Everyone's throwing massive money at the same fundamental bottleneck. So what's Google's differentiator in this infrastructure arms race?

Last week, Google became the first major tech company to publish detailed environmental impact data for AI operations, revealing that median Gemini queries consume 0.24 watt-hours with a remarkable 33x efficiency improvement since May 2024⁵. Critics immediately questioned whether these figures capture the full environmental cost, with some researchers suggesting water consumption could be far higher than Google's "five drops" claim⁶.

But the accuracy debate misses the strategic brilliance entirely. This isn't environmental virtue signaling—it's warfare.

Infrastructure Access, Not Capital, Determines Winners

Here's what the spending announcements don't tell you: money can't solve infrastructure scarcity. When transmission projects take 10+ years to complete and grid connections require 5+ years⁷, even unlimited capital hits hard physical constraints. The International Energy Agency reports that 20% of planned data center projects already face delays due to grid connection issues⁸.

The math is unforgiving. AI data centers could consume 945 terawatt-hours annually by 2030—more electricity than Japan uses in an entire year⁹. That represents nearly triple current consumption, with AI accounting for the fastest-growing segment. When your industry needs infrastructure equivalent to entire countries' electrical capacity, throwing money at hardware procurement isn't enough.

This reality is already reshaping competitive dynamics in ways most analysts haven't recognized.

The Competitive Gap With China

Recent reports from AI experts returning from China reveal a sobering competitive reality: while the U.S. faces "extreme bottlenecks" in power infrastructure, China treats energy availability as a "solved problem"¹⁰. This isn't just about current capacity—it's about the speed of deployment and regulatory efficiency that will determine where AI innovation happens over the next decade.

American companies must overcome this infrastructure disadvantage through superior efficiency and regulatory relationships. Google's environmental transparency strategy makes perfect sense against this backdrop. Rather than simply competing on algorithmic capability or capital deployment, Google is building the regulatory credibility and utility partnerships that will become increasingly crucial as infrastructure constraints tighten.

From Software to Power Generation: The Strategic Shift

The infrastructure constraint runs deeper than grid capacity — it's forcing tech companies to secure energy on across scarce transmission capacity from utilities, building or contracting large-scale solar and wind farms and investing in long-lead, high-capacity reactors as reliable baseload supply.

Examples are already visible: Microsoft signed a 20-year deal to restart Three Mile Island's reactor (835 MW)¹¹. Google partnered with Kairos Power for 500 MW of next-gen nuclear expected by 2030¹². Amazon has committed to 5 GW through renewables and nuclear partnerships¹³. And across the board, hyperscalers are racing to lock in solar and wind PPAs (power purchase agreements) to satisfy customer sustainability requirements.

This represents a fundamental shift: tech companies are no longer just software builders. They are vertically integrating into energy procurement and generation because existing grid capacity alone cannot support AI's scaling plans.

But every source of power — whether solar farms, wind projects, or nuclear reactors — comes with heavy permitting and political complexity. Google's environmental transparency becomes even more strategic in this context, creating credibility with three crucial stakeholders:

Utility executives managing strained grids who must allocate scarce capacity — efficiency signals responsible growth.

Regulators across nuclear, renewables, and transmission who control permit approvals — transparency shows partnership over extraction.

Enterprise customers under sustainability mandates — efficiency metrics combined with clean energy sourcing make Google a lower-risk supplier.

Environmental data, in other words, isn't just corporate responsibility. It's a competitive weapon in the race for energy access.

The Strategic Advantages That Compound

Google's transparency creates several compounding advantages that competitors will struggle to match:

• First-mover advantage in infrastructure politics. While competitors focus on technical capabilities, Google builds regulatory credibility and utility partnerships that determine priority access to scarce grid capacity when infrastructure access becomes the primary constraint.

• Geographic arbitrage through data-driven deployment. Companies that understand their true resource requirements can strategically locate operations in power-rich regions before competitors recognize these advantages. Environmental measurement enables more sophisticated global infrastructure strategies.

• Customer differentiation through compliance leadership. Enterprise sustainability mandates create demand for measurable efficiency. First movers in environmental disclosure can command premium pricing while competitors face margin pressure from regulatory catch-up costs.

• Optimization through measurement. Google's granular environmental tracking reveals system inefficiencies that compound across massive scales—small percentage improvements translate to gigawatt-level savings when operating at country-scale power consumption.

The Economics Revolution: From Subsidies to Sustainable Unit Economics

Today's AI pricing still reflects venture-funded subsidies designed to capture market share, not real costs. But infrastructure demands make this model unsustainable. IBM projects computing expenses will rise 89% by 2025, with AI the main driver¹⁴. McKinsey estimates $5–8 trillion in infrastructure spending is required by 2030¹⁵.

That math doesn't support a subsidy war forever. Efficiency and transparency accelerate the inevitable shift toward sustainable economics — not by guaranteeing higher prices, but by changing the competitive ground rules:

Margin protection: Efficiency lowers Google's own costs. In a market where customers resist price hikes, this means Google can protect margins while less efficient competitors bleed cash or are forced into discounts.

Procurement preference: Enterprises under strict sustainability mandates may choose Google even at a modest premium, because buying "clean compute" reduces their regulatory and reputational risk.

The real economic shift isn't about squeezing more money from customers. It's about who can operate profitably as subsidies vanish, who regulators allow to scale, and which providers win long-term enterprise trust. Efficiency doesn't just cut costs — it determines survival.

Why This Moment Matters

The infrastructure constraints facing AI represent a fundamental shift in competitive dynamics. For the first two decades of the internet era, success came from building better software with existing infrastructure. The AI era requires companies to become infrastructure players, navigating utility politics, permitting processes, and resource allocation decisions that can make or break scaling ambitions.

The companies that recognize this shift and adapt their strategies accordingly will outlast those with superior algorithms but inadequate infrastructure access. Google's environmental transparency initiative demonstrates sophisticated understanding of these changing dynamics, positioning the company for sustainable competitive advantage as infrastructure becomes the primary constraint.

Microsoft's $100 billion in spending and OpenAI's $500 billion in commitments demonstrate the scale of capital required¹³. But capital without infrastructure access is worthless. The companies that secure sustainable partnerships with utilities and regulators will outlast those with bigger checkbooks but weaker stakeholder relationships.

The question isn't whether companies can build better models—it's whether they can secure the massive resources those models need to run.

What do you think? Is Google's environmental transparency genuine corporate responsibility or strategic positioning for the infrastructure wars ahead? Share your thoughts in the comments.

Every great technology wave starts the same way—with raw potential.

But potential alone is not enough.

The early days of the internet? A playground for tech enthusiasts. It only became a revolution when startups figured out how to apply it to real-world problems—e-commerce, search, social media.

The mobile wave? Same story. We didn’t need “mobile apps.” We needed instant messaging, rides on demand, and access to work from anywhere.

And today, AI is following the same pattern.

So where should you, as a founder or product leader, build?

• First, the Big Picture to build a mental model around industry – What exactly are AI agents? How do they differ from good old SaaS or non agentic AI (our chatGPT)?

• Then, the Building Blocks – What do AI agents need to function? Where does the money flow in this new economy?

• Finally, the Human Question (most important one) – What are the problems AI is uniquely positioned to solve? And do we trust it to solve them?

  Thanks for reading Product Boardroom! Subscribe for free to receive new posts and support my work.

What are AI Agents & why do they matter?

The simplest, most human explanation of AI agents I have come across is from AndrewNg (AI Fund, Deep Learning.ai). I paraphrase:

Traditional software—even modern AI chatbots like ChatGPT—functions in a zero-shot fashion. You ask a question, and it gives an answer.
AI agents, by contrast, don’t just answer—they iterate, reflect, and adjust.

Imagine you’re analyzing your monthly business data. If you ask ChatGPT to summarize key insights, it will do so—impressively, even. But is that how you would approach the task?

Likely not. You’d review the data. Think. Compare past trends. Spot anomalies. Write a draft, delete parts, rethink, rewrite.

This is the essence of AI agents—they don’t merely generate an answer; they process, reflect, and evolve.

But this intelligence isn’t magic. It’s built on four foundational patterns:

1. Adaptive execution: AI is super smart but even the smartest person can't get the best output without reflection, iteration. AI agents can adapt, revise, and pivot with reflection or as new information, constraints appear.

2. Tools access: Just as we use TripAdvisor, Skyscanner, and our calendars before booking a trip,AI agents will tap into multiple software tools and connect with external systems (e.g., calendars, financial platforms). Early experiments like AutoGPT and BabyAGI hint at what’s possible. There is a lot to figure out here - how many tools an agent can manage at once (context limit), how they connect (evolving from using human UI to AI optimized interfaces), how permissions work, how they are authenticated.

3. Memory: Most AI models today have fleeting memories, storing only what they need for a single session. Agents, however, can accumulate context over time—learning about your work, projects, and patterns. But this raises bigger questions: How much should AI remember? who decides?. And ofcourse, given so much unstructured data, we will need new databases and other data solutions.

4. Multi-agent collaboration: Human intelligence isn’t singular; it thrives in networks—teams, discussions, debates. AI agents are evolving in the same way, specializing into roles (data analyst, strategist, summarizer) and working together in a multi-agent system.

Where the leverage lies—from Infrastructure to Applications?

Few years ago, even before this GenAI wave, I worked on the Alexa team at Amazon and saw how Amazon essentially “raised” Alexa to be morally and ethically sound—much like raising a child to be a good person, except we wanted her to be a “good” voice assistant.

In AI-speak, that’s the challenging path of training a model from the ground up.

But here’s the thing: Not everyone wants to or can raise a child.

Most people just want a fully-grown expert—a banker, lawyer, or engineer—who already knows what they’re doing.

This analogy carries over to the layers in the AI economy:

1. Infrastructure (Raising the Child) – The foundation models (OpenAI, Google DeepMind, Anthropic), AI chips (NVIDIA, AMD), and compute power (AWS, Azure). This is capital-intensive and dominated by Big Tech.

2. Data & Orchestration (The Bridge to Adulthood) – Model orchestration, Data pipelines, monitoring, and security. In SaaS, billion-dollar companies like Snowflake, Databricks thrived here. In AI, Orchestrators (langchain, crewAI), vector databases (Pinecone, Weaviate) are becoming key players.

3. Applications (The Grown Adult at Work) – Here is where magic gets practical—where we take a “freshly trained generic lawyer” (in AI terms) and hone it to become an extraordinary specialist. SaaS wave gave us Spotify, Uber, Shopify, AI Wave might give us autonomous legal assistants (Harvey AI), digital doctors, digital companions (Character) and more. One key thing to note is that in the Software-as-a-Service wave, we have tools, softwares to solve a problem, perform a task, and do a service. In the AI wave, AI agents will offer the ability to solve that problem, perform that task, perform that service in a piece of software. This will be service-as-software. Aren’t you excited about what the AI wave will bring?

Fall in love with real Problems, not Technology

Technology can dazzle us. But we should resist the temptation to build AI for AI’s sake. AI use is a matter of trust. There are two layers to trust:

• Competence-Based Trust – Which human needs is AI uniquely positioned to solve?

• Relational-Based Trust – Do people trust AI enough to let it solve these needs?

Competence-Based Trust: Can AI agents actually do it better?

AI is typically faster, more cost-effective, and can handle tasks at scale—no question.

But the most important question here is what does better mean here? Broadly it refers to quality.

And quality is multi-faceted: it’s not just about accuracy, but also creativity, ethics, user preference, and context.

So, we have to identify the use cases where AI agents can truly drive value.

Consider work type (physical, thinking, creative, emotional) alongside how standardized what “good” looks like is to map where AI agents could excel or struggle:

• Routine, Repetitive Tasks: Across task types, highly standardized activities with a commonly accepted “right way” of doing things. AI agents excel here; think transactional customer support, straightforward data analytics, or formulaic content generation. Key considerations here are:

  • Scalability: Can your application handle large volumes of repetitive tasks?

  • Simplicity: Basic guardrails often suffice, since the “right” output is well known

  • Time to Value: Is the application intuitive and does it offer relatively quick ROI?

• Framework driven complex tasks: With enough domain training and guardrails, AI can deliver significant benefits in tasks that follow established patterns or best practices but still have room for complexity—think medical diagnosis, self-driving cars. Key considerations here are:

  • Domain Expertise: Do we have enough specific data to sufficiently train AI agents to ensure accurate results?

  • Error Handling: Can the AI detect and gracefully handle edge cases and anomalies rather than assuming all data is standard?

• Highly Complex, Personal, or Uncertain Tasks: This is where human nuance thrives—and where AI still stumbles. Not even humans know exactly how they’ll handle deeply personal or unprecedented scenarios, so teaching an AI to do it is tricky. And if we do give such tasks to AI, there is strong Algorithm aversion which means if an AI agent errs in these realms, people lose trust more quickly than if a human made the same mistake. Key considerations here are:

  • Context Sensitivity: Can AI parse subtle cultural, emotional, or ethical nuances?

  • Self-Correction Ability: Does it know when it's on the wrong track?

  • Algorithmic Fairness: Does the AI reinforce or correct biases from training data?

Relational-Based-Trust: Will people let AI do it?

If something goes wrong—like a self-driving car crash—who’s accountable? The agent? Its maker? Its user? Or no one? It's very important to understand the impact, risk, consequence for the user you are building for. And how it shapes user expectation from the AI agent.

Furthermore in regulated industries, where user confidence alone is not enough—compliance with legal and ethical frameworks is mandatory. In healthcare, for instance, HIPAA in the United States and GDPR provisions in the European Union impose stringent requirements for handling personal data, obliging AI agents to implement encryption, secure logging, and strict access controls. Meanwhile, finance regulators—such as FINRA in the U.S. and emerging directives under the EU’s proposed AI Act—demand transparent, auditable processes, making it essential for AI-driven products (e.g., automated loan approvals) to provide clear justifications for their decisions. These legal standards amplify the need for secure data pipelines, explainability in decision making, fallback mechanisms, and human oversight to ensure that any AI solution remains both trustworthy in the eyes of users and compliant in the eyes of the law

The stakes are incredibly high in certain domains, so trust can’t be an afterthought. In fact, we should regularly ask - Will users realistically trust a machine for this job? And what will it take to increase that trust?

Subscribe now

AI Confidence Map: Automate, Augment or Rethink?

AI will weave itself into every corner of our lives, yet not all paths will lead to meaningful progress. Some solutions will thrive on complete automation, others will demand a human hand on the wheel, and a few will be best shelved—at least until technology and collective intuition catches up. I have put together this AI confidence map as a compass for navigating how deeply we will let machines guide our choices. Ultimately, the measure of success isn’t just innovation for its own sake, but alignment between an AI’s reliable competence and the human trust that breathes life into its potential.

The Golden Zone: AI That Just Works (High Competence, High Trust)

Automated fraud detection, creating marketing copy, reviewing contracts - areas where AI is an obvious winner. Users don’t just tolerate automation here; they actively prefer it.

📌 If you’re here, automate aggressively. AI is the hero.

Hidden Gems: Where AI Needs a Human Touch (High Competence, Low Trust)

Imagine AI-assisted therapy or bedtime stories for kids might be effective with AI but will face Skepticism because the stakes are high or pushback because users love doing these activities.

📌 Be a copilot, not a replacement. Earn trust through transparency where possible.

Blind Faith: AI That’s Not Ready Yet

AI loan reviewers that are biased. AI that screens job candidates but doesn’t understand human potential. This is where reliance on AI might be too early, with big consequences.
📌 Don’t replace humans—assist them. Improve the product until competence catches up.

Dead Ends: AI Should Stay Out of This

AI in High stakes government policy? Creating new art movements? These tasks require nuance, ethics, ingenuity and dealing with human unpredictability that AI will struggle to replicate. Here AI can assist say with a simulation for the impact of a new government policy—but full automation? We will just have to see.
📌 Rethink the problem. AI should inform, but likely not capable of replacing.

The next AI wave isn’t just about tech—it’s about what people will actually use and trust.

Dreams and Nightmares: Ensuring AI really works?

For decades, we’ve built success metrics and guardrails for traditional software, and we still get it wrong sometimes. With AI agents, it’s even harder. Why?

• AI outputs are non-deterministic: meaning the same input can produce different results each time, so standard QA doesn’t always cut it.

• Evaluation criteria change: what’s “good” today might be inadequate or unacceptable tomorrow as user expectations evolve.

• Impact isn’t always immediate: biases, risks, or failures might emerge only after widespread adoption.

Careful, evolving metrics are essential—use case specific, deeply considered, frequently monitored, and continually updated.

The Dream: Headlines we hope to see

"AI-Powered Tutors Provide Free Education to 1 Billion Students Worldwide": AI-driven personalized learning breaks down barriers to access and adapts to each student’s needs.

"AI-Powered Doctors Improve Early Diagnosis by 80% – Millions Gain Access to Better Healthcare": AI reduces diagnostic errors, catching diseases before they become life-threatening.

"AI Sales Agents Boost Revenue by 40%, Automating Lead Generation and Personalized Customer Outreach": AI-powered sales agents identify high-value prospects, craft personalized outreach, and follow up automatically, increasing conversion rates and efficiency.

The Nightmare: Unintended consequences we must mitigate against

"AI Algorithm Denies Millions Access to Loans – Bias in Training Data Exposed": AI trained on historically biased financial data reinforces inequality at scale.

"Deepfake AI Causes Political Crisis – Fake Leader Speech Sparks Global Panic": Disinformation powered by AI-generated deepfakes creates mass confusion. (Seriously, if you haven’t watched The Capture on Netflix, you should.)

"Self-Learning AI Agent Goes Rogue – Automated Trading Wipes Billions Off Stock Market": AI optimizes for profit but disregards systemic risk, triggering a financial crisis.

AI revolution is already here

The AI revolution is happening now. Not in five years. Not next year. Now.

As product leaders, technologists, and thinkers, we must decide: Will we be passive users, or will we be architects of this transformation?

The dream is ours to build.

Subscribe now

4.3 billion people—52% of the world’s population—use some form of digital wallet. This staggering adoption cuts across every generation, and most people reach for their digital wallet more often than traditional payment methods. Only a handful of products match such vast reach—Google Search boasts 4 billion users, Meta’s social apps (like Facebook) serve over 3 billion, and Android devices exceed 3 billion worldwide. or Perhaps it’s more fitting to compare digital wallets to foundational technologies like the internet (5.3+ billion users) and smartphones (6+ billion users).

Whether digital wallets evolve into single-brand products or foundational infrastructure, their impact can be profound, shaping the operating systems of the majority of the planet.

But, Digital wallets need to do more than Simplify

Whether it's buying some bacon or a Burberry, checkout can be painful—not just because you’re about to pay, but because you have to rummage through receipts, multiple cards in a bulky, overstuffed wallet. Digital wallets changed all that and now we pay with a tap. Yet, managing money, identity, and rewards is still a fragmented, disconnected experience.

The missing pieces: where Digital Wallets fall short

• Fragmented Finances: Digitization has rightly fueled competition, leading to innovative fintech solutions. But it’s created a financial puzzle for consumers. Your mortgage might be with HSBC, Forex transfers through Wise. Everyday banking? Monzo. Investments? Vanguard. Each provider operates within its own app, leaving you juggling multiple logins and interfaces.

• Beyond Finance: Even non-financial tasks can spiral into digital scavenger hunts. Need to finalize your life insurance? That might mean fetching your online utility bill, grabbing your physical driver’s license, and tracking down health records in yet another portal.

• Missed Connections: Perks and discounts often get buried in the clutter. Forgot to activate your Amex discount? Left loyalty points unused? C’mon—it’s 2025, we’re talking about Artificial General Intelligence, yet a human still has to manually manage this? Shouldn’t there be a system that connects these dots automatically?

• Accountability Concerns: While digital cards and documents are generally more secure—often relying on tokenization and biometric locks—critical questions remain. Are we simply replacing physical wallets with walled gardens on our smartphones? If one platform stores everything, how safe is your data? And in the event of a security breach—or even just losing your phone—who bears responsibility? Wallet or Individual service?

AI Agents are here: Can you Trust them?

Beyond these gaps, the rise of AI agents presents a new frontier for digital wallets.

For example, OpenAI’s Operator already enables AI-driven shopping, delivery, dining, and travel. Guess what? All of these require user information—age verification, ID proof, and, of course, payments.

Allie K Miller shared from her experience “I like that it punts it back to the user for logins and payments for security… I had several times where the website detected it was an AI and blocked it.”

As AI agents take on more responsibilities, early adopters like Uber, DoorDash, and Instacart are already evaluating how AI agents fit into their terms of service. But for both users and businesses to be comfortable with AI-agents, three major challenges must be addressed:

1. AI agents will need access to more of your life to be useful - access your card or bank account to book that flight, Passport details to check you in, Employment history to fill that loan application. These agents will only be as helpful as the data and permissions they are granted.

2. Users need a way to manage an increasing number of AI agents working on behalf of these users. They will need tools to monitor which agent is doing what, set spending limits or require approval for sensitive actions and revoke permissions or shut down bad actors as needed.

3. Businesses must distinguish Good AI agents from Bad Ones as the risk of malicious agents doing phishing scams, DDoS attacks and more will grow as AI adoption grows.

So, the real question isn’t whether AI agents will manage your money, identity, and assets—but whether you will have the right infrastructure to control them to optimize for you, the user and not the businesses that deployed them? Or someone else? Doesn’t that sound like a Digital Wallet handling your AI-agents?

The Wallet of Tomorrow: Integrated, Intelligent and Secure

Imagine planning a holiday to Greece with your wallet finding the best deals and booking the stay with your budget in mind. At the airport, there’s no fumbling for passports or tickets; your digital wallet with access to your ID, visa generates a secure “travel token” so the airline only sees what they need—no extraneous data. You board the plane with a simple face scan. Upon landing, the wallet instantly converts currency, suggests local tipping etiquette, and handles local expenses seamlessly. Need a dinner reservation? Your wallet handles it effortlessly, ensuring everything fits within your itinerary and budget.

That’s the power of an integrated, intelligent and secure wallet—a tool that unifies critical elements of our life - Finances, Ownership, Access and Identity.

Now, a lot of pieces have to come together to make this vision real but it’s definitely one worth aspiring for. Let’s breakdown each part and see how this might evolve:

1. Effortless Financial Control at Your Fingertips

Journey from fragmented finance to a seamlessly orchestrated single interface - powered by APIs, embedded finance, interoperable systems, and open banking - an interface that can swirl around the concept of managing “your” financial life with “you or on behalf of you”.

• Perfect Payments - Can You Pay Anywhere, Anytime?: Right now, making payments across different platforms, currencies, and methods is still frustratingly fragmented. You shouldn’t have to install Bizum just for a trip to Spain for eg or juggle a dozen apps for different merchants. It all has to come together. That’s why Buy Now, Pay Later (BNPL) providers like Klarna, once standalone services, are integrating directly into wallets like Apple Pay—exactly where they belong. The same is happening with account-to-account (A2A) payments, which today account for 7% of e-com transactions but have yet to gain traction in physical retail. Early adopters like TrueLayer and LinkByStripe are enabling PayByBank, but mainstream adoption depends on wallets making it a default experience. Subsequently, wallets can empower you or your AI agent by selecting the best funding source based on rewards, FX rates, or more making transactions completely frictionless.

• Optimize Opportunity - Can Your Wallet Unlock the Right Financial Resources for You?: As wallets handle an increasing share of your transactions, they’ll evolve into intelligent financial allies. Imagine your wallet recognizing that you’re carrying high-interest debt and surfacing a refinancing offer before you even start searching for one. Instead of users manually applying for credit cards, loans, or mortgages, wallets will proactively bring together microservices to offer the best financial product available when you need it.

• Maximize Wealth - Can Your Wallet Build and Protect Your Wealth?: Future wallets won’t just help you manage money—they’ll help you grow it. They’ll act as autonomous systems, learning your financial patterns, anticipating your needs to optimize how you save, invest, and spend. AI-powered fraud detection will flag weird transactions or block them until you say "okay". This isn’t just about technology; it’s a complete reimagining of how we interact with money, creating a world where financial management is effortless, empowering, and deeply personal.

  

2. Own it, Prove it, Control it - all within your wallet

Our world isn’t just about money—it’s about things we own, the skills we possess, and the access rights we hold. And now, with tech like Non Fungible Tokens (NFTs) and Blockchain, these assets can be digitized, verified, and managed with unprecedented ease. Digital wallets will be at the center of this transformation, empowering users—and their AI agents—to manage and navigate this new world.

• Redefine Ownership - Can you store all that you own?: Digital wallets are evolving beyond just storing credit cards—they can now hold concert tickets, digital keys for your gym or office, and even NFTs (digital collectibles). As trust in blockchain grows, digital wallets will become hubs for all verifiable assets - Digital, Physical, and Intellectual assets. No more hunting for scattered records - your assets live in one secure place, instantly accessible both to you or your AI agent.

• Simplify Proof - Can you prove it Instantly?: Owning something is one thing, proving it seamlessly is another. Need to prove your skills to an employer? Your AI agent, connected to your wallet, could instantly share the relevant skill badge or certification—no PDFs, no paperwork. Thinking of buying or selling a car or real estate? your wallet will instantly share verified details, cutting down time and hassle. And forget about AI agents doing any of this, without a secure, centralized wallet to verify and manage ownership.

• Consolidate Command - Can you Manage and Control It All?: Eventually, your wallet will serve as your universal proof of ownership, qualifications, and access rights, recognized globally. And not just store them but manage who can access, use, or inherit them. Whether it’s transferring ownership of a house, managing digital keys for your office, or planning long-term inheritance strategies, you’ll be able to handle it seamlessly within your wallet—or let your AI agent take over with minimal input.

  

3. The core of it all - the Identity Layer

Identities issued by large corporations (e.g., social media platforms) or governments are already shifting from passwords to biometrics and physical to digital. And new identity frameworks are pushing Identity to be decentralised. Digital wallets are the perfect place to unify a broad array of digitised credentials across sectors and geographies. And as agents join our workforce, identity won’t be limited to humans.

• The Trust Anchor - Are you who You say You are?: With a digital wallet, proving your identity will be instant and private. Selective disclosures will allow users to prove facts without revealing any extra personal information. Renting a car? Your wallet will confirm your age without revealing your full date of birth. Your AI-agent renting a car for you? Digital wallets will issue credentials, similar to OAuth for agents. Merchants can then confidently verify that the agent is your legitimate representative, and accept booking initiated by the agent.

  Jamie smith captured it well - “I believe we’re going to need to evolve KYC even further. We’re going to need ‘KYA’. Know Your Agent.”

• The Eligibility Anchor - Do you Qualify?: From banking to healthcare to education, digital wallets will streamline eligibility checks. Seeking medical treatment? Instead of juggling paperwork and insurance verification, your wallet can securely share pre-approved coverage details, medical history, and policy limits in one click. Plaid’s “Layer” is an early example restricted to financial applications, but wallets will take this further, integrating across sectors for smarter, faster onboarding and processing. This connected approach saves time for users and businesses alike while ensuring compliance and accuracy.

• The Freedom Anchor - Can you take it Everywhere?: Your identity shouldn’t be locked to a single company or government. Future digital IDs will be decentralized—meaning instead of governments or businesses storing your ID, they issue it to you, and you keep it securely in a digital wallet built on blockchain. This makes your identity portable, tamper-proof, and universally accepted. Whether you’re applying for a visa, receiving medical care abroad, or logging in securely online, the same ID could work everywhere. Finally, your identity will work for you—not just for institutions.

  

Digital Wallets: Dreams and Nightmares

As digital wallets evolve, their potential to transform our daily lives is undeniable. But the risks of misuse, security lapses, or overreach are just as significant. Here are some scenarios highlighting the stakes of getting it Right or Wrong.

Headlines we want to see

“AI Agents Revolutionize E-Commerce, with 24/7 Autonomous Transactions”: Digital wallets create trust for businesses and convenience for consumers, enabling seamless, secure AI-driven transactions anytime, anywhere.

“Smart Digital Wallets Drive Financial Empowerment, generating $20 Billion in Financial Gains for its users”: AI-powered digital wallets are automating investments, flagging unnecessary expenses, and optimizing financial decisions, transforming personal finance globally.

“Skill Wallets Revolutionize Hiring, Saving Businesses Millions in Verification Costs”: Instantly verifiable credentials stored in wallets cut hiring timelines and paperwork, unlocking global talent with ease.

"Fraud-Free Future: Wallet Authentication Reduces Risks to Near Zero": Every customer interaction is authenticated via secure, biometric-enabled wallets, slashing fraud and account takeover risks.

Headlines we should fear and mitigate against

"Fraudulent AI Agents Drain Wallet Accounts in Sophisticated Scam": Poor agent verification enables hackers to exploit wallets. The breach compromised not just financial data but identities and assets. Robust cybersecurity frameworks are non-negotiable.

"Wallet Consolidation Raises Privacy and Competition Concerns": Few providers mean higher fees, data misuse, and reduced competition. The line between convenience and surveillance is razor-thin. Users must retain full control to prevent corporate exploitation or government overreach.

"Digital Wallet Blackout Strands Thousands During Natural Disaster": Reliance on centralized servers leaves wallet users unable to access funds or essential services during emergencies. Redundancies and offline capabilities are crucial.

The Battle for your Wallet : Empowerment or Exploitation?

If the timeline of money spans centuries—from coins to paper, paper to plastic—then the digital wallet revolution is only just beginning. Once a simple payment tool, wallets are now evolving into financial hubs, AI-powered identity managers, and gateways to new economic opportunities.

But with transformation comes risk. Will digital wallets empower users, or become walled gardens controlled by a few giants? Will AI agents be trusted allies, or new vulnerabilities waiting to be exploited?

Consumers are still exploring what this technology can do—and how much of their lives they’re willing to entrust to it. Trust will belong to those who offer not just convenience, but true security and user control.

The race is on. Who will own the wallet of tomorrow? And more importantly—who will it serve?