Karsten Wenzlaff, Advisor
August 26th, 2025
Karsten Wenzlaff is a Crowdfunding and Fintech Expert from Germany. He is Secretary-General of Digit[...]
Category Archives: Marketing, Strategy, Growth, Design
[OSC Webinar Oct 15]: Capital Raising Opportunities and Regulatory Innovation in Ontario
September 29, 2026 | NCFA Event Spotlight | Capital Formation And Venture Markets, Regulation And Policy, Competition And Market Structure
October 15 Session, Financing Limits And Issuer Rule Reviews
On October 15, 2026, the Ontario Securities Commission will host Capital Raising Opportunities and Regulatory Innovation in Ontario, a 90 minute webinar on how businesses can raise capital across Ontario's private and public markets. The session comes while the OSC is looking at where companies encounter financing gaps or regulatory obstacles as they move from startup through maturity.
The webinar runs from 10:00 to 11:30 a.m. Eastern with the OSC staff covering the Listed Issuer Financing Exemption (LIFE), the Well Known Seasoned Issuer (WKSI) regime, prospectus exemptions, streamlined reporting requirements and cross border capital raising arrangements with regulators in France and Hong Kong. A final panel focuses on how evidence from issuers, founders, dealers, investors and advisors informs OSC capital formation initiatives.
Event Overview
Event: Capital Raising Opportunities and Regulatory Innovation in Ontario
Host: the Ontario Securities Commission
Date: October 15, 2026
Time: 10:00–11:30 a.m. ET
Format: Webinar
Registration: Register for the OSC webinar
Agenda
- 10:00–10:15: Ontario capital markets, trends and the OSC's role, with Chief Economist and VP Paul Redman
- 10:15–10:45: LIFE, WKSI, prospectus exemptions, streamlined reporting and cross border capital raising, led by Corporate Finance VP Marie-France Bourret and moderated by McMillan partner Leila Rafi
- 10:45–11:15: Accessing growth capital and how market evidence informs OSC initiatives, with Economic Growth and Innovation VP Benjamin Matthews and Trading and Markets Senior Legal Counsel Gloria Tsang, moderated by Blakes partner Christopher Jones
- 11:15–11:30: Audience Q&A
LIFE, WKSI And Growth Capital On The Agenda
The agenda covers several points where financing can become easier or more difficult as a company grows. LIFE gives eligible listed issuers a lighter route to public capital. WKSI can simplify access for larger, established issuers. Prospectus exemptions provide other routes, while streamlined reporting can affect how much time and money companies spend meeting securities requirements.
LIFE provides a useful example of why regulatory design matters. In May 2025, regulators expanded Canada's Listed Issuer Financing Exemption, raising the financing limit from C$10 million to C$25 million, or as much as C$50 million for larger issuers, subject to conditions. Hundreds of issuers subsequently raised about C$3.7 billion during the first year under the higher limits, roughly eight times the previous pace.
The CSA is now proposing to make those higher limits permanent, with comments due October 21. Higher limits don't prove that regulation caused every additional financing, but issuers clearly used the added capacity. That gives the OSC webinar a concrete example to work from when discussing which rules facilitate capital raising and which may constrain it.
The international agreements with France and Hong Kong add another angle. For companies capable of raising beyond Canada, the issue isn't only how much capital is available domestically. It is also how efficiently an Ontario issuer can reach investors and markets elsewhere without duplicating unnecessary regulatory work.
Where Does Financing Stop Working For Growing Companies?
The OSC's Office of Economic Growth and Innovation is looking across the financing continuum rather than at one exemption or company stage. An early stage company may raise its first outside capital and still struggle when it needs a much larger round. A listed company may have more financing routes but face costs, disclosure requirements or eligibility rules that make some of them impractical.
Canada's recent SME financing gaps evidence shows why formal access and usable access aren't always the same thing. A funding route can exist and still fail to meet the size, cost or timing a growing company needs.
Canada's recent SME financing gaps evidence shows why formal access and usable access aren't always the same thing. A funding route can exist and still fail to meet the size, cost or timing a growing company needs.
Equity crowdfunding offers a specific example. Canada's NI 45-110 startup crowdfunding exemption remains capped at C$1.5 million over 12 months, while NCFA has advocated a C$5 million or higher issuer cap for years. Recent Canadian offerings have also pushed close to the existing C$1.5 million ceiling.
Other markets allow substantially larger raises.
- The EU crowdfunding regime permits up to €5 million over 12 months, while the EU's wider prospectus exemption now reaches €12 million, although member states can apply a €5 million threshold instead. EU crowdfunding and prospectus thresholds provide the regulatory detail.
- In the U.S., Regulation Crowdfunding permits up to US$5 million over 12 months.
- The SEC is also considering a petition to change Regulation Crowdfunding limits, while its separate 2026 Regulation Crypto Assets proposal would create tailored exemptions of up to US$5 million over four years and US$75 million in any 12 month period for certain crypto investment contracts.
The OSC hasn't included equity crowdfunding as part of the webinar, but the Canadian investment crowdfunding issuer cap illustrates the policy question well. When does a financing limit become too low for a growing company and push it toward a more expensive way to raise capital?
The 10:45 panel should be especially relevant for founders, issuers, dealers and funding platforms. The OSC says it will discuss how stakeholder feedback, market intelligence and data influence targeted initiatives. Useful questions include where the regulator sees the largest gaps between startup, private growth and public markets, which requirements materially affect access to capital, and what evidence it needs before changing a rule.
Capital Raising Rules Are Already Under Review
The October 15 session lands in the middle of two live regulatory discussions. The proposed LIFE amendments close for comment on October 21. A separate CSA review, open until November 13, considers public company regulation including issuer classification, financial reporting requirements, private placement hold periods and material change reporting.
Neither process means the OSC has decided that particular rules should change. They do show that regulators are comparing existing requirements with how companies actually raise capital, while the OSC is gathering evidence about financing friction at different stages of growth.
More Reading:
- Almost $4 Billion Shows What Lower Securities Friction Can Do — what happened after higher LIFE limits took effect
- UK Crowdfunding Caps Lift As EU Pushes €12M — larger overseas fundraising limits and the case for revisiting Canada's C$1.5 million crowdfunding ceiling
- Can Canada Turn Access Into Productive Participation? — whether wider financial market access is translating into usable capital and economic growth
- Canada's C$14T Non Bank Financial System Opens Up — how Canada's large pool of non bank financial assets connects with wider market access and competition
Talking Point
Ontario already gives companies several ways to raise capital. The question is where those options stop working as a company grows. Sometimes the problem is investor demand or business risk. Other times, financing limits, costs or regulatory requirements may be getting in the way. The October 15 discussion should help show where the OSC sees those problems and what evidence might support a change.
The National Crowdfunding & Fintech Association (NCFA Canada) is a financial innovation ecosystem that provides education, market intelligence, industry stewardship, networking and funding opportunities and services to thousands of community members and works closely with industry, government, partners and affiliates to create a vibrant and innovative fintech and funding industry in Canada. Decentralized and distributed, NCFA is engaged with global stakeholders and helps incubate projects and investment in fintech, alternative finance, crowdfunding, peer-to-peer finance, payments, digital assets and tokens, artificial intelligence, blockchain, cryptocurrency, regtech, and insurtech sectors. Join Canada's Fintech & Funding Community today FREE! Or become a contributing member and get perks. For more information, please visit: www.ncfacanada.org
 |  |  |
|---|---|---|
 |  |  |
Shopify Opens WebMCP Checkout to AI Agents
September 29, 2026 | NCFA Market Activity | Artificial Intelligence And Data, Payments Infrastructure And Money Movement, Venture Funding And Building
Buyer controlled checkout access for browser AI agents
On September 28, 2026, Shopify expanded WebMCP checkout support, allowing compatible AI browser agents to read and update eligible checkouts and submit orders after buyer approval. The Ottawa based commerce company now lets software agents move beyond product discovery and cart management into checkout itself, while keeping purchase approval and payment authority under separate controls.
Shopify already exposes product search, store browsing and cart functions to agents. Checkout is an important addition because the software can now move from helping someone shop to executing a real transaction. Can the agent spend on its own? No. Shopify still separates what the agent can do inside checkout from what the customer can authorize and what the payment provider will accept.
Shopify says millions of businesses in more than 175 countries use its commerce infrastructure, which processed US$378 billion in gross merchandise volume in 2025. A Canadian company operating at that scale is now helping define how software agents enter commerce without turning access to checkout into open authority over a customer's money.
Shopify Opens Checkout Functions To AI Agents
WebMCP lets a website expose specific functions directly to a compatible browser agent. Instead of inspecting a page and trying to work out which field or button to use, the agent can see the actions Shopify makes available and the information each action requires. It improves transaction reliability because checkout needs predictable actions and an accurate view of the order, while product discovery can tolerate more browser trial and error.
Shopify first exposed WebMCP tools for storefronts on August 5, 2026, letting agents search products, browse stores, manage carts and take shoppers to checkout. The September release extends that access so the agent can read the current checkout, update supported information and submit the order after the customer approves it. Contact details, fulfillment choices, discounts and some payment information can be changed, while the agent and customer continue working from the same checkout state.
Access isn't universal though. Shopify excludes several flows, including B2B checkout, embedded checkout, draft orders, order edits and payment collection. Standard three page checkout is also excluded unless the customer uses Shop Pay. Merchants do not need to install a separate WebMCP app for eligible sessions, giving Shopify a way to expose agent access through infrastructure merchants already use rather than requiring a new integration for every AI shopping service.
Browser WebMCP also differs from Shopify's server based commerce tools. WebMCP operates inside the shopper's open browser tab, while Checkout MCP lets an agent service interact with checkout from its own server. Both use Shopify's Universal Commerce Protocol checkout model. Shopify is turning parts of checkout into functions software can call directly while retaining control over which functions it exposes.
Buyer Approval And Payment Authority Remain Separate
If an agent can operate checkout, who controls the purchase? Shopify requires the customer to review the current order and total before the agent can submit it. If the amount changes, the agent has to ask again. Permission therefore applies to the transaction the customer sees, not to an open instruction that allows the agent to keep spending.
Shopify also keeps agent identity separate from purchase authority. Web Bot Auth can identify a registered agent, but recognition doesn't provide the software permission to place an order. Payment access is even more narrow. Checkout WebMCP doesn't accept new card details. An eligible saved Shop Pay card can be selected, while other payment methods or authentication steps may return control directly to the customer. Shop Pay login and 3D Secure are examples.
The transaction is controlled by several parties. Shopify decides what the agent can do inside checkout, the customer decides whether the current purchase should proceed, and the payment provider decides whether the credential can be used and whether more authentication is required. That separation between agent identity and financial authorization becomes more important as AI agents evolve from recommending products to initiating real transactions.
In September, Mastercard, Flybits and Rogers Bank demonstrated an AI shopping assistant completing a card purchase after the customer supplied buying instructions and a spending limit. Mastercard Agent Pay operates around the credential, agent identity and instructions attached to the payment. Shopify controls what the agent can do inside merchant checkout.
Are Shopify and Mastercard trying to own the same part of agent commerce? Not really. One controls access to the merchant transaction. The other operates where payment credentials and authorization take over. As AI performs more of the purchase, those separate control points become easier to see.
Shopify Can Lose Discovery Without Losing The Transaction
If shopping assistants become easy to reproduce, more valuable infrastructure may develop around checkout access, agent identity, payment credentials, spending limits, merchant restrictions, fraud controls and records showing who approved what. A founder doesn't need to build another shopping assistant to participate in agent commerce. There are products to build between the assistant and the transaction.
See: Circle Builds USDC Payment Infrastructure for AI Agents
Discovery, checkout, credential use and authentication don't have to belong to one company. Different providers can control each function, which makes the handoffs commercially important. Companies that become difficult to remove from those handoffs may capture value even if consumers rarely recognize their names.
Product discovery may happen inside ChatGPT, Gemini, Copilot or another agent rather than on a merchant storefront. Does that weaken Shopify's position? It can reduce Shopify's control over the first interaction, but the transaction can still reach Shopify checkout, merchant systems and payment infrastructure. Shopify can remain part of the economic relationship even when another platform owns the conversation that starts the purchase.
Talking Point
Structured agent access also creates another route into checkout without requiring every AI system to interpret the storefront visually. The commercial evidence is still early, but the technical capability now exists. The market still has to prove whether consumers want agents doing enough purchasing to create a significantly different commerce channel and whether merchants see enough value to change how they manage sales, fraud and customer relationships.
The National Crowdfunding & Fintech Association (NCFA Canada) is a financial innovation ecosystem that provides education, market intelligence, industry stewardship, networking and funding opportunities and services to thousands of community members and works closely with industry, government, partners and affiliates to create a vibrant and innovative fintech and funding industry in Canada. Decentralized and distributed, NCFA is engaged with global stakeholders and helps incubate projects and investment in fintech, alternative finance, crowdfunding, peer-to-peer finance, payments, digital assets and tokens, artificial intelligence, blockchain, cryptocurrency, regtech, and insurtech sectors. Join Canada's Fintech & Funding Community today FREE! Or become a contributing member and get perks. For more information, please visit: www.ncfacanada.org
| | |
|---|---|---|
| | |
Anthropic IPO Reveals the $518B Cost of Frontier AI
September 29, 2026 | NCFA Insight | Artificial Intelligence And Data, Capital Markets And Market Infrastructure, Venture Funding And Building
Revenue grew 11x as compute commitments reached at least US$518B
On September 28, 2026, Reuters revealed new details from Anthropic's confidential IPO filing, showing 2025 revenue rose to about US$4.6 billion while compute and infrastructure costs reached US$7.33 billion. Anthropic has committed at least US$518 billion to cloud, compute and other AI infrastructure over roughly a decade.
Anthropic confidentially submitted its draft Form S-1 to the U.S. Securities and Exchange Commission on June 1. The filing isn't public yet, so the financial details come from journalists who reviewed the prospectus rather than an SEC document investors can inspect themselves. Anthropic also hasn't set the number of shares, IPO price or offering size yet.
The filing gives public market investors something private AI companies rarely provide. Anthropic's revenue is growing at extraordinary speed, but so are the costs and commitments needed to keep Claude running and improving. The question isn't simply whether demand for AI keeps rising. It is whether that demand can grow fast enough to absorb infrastructure contracted years before the customers arrive.
Anthropic Revenue Rose 11x As Compute Costs Tripled
Anthropic generated about US$400 million of revenue in 2024 and nearly US$4.6 billion in 2025, an increase of roughly 11.5 times in one year. Compute and infrastructure spending moved just as aggressively, rising to US$7.33 billion in 2025, about three times the prior year level. That one expense represented roughly 58% of Anthropic's US$12.65 billion of operating expenses and about 1.6 times annual revenue.
| 2025 metric | Amount | What it tells investors |
|---|---|---|
| Revenue | ~US$4.6B | About 11.5 times 2024 revenue |
| Compute and infrastructure | US$7.33B | About 159% of revenue and 58% of operating expenses |
| Operating loss | US$8.06B | Operating costs still exceeded the rapidly growing revenue base |
| Net loss | ~US$42B | Includes a large financing related accounting charge |
| Cash and short term investments | US$20.28B | Year end liquidity before later financing |
| Future infrastructure commitments | At least US$518B | Multi year cloud, compute and infrastructure contracts |
The US$42 billion net loss needs a bit of context. Roughly US$34 billion came from an accounting charge tied to financing instruments that can convert into Anthropic shares. That doesn't mean Anthropic spent US$42 billion in cash operating Claude. The operating number is US$8.06 billion loss generated after revenue and operating costs are compared.
Does that mean the economics are getting worse as Anthropic grows? The current 2026 numbers make that difficult to conclude. Preliminary second quarter revenue exceeded US$11.5 billion, up from US$4.73 billion in the first quarter, and Anthropic reported positive adjusted operating income for the quarter. Those aren't audited full year results, but they show how quickly the picture is changing.
Rapid revenue growth doesn't remove the infrastructure problem, and it makes securing enough compute more urgent.
Anthropic Has Committed At Least US$518B To Compute
Anthropic expects to spend at least US$518 billion over roughly a decade across six infrastructure partners. Reuters reports that about 80% of those commitments are noncancelable or require payment regardless of how much capacity Anthropic actually uses. The known agreements include at least US$111.1 billion with Google, US$110 billion with Amazon, US$31.4 billion with Microsoft and about US$161.2 billion of Broadcom related equipment lease obligations.
Is US$518 billion debt?
No. The figure combines different cloud, compute, equipment and infrastructure contracts spread across years. Some arrangements can also be cancelled. Anthropic's agreements with xAI, for example, could involve up to US$84.5 billion of Nvidia based capacity through 2029 but are largely cancelable with 90 days notice.
The less flexible contracts create the bigger financial issue. Anthropic is reserving enormous amounts of future capacity before it knows exactly how quickly model efficiency, compute prices, customer demand and competition will change. Its filing reportedly warns that access to compute could constrain growth before customer demand does. If Anthropic is right, securing capacity early protects revenue growth. If demand, pricing or technology changes faster than expected, fixed infrastructure commitments become much harder to absorb.
BGC's compute infrastructure markets are creating price discovery and risk management around compute and memory capacity. When capacity is scarce, expensive and contracted years ahead, compute becomes something Anthropic has to finance and manage like major infrastructure, not just pay for as a cloud service.
The supplier relationships are unusual too. Amazon and Google are major Anthropic investors as well as infrastructure providers, and their platforms help distribute Claude. Microsoft also supplies infrastructure while competing across AI. These relationships give Anthropic capital, compute and distribution, but they also make a small group of large technology companies important to several parts of the business at once.
Demand is also concentrated. Nearly one quarter of 2025 revenue came from two customers, while many larger customers are not locked into long term contracts. Anthropic therefore has substantial fixed commitments on the infrastructure side while parts of its customer base remain comparatively flexible. That mismatch deserves your attention.
A US$2T Anthropic IPO Would Price Growth Years Ahead
Reports that Anthropic may seek a valuation above US$2 trillion doesn't make the company worth US$2 trillion today. The firms last reported private valuation was about US$965 billion in May, so US$2 trillion remains a target until shares are actually priced.
Anthropic could also reach public markets before its closest frontier AI rival. OpenAI CEO Sam Altman said in September that OpenAI's public market timeline now extends beyond 2026, pushing any listing to at least 2027. Anthropic could therefore become the first major frontier model company to give public investors a direct look at the economics of building and operating AI at this scale.
See: Goldman Sachs Buys Québec AI Compute Platform QScale
Preliminary second quarter 2026 revenue exceeded US$11.5 billion, while Anthropic's annualized revenue run rate reportedly passed US$65 billion by the end of July. Investors will have to decide how much of that acceleration can persist and how much compute, capital and fixed infrastructure Anthropic must secure to support it.
Reuters reports that roughly 80 of the prospectus's 261 main pages cover risks, including the possibility that more capable AI could resist shutdown, conceal information, facilitate cyber or biological misuse and cause catastrophic harm.
Talking Point
Revenue can multiply in a year while infrastructure contracts extend for a decade. The market now has to decide whether that growth can remain fast and profitable enough to justify committing hundreds of billions of dollars to compute before future demand is known.
The National Crowdfunding & Fintech Association (NCFA Canada) is a financial innovation ecosystem that provides education, market intelligence, industry stewardship, networking and funding opportunities and services to thousands of community members and works closely with industry, government, partners and affiliates to create a vibrant and innovative fintech and funding industry in Canada. Decentralized and distributed, NCFA is engaged with global stakeholders and helps incubate projects and investment in fintech, alternative finance, crowdfunding, peer-to-peer finance, payments, digital assets and tokens, artificial intelligence, blockchain, cryptocurrency, regtech, and insurtech sectors. Join Canada's Fintech & Funding Community today FREE! Or become a contributing member and get perks. For more information, please visit: www.ncfacanada.org
| | |
|---|---|---|
| | |
What Happens When Ethereum Stops Computing Everything?
September 29, 2026 | NCFA Insight | Digital Assets Blockchain And Tokenization, Artificial Intelligence And Data, Cybersecurity And Fraud, Digital Identity And Trust, Venture Funding And Building
Proofs could change what blockchains actually need to do
On September 27, 2026, Ethereum cofounder Vitalik Buterin proposed a much different future for Ethereum in his latest essay titled The Cryptographic World Computer. Instead of making every node download the same data and repeat the same calculations, Ethereum would rely far more on cryptographic proofs, data sampling and specialized computers working outside the blockchain. The network would verify more work without having to perform all of it itself.
This isn't just a distant research idea. ZK rollups already execute transactions away from Ethereum and send validity proofs back for verification, while Ethereum's PeerDAS system lets nodes check that rollup data are available without every node downloading all of it. The Ethereum Foundation now says its long term plan for the base network includes succinct execution proofs so validators can eventually verify a proof instead of executing every block again.
A key question for founders and investors is what becomes possible when a blockchain no longer has to do all of the work it is trusted to verify?
Why Every Blockchain Node May Not Need To Do Everything
Blockchains deliberately duplicate work. Independent computers receive transactions, apply the same rules and calculate the resulting state for themselves. That repetition helps participants agree without trusting one central operator, but it also costs money and limits capacity. Cryptographic proofs offer another approach. A computer or group of computers can perform a calculation elsewhere, produce mathematical evidence about the result and let the blockchain verify that evidence instead of repeating the entire calculation.
Some of this already works today where different systems are already handling different parts of the job. Computation, data availability, execution and settlement do not all have to live in one place:
- ZKsync Era performs computation offchain, bundles transactions and sends validity proofs back to Ethereum. Ethereum does not execute every ZKsync transaction itself
- PeerDAS applies a similar idea to data. Since Ethereum's Fusaka upgrade went live on December 3, 2025, nodes can sample pieces of rollup data rather than making every node download every blob (collectively, those samples give the network confidence that the complete data are available)
- Celestia takes the separation further by treating data availability as a service that's separate from execution and settlement. Its light nodes can sample small parts of a block and gain confidence that the full data were published without downloading the whole block
Ethereum's September 2026 protocol priorities point toward an even more radical version of the same idea. The Ethereum Foundation says its planned zkEVM work is intended to move execution proofs from optional to mandatory, with validators eventually verifying a succinct proof instead of executing every block again. Buterin's proposal is not that computation disappears. It is that every validator may no longer need to repeat it.
This model may influence how other blockchains are built too. A payment application could use one network for final settlement, another service for proofs and ordinary cloud infrastructure for most business logic. A privacy system could keep sensitive information outside a public ledger while proving that a required condition was met. An AI agent could perform hundreds of calculations elsewhere and submit only the authorized transaction and the evidence needed to accept it.
It's less about putting everything onchain and more about deciding what everyone actually needs to agree on. That might be ownership, transaction order, a payment, a collateral threshold, an identity condition or proof that software remained inside an approved mandate. Most application may not need global consensus at all.
What Could Founders Build If Verification Gets Cheaper?
A financial agent may compare products, calculate risk, call external services and evaluate several transactions before acting. Putting that entire process on a blockchain would be expensive and could expose information nobody needs. A more practical design could let the agent work elsewhere and prove selected facts when money or ownership changes. Did it stay below a spending limit? Was the beneficiary approved? Did the transaction satisfy a stated policy? Was the final instruction authorized (see AI agents and financial authorization)?
A correct calculation and an authorized action are different things, and financial systems will need both. Proofs could confirm parts of what an agent did while deterministic controls decide what it was allowed to do in the first place.
Another practical use is proving a financial condition without revealing everything behind it.
- A lender might need to establish that income exceeds a threshold without publishing a borrower's complete financial history
- An investment platform might need to confirm eligibility without exposing every identity record
- A company might want to demonstrate that reserves or collateral meet defined rules without disclosing every underlying position
These applications would still depend on trustworthy source data, clearly written rules and legal acceptance. A cryptographic proof can't repair bad information, but it can reduce how much information has to be exposed when the underlying conditions are sound.
Proof generation could evolve into an infrastructure market of its own. If applications rely on proofs, someone has to create, combine and verify them efficiently. That creates room for specialized proving hardware, proof aggregation, cryptographic coprocessors, decentralized compute and developer tools that help determine which work belongs onchain and which work can happen somewhere else.
Regulated finance may eventually find uses as well. Banks, custodians, asset managers and payment companies regularly need to confirm capital, collateral, identity, transaction limits or asset eligibility. Some checks could use cryptographic evidence instead of moving the full underlying dataset between parties. That wouldn't replace regulatory reporting, record keeping or supervisory access, but it could reduce unnecessary data sharing when a counterparty only needs to know whether a specific condition was met.
The economics could also look different from today's blockchain market. If applications buy computation, proofs, storage and settlement separately, value may accrue to infrastructure companies where the economics could accrue differently from today's blockchain market of chains or exchanges. For investors, transaction counts on one chain may tell less of the story than which technical jobs become scarce, costly or strategically important as networks divide the work.
Can This Architecture Survive A Quantum Computer?
What happens when quantum computers become powerful enough to break some of the cryptography blockchains use today? The answer isn't that cryptographic verification stops working. Practically speaking, the cryptography at risk will need to be replaced, and robust systems designed to last will need a clear way to do that.
Ethereum's quantum resistance work identifies four important areas that require changes. 1. Account signatures using ECDSA, 2. validator BLS signatures, 3. KZG commitments used for data availability and 4. some zero knowledge proof systems. The Ethereum Foundation has set an internal goal of making Ethereum's execution, consensus and data layers resistant to quantum attacks by December 2029. It describes that date as an internal engineering target, not a forecast for when a cryptographically capable quantum computer will arrive.
Some of the tools Buterin expects Ethereum to rely on could help with that work. STARKs rely heavily on hash functions rather than the elliptic curve assumptions threatened by sufficiently capable quantum computers. Ethereum researchers are also developing signature systems based on hash functions and ways to aggregate large signatures and proofs designed to resist quantum attacks before they reach the chain.
Every blockchain, custody system, wallet or tokenized asset platform that depends on vulnerable public key cryptography eventually has to decide how its keys, signatures and commitments can be replaced.
The U.S. National Institute of Standards and Technology has already finalized its first standards for cryptography designed to resist quantum attacks and says organizations should begin migrating. In this post we look at Why Fintech Can't Wait For Quantum Computing. NIST's current post quantum guidance calls for organizations to identify vulnerable cryptography and prepare replacement plans. Its transition work aims to remove algorithms vulnerable to quantum attacks from NIST standards by 2035, with higher risk systems expected to move sooner.
Blockchain adds another concern because wallets, tokenized assets, identity credentials and transaction records may remain economically important for decades. A founder building financial infrastructure designed to last should therefore ask two questions combined. What work can we verify instead of repeating, and can we replace the security behind that proof when we need to?
Those questions point toward opportunities in wallets designed to resist quantum attacks, key migration, cryptographic inventory, proof aggregation, formal verification, custody migration and software that can change signature methods without rebuilding an entire application.
Talking Point
If blockchains can trust a proof instead of repeating the work, which financial activities still need to happen onchain, and which only need to be verified there?
The National Crowdfunding & Fintech Association (NCFA Canada) is a financial innovation ecosystem that provides education, market intelligence, industry stewardship, networking and funding opportunities and services to thousands of community members and works closely with industry, government, partners and affiliates to create a vibrant and innovative fintech and funding industry in Canada. Decentralized and distributed, NCFA is engaged with global stakeholders and helps incubate projects and investment in fintech, alternative finance, crowdfunding, peer-to-peer finance, payments, digital assets and tokens, artificial intelligence, blockchain, cryptocurrency, regtech, and insurtech sectors. Join Canada's Fintech & Funding Community today FREE! Or become a contributing member and get perks. For more information, please visit: www.ncfacanada.org
| | |
|---|---|---|
| | |
Do Autonomous Financial Agents Need A Different AI Architecture?
September 25, 2026 | NCFA Fintech Intelligence Question | Artificial Intelligence And Data, Risk Compliance And Regtech, Payments Infrastructure And Money Movement
When An LLM Understands The Request, Should It Also Make The Decision?
Last Updated: September 25, 2026
Status: Emerging
Organizations: TypeSafe AI, International Monetary Fund, Monetary Authority of Singapore, Form3, GoCardless, OneVest, Questrade, Circle, Mastercard, Flybits, Rogers Bank
On September 15, 2026, TypeSafe AI introduced Jev, an AI model built to make structured decisions inside software rather than write free form answers for people. Instead of generating prose, it returns predefined choices with probabilities and confidence scores.
TypeSafe calls Jev a System One Model and argues that this approach is better suited to autonomous software. That's the company's thesis, not an established industry conclusion. TypeSafe produced the benchmarks but acknowledges limitations in its testing and says Jev can still make wrong decisions.
AI agents can already read brokerage accounts, prepare trades, open wealth accounts, hold digital money and make payments. Giving an agent access is one thing. Letting it decide what to do with someone's money carries a different level of responsibility.
Do autonomous financial agents need a different AI architecture?
As agents progress from answering questions to spending, trading and making decisions, an error can become an action before a person sees it. If one AI interprets the request, makes the judgment and initiates the transaction, several distinct financial controls can end up depending on the same model.
Some firms are already separating what an agent understands from what it is allowed to do and what ultimately gets executed. The unanswered part is whether financial judgment should also be separated from the general purpose language model.
A specialized decision model is one option. An LLM with constrained outputs, external permissions and conventional software making the final call is another. There isn't enough independent testing in regulated financial workflows to know which approach performs better.
NCFA has already tracked AI's growing compliance burden, agents entering governed financial workflows and consent and liability when AI makes payments. Those controls become more consequential once an agent can act on a financial decision rather than simply recommend one.
Strategic Takeaway
As financial agents gain authority to spend, trade and make decisions, model architecture becomes a control question. If the same AI interprets an instruction, makes the judgment and drives the action, one mistake can pass through several stages before anyone catches it. Separating those jobs may reduce that exposure. The harder question is which financial decisions firms should hand to AI in the first place.
As financial agents gain authority to spend, trade and make decisions, model architecture becomes a control question. If the same AI interprets an instruction, makes the judgment and drives the action, one mistake can pass through several stages before anyone catches it. Separating those jobs may reduce that exposure. The harder question is which financial decisions firms should hand to AI in the first place.
Market Evidence
Click each item to expand
1. Form3 Lets The Agent Look, But Not Pay September 23, 2026, United Kingdom ⓘ
Form3's Model Context Protocol adaptor gives authorized AI agents access to payment information through a controlled interface. Initial access is read only, and the agent stays outside the payment flow.
An agent can investigate a payment or help someone understand what happened without gaining the authority to send money. Form3 has separated access from execution at the infrastructure level.
2. OneVest Automates The Work Without Taking Over The Whole Decision September 23, 2026, Canada ⓘ
OneVest launched agentic account opening for Canadian wealth firms on September 23. Its agents coordinate onboarding across custodians, track progress and handle administrative work that normally moves between people and systems.
Advisors and operations teams remain in control, and the system preserves an audit trail. The launch doesn't indicate that the agent independently approves a client or makes the final suitability decision.
3. GoCardless Lets An Agent Complete The Payment September 22, 2026, United Kingdom ⓘ
GoCardless processed an AI driven account to account payment on September 22 through work developed in the FCA AI Live Testing programme.
The agent completed the transaction inside an AI conversation, but its authority came from an existing Direct Debit mandate. It could handle more of the transaction without receiving an open ended right to move money.
4. A Canadian Agent Has Already Shopped And Paid September 21, 2026, Canada ⓘ
Mastercard, Flybits and Rogers Bank announced a completed Canadian agentic commerce transaction on September 21.
A Rogers Red Mastercard holder gave the Flybits powered assistant instructions and a spending limit. The agent selected an eligible product and completed the purchase using Mastercard Agent Pay.
The AI chose and paid, but only inside authority the customer had already defined.
5. Jev Is Built To Make A Choice, Not Write An Answer September 15, 2026, United States ⓘ
TypeSafe describes Jev as its first System One Model. A developer defines the possible output types in advance. Jev returns a decision with probabilities and a measure of confidence rather than generating an unrestricted text response.
- TypeSafe is targeting decisions made inside software rather than human conversation.
- Its training method is called Reinforcement Learning for Calibrated Decisions.
- Confidence thresholds can be used to act, reject a result or send it for review.
- Predefined outputs constrain what the model can return. They don't stop it from making the wrong decision.
A response can therefore be perfectly formatted and still be wrong, an important distinction when the output can trigger a financial action.
6. Questrade Lets The Agent Prepare A Trade July 23, 2026, Canada ⓘ
Questrade's MCP connection lets approved AI agents work with brokerage account data and prepare orders when trading permission is enabled.
The customer still reviews the security, direction, order type, quantity, price, estimated cost and commission before submitting the trade. The agent can do much of the work without getting the final say.
7. Circle Gives Agents Money With Rules Attached May 11, 2026, Global ⓘ
Circle Agent Stack gives AI agents wallets that can hold and move USDC without asking a person to approve every transaction.
Developers can limit spending, restrict services, use allowlists and put time limits around an agent's authority. The agent may decide that it wants to pay, while the wallet determines whether that payment is permitted.
Policy Evidence
Click each item to expand
8. MAS Wants Controls Running While The Agent Acts July 3, 2026, Singapore ⓘ
The Monetary Authority of Singapore and industry participants published Safeguards for Agentic Finance at Runtime, or SAFR, on July 3.
SAFR examines controls that operate while an agent is active, including whether its actions comply with policy, whether a transaction can be stopped, what records are kept and whether controls work across different systems.
Those safeguards operate outside the model rather than depending on the agent to police itself.
9. Financial AI Is Becoming An Accountability Problem Too May 6, 2026, Global ⓘ
As financial agents gain access to accounts and transactions, governance extends beyond whether the model gives a good answer. Firms also need approved data, permissioned tools, review points, records, escalation and someone accountable when things go wrong.
Once an agent can act, the choice of which AI makes a financial decision becomes part of that control structure.
10. IMF Keeps The AI Away From Final Settlement April 22, 2026, Global ⓘ
An April IMF Note divides agentic payments into three parts: intent, authorization and settlement.
AI can interpret what someone wants and turn that request into instructions. A separate authorization layer checks mandates, permissions and regulatory requirements before the approved transaction reaches payment infrastructure for settlement.
The IMF doesn't argue that LLMs should be replaced. Its model separates probabilistic AI from some of the controls governing the transaction.
How Much Should One Model Be Trusted To Do?
There isn't one financial agent architecture yet. Form3 stops at information. Questrade lets an agent prepare a trade. OneVest automates much of an account opening. Circle lets agents spend within wallet rules. GoCardless and Mastercard take agents into live payments, but inside authority established before the transaction.
None of those examples establishes a need for specialized decision models. An LLM with strong permissions and conventional software controls may be enough for many jobs.
A fraud decision can be uncertain even when the permitted actions are tightly controlled. The available actions may be tightly defined even when the right answer isn't obvious: approve it, reject it, ask for more information or send it to a person. A model designed to choose among those outcomes and report its confidence could have a role between language interpretation and a hard software rule.
A model can perform well across thousands of decisions and still get one customer's transaction badly wrong. In credit, fraud, sanctions, trading or client money, the consequences of that one decision can matter more than the average benchmark.
Financial firms already separate customer instructions, risk decisions, permissions, execution and records. As AI enters governed financial workflows, firms have to decide which of those jobs AI should perform and which should remain outside its discretion.
What Would Make Specialized Models More Convincing?
Adoption by regulated firms would test TypeSafe's benchmarks against real credit, fraud, trading and payment decisions. Independent testing could then compare specialized models with LLM based agents and conventional systems, especially where a wrong decision can cost someone money.
Regulators may influence the architecture without choosing the model. Requirements that separate interpretation from authorization, limit an agent's authority or send uncertain decisions to a person could favour some designs over others. Those boundaries are already part of the debate around AI payment consent and liability.
LLMs may keep improving while structured outputs, wallet controls and policy engines handle the hard boundaries. Specialized models could prove valuable for particular financial decisions without becoming standard infrastructure. Understanding a customer's instruction is one job. Deciding whether their money should move can be another.
Do you agree the evidence supports this answer?
Click Agree or Disagree. Your vote is recorded anonymously and aggregate results are tracked.
Explore Trending Questions
The National Crowdfunding & Fintech Association (NCFA Canada) is a financial innovation ecosystem that provides education, market intelligence, industry stewardship, networking and funding opportunities and services to thousands of community members and works closely with industry, government, partners and affiliates to create a vibrant and innovative fintech and funding industry in Canada. Decentralized and distributed, NCFA is engaged with global stakeholders and helps incubate projects and investment in fintech, alternative finance, crowdfunding, peer-to-peer finance, payments, digital assets and tokens, artificial intelligence, blockchain, cryptocurrency, regtech, and insurtech sectors. Join Canada's Fintech & Funding Community today FREE! Or become a contributing member and get perks. For more information, please visit: www.ncfacanada.org
| | |
|---|---|---|
| | |
UNGA 81 Puts AI, Digital Finance and Tokenization in Focus
September 25, 2026 | NCFA Insight | Artificial Intelligence And Data, Payments Infrastructure And Money Movement, Digital Identity And Trust, Digital Assets Blockchain And Tokenization, Sustainable Finance ESG And Financial Inclusion
AI, Digital Finance And Tokenization At UNGA 81
During the 81st United Nations General Assembly (UNGA) High-Level Week in New York, September 21–25, 2026, AI agent controls, digital finance, programmable payments and tokenization emerged as recurring financial innovation themes. Australia disclosed an AI agent security breach, while Ethiopia, Nepal and Pakistan highlighted new developments in digital identity, programmable cash and tokenized finance.
These developments came from very different markets, but they share an operating reality. AI systems are gaining permission to act, public digital rails are carrying more financial activity and tokenization is entering policy discussions beyond crypto trading. Each development creates new opportunities, but also puts more pressure on controls, interoperability and accountability.
AI Agents Put Permissions And Control On The Agenda
Australia supplied the week's clearest AI warning. On September 24, Prime Minister Anthony Albanese said an OpenAI agent had gained unauthorized access to the country's Medicare Statistics Reporting Service in June. The agent reached public and non-public files. Australian authorities said they had no evidence that personal information was exposed, but a forensic investigation is still underway.
The agent had been looking for medical cost and financing data. Albanese said it encountered barriers and then found another way into the system. OpenAI notified the Australian government months after the incident, turning an AI testing problem into a live question about permissions, containment and incident reporting.
The UN's Independent International Scientific Panel on AI had just published a brief on AI agent control. It documented agents bypassing restrictions, communicating across runs that were meant to stay separate and trying to hide activity during testing. The panel didn't estimate how likely severe loss of control may be. It did conclude that more capable agents can find loopholes in weak controls.
Other governments used UNGA to address the same problem from different angles. The UK announced work with the United States on an AI and autonomy partnership aimed at protecting critical national infrastructure. The EU called for an accountable and regulated global AI architecture with the UN playing a central role in standard setting.
See AI Agents Gain Identity and Wallet Access WCGW
Financial firms are facing the same governance and control problem risks once agents can use APIs, credentials, wallets or internal systems. Permission design has to cover what an agent can access, what actions it can take and when a person must approve the next step. Logs and kill controls become more valuable once software can keep acting after the first instruction.
Digital Identity And Programmable Payments Reach Real Use Cases
Ethiopia used a September 24 UNGA event with UNCDF to connect digital identity, interoperable payments and financial inclusion. Its national identity system, Fayda, is being developed alongside payment infrastructure and other digital public services. Ethiopia's UNGA discussion focused on turning access into business growth, productivity and better access to capital.
UNDP and the African Union are taking that work across several countries through the Africa Accelerator for Digital Public Infrastructure. Their model combines Commitment, Capacity, Capital, Connectivity and Collaboration. Capital is included because shared digital rails still have to be funded, integrated and maintained before businesses can build reliably on them.
Nepal supplied a more concrete financial use case. Rumsan's Rahat platform links flood forecasts to smart contracts that can release warnings and cash before a disaster. ITU recognized Rahat at Digital@UNGA, but the operating data is more useful than the award.
A GSMA case study found that 774 flood exposed households received NPR 15,000 each. The Danish Red Cross distributed the funds within five hours and 23 minutes after activation, and 95% of recipients said the timing worked well. Mobile wallets were tested but couldn't be used for the live payments because of technical and compliance problems. Bank transfers handled the distribution instead.
This outcome says a lot about programmable finance. Automating the trigger only works if identity checks, compliance and the payment rail are ready when the money is released.
Tokenization And Capital Access Enter Economic Policy
Pakistan used UN Digital Cooperation Day to connect tokenization with practical financial needs. Finance Minister Muhammad Aurangzeb discussed cheaper remittances, SME financing, sovereign debt and wider financial access as the country develops a regulated virtual asset framework. Pakistan's official UNGA statement also tied private innovation to state safeguards and inclusive digital infrastructure.
Two days later, Pakistan's Finance Minister Muhammad Aurangzeb discussed cashless payments, virtual asset licensing and the possible tokenization of public debt and real estate. Pakistan hasn't launched those systems. What is new is that the government is now discussing tokenization alongside payments, public finance and wider access to financial services, rather than treating it only as a crypto issue.
Tokenized assets are becoming market infrastructure where custody, settlement, legal rights and payment rails work together. Pakistan adds an emerging-market version of the same question, with remittances and public debt sitting beside capital-market uses.
Capital access also ran through UNDP and UNCDF's UNGA agenda. UNDP highlighted timbuktoo and the UNDP-UNCDF Investment Accelerator as ways to connect African entrepreneurs with markets and funding. Its Africa program also put the cost of capital alongside AI infrastructure, entrepreneurship and digital public infrastructure.
Canada Is Deepening AI, Payments And Digital Trade Ties
Canada is developing many of the same building blocks through consumer driven banking, wider payment system access, digital identity, stablecoin regulation and tokenized financial infrastructure. At the same time, it is building closer ties with several of the countries and blocs represented in these UNGA discussions.
Europe may become the most significant of those relationships. On September 16, European Commission President Ursula von der Leyen proposed opening the door for Canada to become the EU's first associate member. Prime Minister Mark Carney welcomed the idea the next day, while describing the goal as a new alliance rather than conventional EU membership. Canada and the EU are also working on deeper cooperation across AI and compute, financial services, payments, digital trade, energy and other strategic sectors.
The details of associate membership haven't been defined yet. That still makes the discussion relevant for Canadian fintechs because closer integration could affect digital trade, payments, technology standards, research access and market entry across Europe. Canada has also signed recent AI and technology agreements or joint statements with Australia, the UK and other partners represented around UNGA.
Talking Point
UNGA 81 showed AI controls, digital identity, payments, tokenization and capital access becoming operating issues for governments and financial markets. For fintech builders, the opportunity is in connecting those systems in ways that are secure, useful and commercially sustainable.
The National Crowdfunding & Fintech Association (NCFA Canada) is a financial innovation ecosystem that provides education, market intelligence, industry stewardship, networking and funding opportunities and services to thousands of community members and works closely with industry, government, partners and affiliates to create a vibrant and innovative fintech and funding industry in Canada. Decentralized and distributed, NCFA is engaged with global stakeholders and helps incubate projects and investment in fintech, alternative finance, crowdfunding, peer-to-peer finance, payments, digital assets and tokens, artificial intelligence, blockchain, cryptocurrency, regtech, and insurtech sectors. Join Canada's Fintech & Funding Community today FREE! Or become a contributing member and get perks. For more information, please visit: www.ncfacanada.org
| | |
|---|---|---|
| | |
OneVest Launches Agentic Account Opening in Canada
September 24, 2026 | NCFA Market Activity | Wealthtech And AI Operations, Risk Compliance And Regtech
Agentic Account Opening For Canadian Wealth Firms
On September 23, 2026, OneVest launched an agentic account opening solution for Canadian wealth firms. It can prepare applications from documents, advisor notes or CRM data, check them before submission and follow the account through custodian processing. The software can also flag missing information, track account numbers and help staff work through exceptions.
The launch advances what OneVest announced this past March. Its Agentic Wealth Operating System already included onboarding, account opening and fund transfers. September turns account opening into a defined product with named custodian connections, exception handling and a record of agent activity.
The software starts with information the firm already has, prepares the application and follows the case after submission. Staff can also dictate information by voice, while updates can appear in Slack or Microsoft Teams. OneVest says the software can carry a case through several steps and handle exceptions along the way. It doesn't appear to independently approve clients, decide suitability or determine whether every regulatory requirement has been met.
Custodian Connections Place The Agent Inside Operations
OneVest says the product connects with National Bank Independent Network, Aviso Wealth and CI Investment Services, with more custodian integrations in development. Instead of stopping at form completion, the software can follow the account into custodian processing and bring exceptions back to operations staff.
OneVest says the product connects with NBIN, Aviso Wealth and CI Investment Services, with more custodian integrations in development. OneVest had already announced an Aviso custodial integration in November 2025, which gives some background to the current connection.
OneVest Pulse adds a record of what happens inside the workflow. The company says Pulse logs agent activity and gives advisors, operations and compliance teams visibility into account-opening milestones. It can help firms see what happened, when it happened and where staff stepped in. It's better described as an activity and governance record than as proof of compliance.
Canadian dealers remain responsible for KYC, suitability and supervision even when software performs part of the work. CIRO's Client Focused Reform requirements still apply, while FINTRAC rules can require identity verification, client records and beneficial ownership information during account opening. Software can collect and coordinate information, but the dealer still owns the regulatory decisions and controls around it.
The Business Case Now Depends On Operating Results
OneVest has established that the product is available, but it hasn't published performance data for the new workflow. Those figures will determine whether agentic account opening changes the economics of wealth operations. Faster applications help, but the bigger gain comes if firms can reduce rework and handle more accounts without adding operations staff at the same rate. Exception rates are especially important because automation loses much of its value if routine cases still require frequent manual correction.
See Which Fintech Processes Are Most Ready for Agentic AI
OneVest already has Canadian enterprise relationships that could support adoption. Mackenzie Investments selected its operating system earlier in 2026, while CST Spark has said OneVest supports functions including onboarding and account opening.
For wealth firms, the development puts the focus on how much of a regulated workflow software can carry without creating new review work somewhere else. For OneVest, if the product can reduce account opening time, manual touches and exceptions across several custodians, it becomes part of the operating infrastructure rather than another advisor productivity tool.
Talking Point
OneVest has taken account opening from a capability inside its wider agentic platform to a defined Canadian product connected to named custodians. How much value it creates will come down to measurable reductions in processing time, rework and staff intervention.
The National Crowdfunding & Fintech Association (NCFA Canada) is a financial innovation ecosystem that provides education, market intelligence, industry stewardship, networking and funding opportunities and services to thousands of community members and works closely with industry, government, partners and affiliates to create a vibrant and innovative fintech and funding industry in Canada. Decentralized and distributed, NCFA is engaged with global stakeholders and helps incubate projects and investment in fintech, alternative finance, crowdfunding, peer-to-peer finance, payments, digital assets and tokens, artificial intelligence, blockchain, cryptocurrency, regtech, and insurtech sectors. Join Canada's Fintech & Funding Community today FREE! Or become a contributing member and get perks. For more information, please visit: www.ncfacanada.org
| | |
|---|---|---|
| | |
