cash effects. Keep additional revenue separate from reduced costs so that the same improvement is not counted twice. If a task takes fewer staff hours but payroll and revenue remain unchanged, report capacity released rather than an immediate cash saving.
A tool can still be worth buying for quality, responsiveness or capacity. Those are legitimate goals when stated openly. The mistake is relabeling every hour saved as money already earned, then comparing that modeled figure with an investment’s cash return.
There are two useful ledgers, not one universal score
The investment ledger tracks acquisition outlay, market value, distributions and exit costs for the instrument being modeled. The operating ledger tracks implementation outlay, recurring costs and realized business benefit for the workflow. Both help make assumptions visible. Neither creates a fair universal ranking without a common horizon, an appropriate treatment of risk and evidence supporting the input values.
SEC investor guidance explains that transaction and ongoing expenses reduce investment returns. DN applies the same discipline to the scenario exercise: the quoted change in an asset’s price is not the investor’s net result after costs. For a software workflow, the vendor’s subscription price is likewise only one part of the operating case.
Do not force both ledgers to produce the same label. A quoted asset value can remain unrealized. An estimated productivity benefit can remain unrealized in cash. A distribution actually received and a customer payment actually collected provide different evidence from an assumption entered in a calculator.
Worked investment scenario: a 20% price rise is not a 20% net return
Fictional arithmetic: an investor allocates $1,000, pays $10 in entry costs and therefore acquires $990 of an asset. Assume its price subsequently rises by 20%, there are no distributions and exit costs are $10. The modeled asset value becomes $1,188; net sale proceeds are $1,178; and the modeled gain against the original $1,000 outlay is $178, or 17.8%.
If the assumed price instead falls by 20%, asset value becomes $792. After the same assumed exit cost, proceeds are $782 and the modeled loss is $218, or 21.8%. These are arithmetic illustrations. They do not describe the expected behavior of an equity, token or particular market.
The formula can be applied to either a share or a token for a simple unleveraged scenario, but identical arithmetic does not make their rights or risks identical. The researcher must still examine the instrument, custody, liquidity, disclosures and applicable terms. Variable fees, market impact, taxes, currency conversion and financing can change the realized result.
Worked operating scenario: capacity only becomes cash when something changes
Fictional business case: a workflow requires $1,000 to implement, costs $150 per month to operate and produces $300 per month of verified incremental cash benefit. Monthly net contribution is $150. Over 12 months, the modeled contribution after implementation is $800: $3,600 of benefit, minus $1,800 of recurring costs, minus $1,000 of setup.
At a constant $150 monthly net contribution, simple payback on setup is about 6.7 months. That is a scenario, not a promised service outcome. It assumes the benefit repeats, the workflow keeps operating, the stated costs are complete and the business can realize the modeled cash effects.
Now suppose the team originally called $300 of estimated staff capacity a cash saving, but the pilot demonstrates only $80 per month of actual cash benefit. With operating costs unchanged, the workflow contributes negative $70 per month. The 12-month result after setup is negative $1,840. The tool may still improve quality or free capacity, but the original cash-payback claim is unsupported.
This sensitivity is why DN recommends recording the mechanism of benefit. Reduced external invoices, collected incremental customer revenue and avoided paid overtime are different from a theoretical hourly value placed on time. An honest ledger can show both, provided they are not silently added together or presented as equivalent cash.
Use the DN Investment and Operating Value Ledger
The instrument presents the two scenario models separately. The investment side asks for the initial outlay, entry costs, assumed price change, distributions and exit costs. It calculates net proceeds and gain or loss. It does not choose a token, estimate a probability or retrieve a market price.
The operating side asks for setup cost, monthly operating cost, monthly cash benefit and the modeled period. It shows monthly contribution, cumulative contribution after setup and simple payback when contribution is positive. Use observed cash benefits for an operating decision; if you enter a capacity estimate, the result remains a capacity-valued scenario and must not be presented as realized profit.
The output deliberately avoids a winner label. A decision to buy software can complement a decision to hold an investment, and neither is necessarily suitable for the same money or objective. Use the instrument to expose missing assumptions, then evaluate the opportunity under its own rights, liquidity, obligations and evidence.
DN Investment and Operating Value Ledger
Change the illustrative assumptions below. This worksheet runs locally in your browser. It records your declarations; it does not verify evidence, connect to accounts, submit instructions or grant approval.
The initial numbers are synthetic. Unknown evidence remains unresolved. Downloads contain the values you enter; use aliases and avoid private or identifying information.
Ask what happens when the optimistic case fails
For equity, identify the information that would weaken the company thesis and the practical conditions for changing or exiting the position. For a token, ask which holder mechanism remains if promotional activity or speculative demand slows. For software, define the cost of shutting down the workflow, exporting records and returning to a manual or alternative process.
Time commitment belongs in the comparison. An investment can require monitoring and administration. An operating workflow can require customer acquisition, maintenance, exception handling and quality review. Calling one “passive” and another “automated” should not remove those tasks from the decision record.
DN recommends three written cases: the minimum outcome that justifies the purchase, a plausible disappointment case and the evidence that would trigger reassessment. These are decision rules chosen by the reader, not probability estimates. Where probabilities are unknown, leave them unknown instead of decorating a subjective ranking with precise risk-adjusted scores.
What evidence should you request before paying?
For a security, resolve the exact issuer, instrument, rights, disclosures, costs and intermediary arrangement. For a token, resolve the contract, actual holder rights, supply and control rules, distribution evidence and executable market access. For software, resolve the workflow, service limits, total operating costs, support, data handling and what counts as an accepted output.
A provider claiming that “AI growth benefits everyone in the ecosystem” should still explain the route to this buyer’s benefit. If the buyer must supply labor, capital, customers or infrastructure to obtain that benefit, include those contributions. A commercial opportunity can be real while being very different from the passive return implied by its marketing.
Preserve the dated terms and evidence used in the decision. If the rights, pricing or service change, the old calculation may no longer describe the current purchase. The decision record should say which assumption changed and whether that change affects the expected benefit, the cost of continuing or the ability to exit.
Methodology and limitations
This article combines primary definitions and fee guidance with an original DN comparison framework and fictional arithmetic. It makes no claim about which asset class will outperform, whether a particular token is a security, or how any reader should allocate a portfolio. It is not personalized investment, legal, tax or accounting advice.
The ledgers omit taxes, inflation, discounting, currency changes, financing, uncertain distributions and variable market impact unless the reader includes relevant amounts in the stated cost assumptions. They do not model liquidity failure, default, dilution, business interruption or a probability distribution. Simple payback also ignores the timing and risk of cash flows beyond the stated constant-benefit scenario.
DN has not measured returns or productivity from a live purchase for this study. A useful next step is a small, bounded evidence exercise appropriate to the decision: inspect the instrument’s rights, reconcile a claimed token distribution, or run a software pilot with accepted outcomes and complete costs. The evidence should determine the conclusion, rather than the common AI label.
Editorial scope: the comparison does not rank or recommend any issuer, token project or software provider. No investment transaction or subscription purchase is made by the instrument.
Sources and evidence register
Primary sources reviewed on 21 November 2026. The article distinguishes documented facts, DN analysis and synthetic examples. Recheck changeable rules and configurations before acting.
- Investor.gov: Stock — Basic equity-ownership definition; inspect the exact instrument.
- Investor.gov: Fees and Expenses — Transaction and recurring costs affect investment results; the article uses its own fictional numbers.
- ERC-20 — Token interface does not establish a universal company-equity or payout claim.
- n8n pricing — Example of execution-based workflow pricing; no specific plan price or recommendation asserted.
Frequently asked questions
Is buying an AI-agent token the same as investing in its company?
Not necessarily. Review the specific instrument and legal terms. A token may provide access, governance or another mechanism without giving the holder equity in the operating company.
Can an AI subscription be a better use of money than a token?
It can serve a different goal. A subscription should be assessed by the useful work and realized business benefit it produces after full costs; a token should be assessed by its rights, economics and investment risks. The article does not rank them for a particular reader.
Does time saved equal profit?
No. Time saved can release capacity without changing cash income or costs. Record how the released capacity actually changes revenue, spending or another explicitly stated business goal.
Does a profitable AI product guarantee a profitable token?
No. The product’s receipts, the operating company’s profits and the token holder’s rights are separate records. The economic link must be documented and verified.
Does the DN ledger predict investment returns?
No. It performs arithmetic using user-supplied hypothetical assumptions and keeps the investment and operating models separate. It does not estimate future prices or select an investment.




