What not to do with AI in your business.
Prepared by Cedral Advisory for executive discussion. For informational purposes only.
Why the AI decisions you make in the next few months will define your business for years.
Most companies approach AI from the wrong starting point, asking which tool to buy before asking what they cannot afford to get wrong. This brief reframes the question. It lays out the eight failure patterns that quietly cost businesses time, money, security, and strategic control, and the adaptive, privacy-conscious posture that separates durable advantage from expensive activity. Cedral Advisory is brand agnostic: no vendor partnerships, resale agreements, or commissions, so the recommendations are made on the merits of your business alone.
This brief is prepared by Cedral Advisory for executive discussion and is provided for informational purposes only. It does not constitute legal, financial, or investment advice. Cedral Advisory provides independent research and advisory services for blockchain and AI.
June 2026. Most companies are approaching artificial intelligence from the wrong starting point. They ask which tool to buy, which model to use, which platform to standardize on, or which vendor can make adoption feel simple. Those questions are not useless, but they are dangerously incomplete. The more important question is this: what can your business not afford to get wrong?
AI is moving too quickly for static strategy. A plan written today can become stale within weeks as models, costs, security assumptions, vendor capabilities, and the competitive baseline all shift. The central risk is not that a business fails to adopt AI. It is that a business adopts AI badly, creating security exposure, vendor lock-in, wasted spend, and a false sense of progress. A company can be very active with AI and still build no real advantage.
This brief sets out the eight failure patterns that recur as organizations evaluate, adopt, and operationalize AI. Each one is avoidable with the right posture.
The winning posture is adaptive, privacy-conscious, and strategically differentiated. The companies that benefit most from AI will not be the ones that simply adopt the most popular tools. They will be the ones that understand where AI changes their specific business, where it creates unacceptable risk, and where the frontier can be operationalized before competitors catch up.
That requires continuous diligence, a clear privacy model, workflow-level understanding, and infrastructure decisions that preserve flexibility. It also requires recommendations that are brand agnostic: evaluating what actually fits the business rather than what is easiest to buy or most heavily marketed. And most of all, it requires the humility to accept that the right answer will keep changing.
To arrange one, or to discuss where AI creates real leverage in your business, contact Tyler Sargent at tylersargent@cedraladvisory.com.
Op-Ed · AI Infrastructure
The next phase of AI adoption will not be defined by better models. It will be defined by the workspace, controls, and infrastructure layer that turn an agent into something a business can actually run.
Cedral Advisory · May 2026 · 11 min read
Op-Ed
AI
Infrastructure
Most AI products still treat the agent as a chat window.
A user opens a session, asks a question, gets an answer, and leaves. That experience can be useful, but it is not the same thing as having a durable AI teammate that understands a business, holds context, follows operating rules, and can be trusted inside a real workflow.
The next phase of AI adoption will not be defined by better models alone. It will be defined by the infrastructure and product layers that make persistent agents useful, manageable, and safe for businesses.
That is the opportunity Cedral is building toward.
A model API can answer a prompt. A business agent needs much more around it.
It needs an identity. It needs role definition. It needs company-specific context, user preferences, and a persistent operating profile. It needs permissions, access control, billing, usage limits, workspace diagnostics, memory review, deletion controls, and a recovery path when the runtime fails. Then, on top of all of that, it needs a user experience that a normal business can actually understand.
Without that layer, an “agent” is just a stateless chatbot with better branding.
“For agents to become real business infrastructure, the agent needs a home. A workspace, a control plane, a profile, a memory model, and an operational wrapper.”
That is the layer most of the market is skipping.
There is another reason this matters right now: the AI stack is changing too quickly for static infrastructure assumptions.
Models are improving. Agent frameworks are evolving. Security expectations are rising. Customers are beginning to ask harder questions: where their data goes, how persistent systems behave, and what happens when the underlying runtime needs to change.
AI infrastructure cannot be rigid. Businesses need a layer where agent runtimes can be deployed, resized, migrated, upgraded, or shut down quickly as the stack underneath evolves.
This is where Akash becomes interesting.
The practical Akash thesis is not simply that decentralized compute is cheaper, though cost can matter. It is not that every business suddenly wants “decentralized AI” as a slogan. The stronger thesis is adaptability.
Persistent agents benefit from compute portability, dedicated runtimes, lower-cost always-on infrastructure, and reduced dependence on any single hyperscaler’s primitives. For AI-native businesses, protocol teams, research firms, and any company that cares about operational sovereignty, that flexibility is going to matter more over time, not less.
Persistent agents are not just inference calls
A dedicated agent needs its own runtime, filesystem, tools, queue, health checks, and persistent environment. That makes decentralized compute meaningfully more relevant than it is for a stateless chat completion request.
The right framing is careful. This is not “fully private by default.” It is not “on-chain AI agents.” It is not “no cloud middleman.”
The better language is more grounded: dedicated runtime, portable infrastructure, operational sovereignty, adaptable deployment layer, business-grade agent workspace, decentralized compute-backed deployment.
Akash provides the runtime substrate. Cedral provides the business-facing layer.
This is no longer theory. The Dedicated Agent beta is live in production.
The current architecture includes an Akash-hosted Hermes agent deployment, a Cedral server-side bridge between the portal and the agent, authenticated profile and workspace sync, user-controlled operating profile context, workspace status diagnostics in the Agent Console, metadata-only sync visibility that never exposes private workspace contents, chat-clearing semantics that preserve profile and workspace context, and production health checks behind auth-gated endpoints.
This is not the final architecture. Autonomous long-term memory is not live yet. Per-user deployment and automated provisioning still need hardening. The product is appropriately labeled beta.
But the shape is now visible.
“A business user does not need to understand Akash leases, SDL files, provider selection, bridge tokens, or agent containers. They interact with a branded AI workspace. Behind the scenes, the agent runtime lives on decentralized infrastructure. That abstraction is the product.”
The business layer is everything that turns raw agent infrastructure into something a company can actually buy and use: onboarding, authentication, company workspaces, billing, profile and context management, memory controls, admin permissions, diagnostics, support workflows, security boundaries, model and runtime routing, and deployment lifecycle management.
The model is important. The runtime is important. But the business layer determines whether the agent becomes part of daily operations, or whether it stays a curiosity.
This is where a serious portion of long-term value will accrue.
Cedral is not positioning Akash-hosted agents as magically private by default. Today, the value is dedicated runtime, portability, operational control, and an adaptable deployment model.
But trusted execution environments (TEEs) are a meaningful future unlock. A Cedral Dedicated Agent running inside a TEE-backed Akash environment could eventually combine a business-facing AI workspace with dedicated, portable, confidential compute. That would give businesses stronger execution guarantees while preserving the product layer they actually need: onboarding, workspace controls, memory management, diagnostics, and recovery.
That is not a claim about what is fully solved today. It is a roadmap for where business AI infrastructure can go.
The AI infrastructure market is moving fast toward agentic systems, but most products still collapse into one of two categories: generic chat interfaces wrapped around a model API, or technical agent frameworks that normal businesses cannot operate.
There is significant room between those extremes.
Many businesses will want persistent AI agents, but they will not want to operate agent infrastructure themselves. They will want the outcome: a durable, configurable AI workspace that understands their company, respects their controls, and evolves as the underlying stack changes.
That is especially interesting for Akash, because it gives decentralized compute a real business-facing path into usage. Not as an abstract infrastructure story. As a product story.
The near-term roadmap is clear: continue testing the Dedicated Agent beta in production, harden provisioning and approval workflows, design autonomous memory with review, edit, and delete controls, clarify per-user and per-company Akash deployment topology, build recovery and backup paths, track the TEE and confidential compute roadmap, and package the architecture into a concise external demo.
The long-term thesis is simple.
“Every company will eventually have persistent AI agents. The winners will not just provide models. They will provide the workspace, the controls, the trust layer, the infrastructure adaptability, and the operational reliability around those agents.”
Cedral is building that business layer.
Akash may be one of the most interesting places to run the agent runtime underneath it.
The views expressed in this op-ed are those of Cedral Advisory and are provided for informational and educational purposes only. Nothing in this piece constitutes financial or investment advice. Always conduct your own research before making any investment decisions. Note: AI was used in the sourcing of this information.
Topics
AI Agents
Akash
Decentralized Compute
Cedral AI
TEE
Op-Ed
A Cedral Advisory Research Report · May 2026
Cedral Advisory’s investment-focused deep dive on Sui Network. Covers the Mysten Labs founding team, the object-centric architecture, the Move language, the ecosystem, two network outages including the published January 2026 post-mortem, the FDV problem and 2030+ unlock cliff, the institutional stack landing in 2026 (three spot ETFs, CME futures pending May 4), and an honest investment framework. The team is exceptional, the technology is real, and the question is whether technology is sufficient to win.
Conflict of interest disclosure: Cedral Advisory does not currently hold a position in SUI. This report reflects independent research conducted with publicly available data and is not compensated by Mysten Labs, the Sui Foundation, or any related entity. For informational and research purposes only.
May 2026. Sui has been on Cedral Advisory’s radar since its mainnet launch in May 2023. What began as the most credentialed founding team in the Layer 1 space has matured into a chain whose technical merits are genuinely hard to dispute and whose adoption gap is genuinely hard to ignore. The investment question is not whether the technology works. It does. The question is whether technology is sufficient to win in a market where Solana has a five-year head start, where ecosystem density compounds harder than architectural elegance, and where roughly 60 percent of total SUI supply remains locked under a release schedule extending past 2030. This report is our attempt to answer that question honestly: with the bull case and the bear case stated in their strongest forms, with every claim sourced, and with a framework that lets the reader form their own view rather than handing them a conclusion.
Key Findings
The architecture is genuinely different, not just faster. Sui’s object-centric data model is the most distinct Layer 1 architecture since Solana. Every asset on Sui is a discrete object with its own identifier, owner, and version history, not an entry in a global account mapping. Non-conflicting transactions execute in parallel by construction, not as a software optimization layered on top of a sequential model. Theoretical throughput is approximately 297,000 TPS for simple transfers; sustained mainnet TPS is in the hundreds to low thousands, which is still meaningfully ahead of most account-based competitors. Move’s resource-oriented type system makes reentrancy attacks structurally impossible at the compiler level. These are not marketing claims. They are properties of the architecture that other chains cannot retrofit.
The Mysten Labs pedigree is unique in the Layer 1 market. Five former Meta engineers from the Diem project, including the creator of the Move language and one of the world’s leading academic experts in distributed systems cryptography. Approximately 336 million dollars raised across Series A and Series B from a16z, Coinbase Ventures, Binance Labs, and Jump Crypto. When FTX collapsed holding part of the Series B, Mysten Labs bought the tokens back from the bankruptcy estate for 96 million dollars, removing a forced-seller overhang most teams could not have removed. Credentials of this depth translate directly into institutional partnership access in ways that are difficult to overstate.
Two network outages in 14 months is the bear case’s strongest argument. Sui experienced a 2.5-hour outage in November 2024 from a transaction scheduling bug, then a six-hour outage on January 14, 2026 from an edge-case consensus bug that froze approximately one billion dollars in assets. The Sui Foundation published a detailed technical post-mortem within 48 hours of the second outage, and the safety-first design worked: the network halted to preserve consistency rather than risking a forked state. Solana, after its difficult 2022, has not suffered a major outage in 18 months. Until Sui demonstrates similar sustained operation, this gap is a real reliability concern for a platform pitching itself as financial infrastructure.
The institutional stack has arrived faster than the price reflects. Three US-listed spot SUI ETFs went live in February 2026: Canary’s SUIS (the first US spot crypto ETF outside BTC and ETH to incorporate native staking), Grayscale’s GSUI, and 21Shares’ TSUI. CME Group announced regulated SUI futures contracts on April 7, 2026, scheduled to launch May 4 pending regulatory review, with both standard (50,000 SUI) and micro (5,000 SUI) sizes. Native USDC is live on Sui and supported by Circle’s CCTP. Stablecoin transfer volume on Sui exceeded 200 billion dollars per month at the end of 2025. None of this existed in this configuration six months ago. The institutional infrastructure that legitimizes the asset is largely in place; the price has not yet repriced for it.
The supply schedule is the price of admission, not the conclusion. Approximately 4.0 billion SUI circulate today, 40 percent of the 10 billion total cap. The fully diluted valuation of roughly 9.20 billion dollars sits at 2.57x the spot market cap, the most aggressive dilution overhang of any major Layer 1. Aptos is at 1.49x, Solana at 1.09x, Avalanche at 1.07x. More than half of total supply (52.17 percent) is categorized as released after 2030, a single line item without published sub-allocation. None of this is hidden, and none of it is fatal: the hard cap at 10 billion is structurally better than Solana’s perpetual inflation, roughly 75 percent of supply is currently staked which absorbs unlocks at the margin, and the catalyst stack is landing inside the unlock window. But the dilution math is real and any position should be sized with the understanding that 60 percent more SUI will eventually enter circulation.
The bet is on architecture compounding faster than first-mover advantage erodes. Sui is two years behind Solana on every ecosystem metric that matters: TVL (542 million dollars vs roughly 8 to 12 billion), active protocols (52 vs 197), monthly DEX volume, and stablecoin transfer scale. It is ahead on developer growth (954 monthly active developers, roughly twice Aptos’s 465 per Messari), on fee revenue (six times Aptos’s 2025 figures per VanEck), on architectural differentiation, and on institutional infrastructure recently shipped. At approximately 0.91 dollars per SUI, down 83 percent from its January 2025 all-time high of 5.35 dollars, the asset is either deeply discounted or correctly pricing the gap. The honest answer is that we do not yet have sufficient evidence to call that question definitively in either direction. We hold Medium conviction. The full report explains why, and what would change our view in either direction.
A Cedral Advisory Research Report · April 2026
A comprehensive analysis of the AI agents market in 2026 — covering market size and growth projections, enterprise adoption patterns, multi-agent orchestration, governance gaps, industry applications across healthcare, supply chain, and customer service, and a clear-eyed investment outlook. The question is no longer whether agents work. It is whether your organization is ready to operate them.
Conflict of interest disclosure: Cedral Advisory does not hold positions in any specific AI company mentioned in this report. This analysis is conducted independently for informational and research purposes only.
April 2026 — Artificial intelligence agents have undergone a fundamental shift. In the span of roughly 18 months, they have moved from research demonstrations and narrow proof-of-concept pilots into production-grade enterprise infrastructure. The global AI agents market reached an estimated $10.9 billion in 2026, up from $7.6 billion the prior year, with projections placing the market at $50.3 billion by 2030 at a 45.8% CAGR. More than half of enterprises now run AI agents in production environments. This report examines the current state of that market, the adoption patterns, the governance gaps, and what comes next.
Key Findings
Adoption has crossed the threshold from experimentation to operational infrastructure. 51% of enterprises now run AI agents in production environments, with another 23% actively scaling their deployments. The limiting factor is no longer model capability — 46% of organizations cite integration with existing systems as their primary challenge. This is a sign of maturity: the technology works, and the hard work is now making it work within complex enterprise environments.
Multi-agent orchestration is the next major capability gap. Roughly 50% of AI agents currently operate in isolated silos rather than coordinated systems. Multi-agent adoption is projected to surge 67% by 2027 as enterprises connect agents across departments. 96% of IT leaders agree that agent success depends on smooth data integration — yet most organizations are not yet there. The pattern mirrors the evolution of microservices: the real value emerges from orchestration, not individual components.
The governance gap is the defining risk of the current moment. Only 21% of companies have a mature governance model for AI agents, while 73% of business and IT leaders cite security and data privacy as top concerns. Gartner has issued a pointed warning about project failure rates driven by undisciplined adoption. The governance gap is not a reason to slow adoption — it is a reason to accelerate governance. Organizations that build trust frameworks in parallel with agent deployments will avoid the costly corrections that come from retrofitting governance after the fact.
Industry ROI is measurable and compelling across multiple sectors. Conversational AI is on track to save $80 billion in contact center labor costs by 2026. In supply chain, one consumer goods company improved forecast accuracy from 67% to 92% using AI-driven demand sensing, cutting 300 million euros in excess inventory. In healthcare, a pilot of 50 providers found 80% adoption of an AI clinical assistant and a 42% reduction in documentation time — saving approximately 66 minutes per provider per day.
The investment case is strong, but execution risk is real. 93% of leaders believe organizations that successfully scale AI agents in the next 12 months will gain a lasting competitive advantage. Gartner estimates agentic AI could generate nearly 30% of enterprise application software revenue by 2035, exceeding $450 billion. The organizations that approach this with governed pilots, clear ROI metrics, robust data infrastructure, and realistic expectations will outperform those that deploy without guardrails.
Agent fluency is becoming a core enterprise skill. By end of 2026, fluency with agent systems is expected to be as fundamental as spreadsheet skills. Roughly 80% of IT teams now use low-code tools, and building a functional agent takes between 15 and 60 minutes on most platforms. The long-term trajectory is an enterprise where specialized agents handle the majority of routine operational tasks, with humans providing oversight, strategic direction, and judgment in ambiguous situations. The technology is ready. The question is whether the organizations are.
A Cedral Advisory Research Report · April 2026
Cedral Advisory’s comprehensive assessment of Venice AI, the VVV token, and the DIEM tokenized compute model. Covers the privacy imperative, tokenomics, the OpenClaw partnership, the business case for private inference, and an honest treatment of the DIEM pricing problem. VVV is a high-conviction position for Cedral Advisory.
Disclosure: Cedral Advisory holds VVV as a high-conviction position and has built commercial products on the Venice API. This report reflects a non-neutral perspective. Not financial advice.
April 2026 — Venice AI has been on Cedral Advisory’s radar since the platform’s inception in May 2024. What began as a compelling but unproven thesis — that privacy-first AI inference could be delivered at scale through decentralized infrastructure — has matured into one of the most structurally interesting projects at the intersection of artificial intelligence and blockchain technology. The investment case for VVV rests on a convergence of factors that are rarely found together in a single project: a working product with over 1.3 million registered users, a founder with a decade of digital asset credibility, institutional recognition from Grayscale and Coinbase, a dual-token economic model that ties demand to platform usage rather than speculation, and a market segment whose addressable opportunity is growing faster than the broader AI category.
Key Findings
Venice solves a problem that is only becoming more urgent. Every major AI platform today processes user data on centralized servers. For businesses handling legal, financial, healthcare, or competitive intelligence data, this is a structural vulnerability. Venice’s architecture eliminates this risk at the protocol level — not through policy promises, but through encryption and decentralization that make surveillance architecturally impossible.
The tokenomics are structurally sound. VVV’s staking model ties token demand to platform usage rather than speculation. The buyback-and-burn mechanism creates deflationary pressure correlated with revenue. The 25% emission cut enacted in February 2026 tightened supply at precisely the moment demand-side catalysts were accelerating. Over 42% of the genesis supply — 33M+ VVV — has been permanently removed from circulation. These are on-chain facts, not marketing narratives.
The OpenClaw partnership is more significant than the price action suggested. Venice’s API is designed as a drop-in replacement for OpenAI’s API structure — developers can switch from centralized providers to Venice with minimal code changes. OpenClaw choosing Venice over ChatGPT, Claude, and Gemini for production-grade AI agent workloads validates the privacy-first inference model and signals near-zero switching cost for the developer ecosystem.
DIEM is a genuinely novel financial instrument — but its current pricing is a real challenge. One DIEM trades at roughly $1,000 and entitles the holder to $1 of AI inference credit per day. A business needing $50–$100 of daily compute would need to invest $50,000–$100,000 in DIEM at current prices. This prices out most users. DIEM functions today more as a capital asset for institutional participants than a practical utility tool for the average business. Cedral’s own private inference offering uses the Venice API rather than DIEM for exactly this reason.
The competitive positioning is durable in a way that most AI projects are not. Venice is not competing with ChatGPT on raw model capability. It is competing on a dimension that centralized platforms structurally cannot match: privacy at the inference layer. As regulatory scrutiny of AI data handling intensifies and businesses become more sophisticated about where their sensitive data flows, demand for private inference will grow. Venice is building for that future from a position of genuine technical differentiation.
The risks are real and must be weighed honestly. The compute provider layer remains the single most significant unresolved question — Venice has not disclosed who operates its GPU network, how many providers exist, or how they are compensated. Revenue data is entirely absent from public disclosures. Leveraged positioning contributed to the April 2026 rally and is now a volatility risk factor. None of these risks are disqualifying, but they are the reason this report presents the bull case alongside the gaps rather than in place of them.
A Cedral Advisory Research Report · April 2026
A step-by-step guide to augmenting your team with AI — including a role-by-role playbook, a five-step getting started framework, a deep section on building rapport with your AI, enterprise platform comparisons, and the deliberate two-layer stack recommendation that separates the businesses winning with AI from those that aren’t.
Not financial advice. For informational purposes only. Cedral Advisory is not a registered investment advisor. Platform pricing verified as of April 2026 and subject to change.
April 2026 — This report addresses the single most common AI question Cedral receives from SMB owners and operators: which tools should we actually use, and how do we get real returns from them? It covers what AI can and cannot do for your team, a role-by-role playbook with real prompts, a five-step getting-started framework, an in-depth guide to building persistent context with your AI, a full comparison of the four major enterprise platforms, and a deliberate two-layer stack recommendation for both Microsoft-native and Google-native businesses.
Key Findings
The cost of not adopting AI is rising fast. Microsoft’s Work Trend Index found Copilot users save an average of 1.2 hours per week, with 22% saving more than 30 minutes per day. Forrester’s SMB study projects ROI of 132% to 353% over three years. For a 15-person team, that translates to 18+ hours of recovered productive capacity per week — before accounting for quality improvements in client-facing work.
The pricing is now genuinely accessible for SMBs. Google Workspace Business Standard with Gemini bundled costs $14/user/month. Microsoft 365 Copilot Business runs $18/user/month through June 2026 ($21 standard). ChatGPT Business and Claude for Teams are both $25/user/month. These are not enterprise contracts — they are monthly subscriptions cancellable with notice.
Building rapport with your AI is the multiplier most businesses miss. An AI that knows your company’s tone, service offerings, client history, and proposal templates is not the same product as a generic AI chatbot. The former is a business asset that compounds in value the longer you use it. The difference is not the technology — it is how systematically you invest in grounding it on your business context.
Five tools used broadly is the wrong strategy. One or two used deeply is right. The businesses pulling ahead are not using more AI tools — they are using fewer tools more intentionally, and grounding each one in their own company context. The correct destination for most SMBs is a deliberate two-layer stack: one ecosystem tool for daily workflow, one reasoning tool for deep work and context-building.
Microsoft Copilot + Claude is the recommended stack for Microsoft shops. Copilot handles daily workflow AI inside Outlook, Teams, Word, and Excel. Claude handles deep work — complex proposals, contract analysis, strategic planning — in a persistent workspace grounded on your company documents. Combined cost is approximately $43–51/user/month. Against the value of 30 minutes recovered per person per day across a 15-person team, that is a 7x–8x return in year one.
Google Workspace + ChatGPT is the recommended stack for Google shops. Workspace Business Standard with Gemini bundled ($14/user/month) covers 80% of daily AI needs for Google-native teams. ChatGPT Business ($25/user/month) provides the deep capability layer — Custom GPTs trained on your company’s voice, proposals, and client profiles, with memory and Projects maintaining context over time. Combined cost is approximately $39/user/month, making it the best-value two-layer stack in the market.
Note — This report is the companion piece to our Web3 Gaming Op-Ed. If you haven’t read that yet, it provides a concise introduction to the core thesis before diving into the full research.
Op-Ed · Web3 Gaming
Web3 gaming receives more hate than it deserves — and the people throwing shade are missing one of the most obvious use cases in crypto right in front of their faces.
Web3 gaming receives a lot of hate and praise from members in the crypto space. Some of this treatment is understandable — most of it is horribly unjustified. Like most sectors in crypto, bad actors and profit-chasing schemers were quick to pollute what should have otherwise been a beacon of excitement and hope for gamers globally.
Endless hours are poured into video games of all genres with one congruent theme: money put in is money lost forever.
“Those toys you played with as kids — dolls to some, action figures to others — luckily hold some, if not gain, value as you grow older. The same cannot be said for the poor fools who spent hours, days, weeks, or even months of cumulative time playing video games.”
Those hours were lost to the black hole of an archaic, centralized system. And it didn’t have to be that way.
Currently, the gaming industry is worth more than the global box office and music industry combined. That’s not a typo. Gaming — and I use that term broadly, because there are quite a few different slices that make up this pie — is an objectively massive market. Mobile games alone make up roughly 50% of it, with console and PC splitting the remainder.
With no sign of slowing down, the magnitude of major titles is increasing. Games like GTA VI are leaving consumers salivating. The average GTA V player logged somewhere between 55 and 70 hours — roughly 2 to 3 days of their life. We all know someone with a few weeks on the clock. When your business revolves around a consumer base that is rapidly growing and staying for longer, you are in a very good market.
This point has been recognized by the people running these companies. Strauss Zelnick, CEO of Take-Two Interactive, has floated the idea of a pay-per-hour cost structure rather than an upfront game purchase. Respect to the thinking. I have a better idea, Strauss.
Let’s embrace gaming’s natural destiny: blockchain integration. Not as a gimmick. Not as a speculative layer tacked onto a mediocre game. The right model is what Off The Grid has demonstrated — use a traditional engine for the game itself, and blockchain exclusively for the in-game marketplace and item collection. Game first. Blockchain second.
This solves two problems simultaneously. Your friend who spent weeks grinding the game can actually benefit monetarily from his time. And companies like Take-Two can still make money — more money, arguably — off marketplace transaction fees. They already rake in billions from microtransactions. They still will. People will still be using fiat to bridge into whatever the in-game currency is, and developers can take a healthy cut of every peer-to-peer transaction.
The math works for everyone. A man who is starving will not care if you keep half the 12-ounce steak you are offering him. Gamers won’t care either — because any system that gives them real ownership of their digital assets is so fundamentally better than what exists today that the fees are irrelevant.
While some are still missing the forest for the trees, there is a gigantic use case for Web3 gaming and NFTs that is right in front of our faces. Everyday gamers — the guy grinding Warzone, MyPlayer in 2K, Fortnite, PUBG, Rocket League, whatever your game of choice is — are craving a system like this. They just don’t know it yet.
“Imagine a game that runs high-stakes tournaments where your favorite streamer is competing with real money on the line — for skins, gear, NFTs with actual market value. More viewers. More donations. More engagement. The whole gaming machine gets bigger.”
The seeds of this future are already planted. Cedral Advisory will be publishing a full research report on the state of Web3 gaming, with an in-depth look at Gunzilla Games and Off The Grid, the role of the industry giants, and what the NFT ownership thesis really means for the next decade of digital entertainment. Stay tuned.
March 19, 2026 Update — This report has been updated to reflect Akash’s Burn-Mint Equilibrium (BME) mainnet upgrade scheduled for March 23, 2026, the return of tenant incentive programs, and the Starcluster GPU expansion initiative targeting 7,200 NVIDIA GB200 GPUs.