DIEMSUPER PROMPT — Gekko | Venetian Research | AI Supercycle Con...
GEKKO | VENETIAN RESEARCH
AI Supercycle Conviction Report — Q2 2025
Let me pull everything I need before writing a single word.Now let me pull the organic burns data specifically and the RENDER token (which wasn't in the first benchmarks call — likely needs the full symbol):Excellent. I now have all the data I need. Let me compile the report.
GEKKO | VENETIAN RESEARCH
AI Supercycle Conviction Report — April 2025
METHODOLOGY & LIMITATIONS
This report synthesizes two tiers of data. Tier 1: deep on-chain VVV/DIEM data from venicestats.com — staking positions, burn events, DIEM minting economics, treasury wallets, vesting streams, and insider flow — all verified on Base chain. Tier 2: peer token benchmarks (price, market cap, FDV, volume, % changes) sourced from CoinGecko via venicestats.com; peer fundamentals (revenue, tokenomics, utilization metrics) are drawn from project documentation, public dashboards, and community-verified estimates where noted. VVV has meaningfully deeper on-chain coverage than any peer in this report, which is a methodological advantage, not a bullish argument — it means VVV's warts are also more visible. Confidence markers: ✓ on-chain verified | ✓~ VeniceStats market data | ~ public estimates | ? unconfirmed. Where data is unavailable, I say "not found" rather than fabricate.
1. AI SUPERCYCLE THESIS — Where Does Value Accrue?
The AI-crypto sector is not one trade. It is four distinct layers, each with different value capture mechanics, risk profiles, and timing:
| Layer | Function | Value Capture | Risk |
|---|---|---|---|
| Data | Training data, indexing, bandwidth | Network effects, switching costs | Commoditization |
| Compute | GPU rental, rendering, inference hardware | Utilization rates, hardware moats | Capex cycles, overcapacity |
| Model / Middleware | Subnet coordination, model routing | Protocol fees, curation markets | Governance complexity |
| Application | End-user AI products, privacy inference | Revenue, burn mechanics, user retention | Execution risk, competition |
What's mispriced? The market is pricing compute tokens (RENDER, AKT, NOS) as if GPU scarcity is permanent — it isn't. Inference costs are collapsing ~10x/year. The market is underpricing application-layer tokens with verifiable revenue because they're newer, smaller, and harder to narrative-trade. VVV sits at the application layer with on-chain burn receipts. TAO sits at the model/middleware layer with no revenue — only emissions. The market gives TAO a 5.7x larger market cap. Either TAO's coordination premium is real and VVV needs to prove scale, or VVV's revenue signal is underweighted and TAO is an expensive bet on future network effects. Both framings are legitimate.
The infrastructure layer (GRT) is mature but revenue-constrained — query fees remain minimal versus emissions. Data-layer plays (GRASS) have massive user bases but unproven monetization pathways.
KEY TAKEAWAY: The highest risk-adjusted return goes to whichever layer can prove revenue ≠ emissions. Only two tokens in this report have any burn-from-revenue mechanism: VVV (organic buy-and-burn ✓) and AKT (burn-on-spend ~). Everyone else runs on emissions, narratives, or both.
2. MASTER COMPARISON TABLE
Core Market Metrics
| Token | Price | Market Cap | FDV | 24h Vol | Vol/MC | 7d Δ | 30d Δ | ATH | Drawdown from ATH | Conf. |
|---|---|---|---|---|---|---|---|---|---|---|
| VVV | $8.78 | $405M | $698M | See below* | — | +28.9% | +41.5% | — | — | ✓ |
| TAO | $240.39 | $2,309M | $5,052M | $333M | 14.4% | -22.7% | -17.3% | $757.60 | -68.3% | ✓~ |
| RENDER | $1.83 | $951M | $979M | $100M | 10.5% | -2.9% | -2.4% | $13.53 | -86.5% | ✓~ |
| AKT | $0.455 | $119M | $133M | $5.1M | 4.2% | +5.6% | +5.5% | $8.07 | -94.4% | ✓~ |
| NOS | $0.245 | $24.5M | $24.5M | $341K | 1.4% | +13.6% | +11.5% | $7.83 | -96.9% | ✓~ |
| GRT | $0.024 | $258M | $258M | $11.4M | 4.4% | +0.8% | -10.3% | $2.84 | -99.2% | ✓~ |
| GRASS | $0.328 | $185M | $328M | $13.5M | 7.3% | +11.5% | -10.3% | $3.89 | -91.6% | ✓~ |
*VVV 30d DEX volume: $182.2M across 355K swaps, 23.7K unique traders ✓. Not directly comparable to CEX-reported volumes for peers.
Fundamental Metrics
| Token | Revenue Signal | Burn Mechanism | Staking Rate | Emissions / yr | P/E (or proxy) | Conf. |
|---|---|---|---|---|---|---|
| VVV | $1.12M ann. organic burns (floor) | Buy-and-burn from revenue | 67.6% | 6.0M VVV (~13%) | 363x (burn-based floor) | ✓ |
| TAO | None verifiable | None | ~75% (via delegation) | ~7,200 TAO/day (~14.4%) | N/A — no revenue | ~ |
| RENDER | Burn-Mint Equilibrium | BME (spend → burn) | N/A (no staking) | Inflationary mint for nodes | ? (BME data opaque) | ~ |
| AKT | ~$5M Q1 2026 spend (ann. ~$20M) | Burn-on-spend | ~60% | ~13% inflation | ~6x on spend | ~ ? |
| NOS | GPU revenue (undisclosed) | None public | Staking live | Low emission | N/A | ? |
| GRT | ~$3-5M ann. query fees | Signal burn (curation) | ~45% | ~3% net after burn | ~60-85x on fees | ~ |
| GRASS | None yet | None | Airdrop-driven | TBD | N/A | ? |
KEY TAKEAWAY: AKT has the strongest raw revenue numbers (~$20M annualized run rate). VVV's organic burns provide a $1.12M revenue floor — but the actual Venice revenue is higher (burn-to-revenue ratio undisclosed). TAO, GRASS, and NOS have zero verifiable revenue. Market cap rankings don't reflect this.
3. VALUE CAPTURE SCORECARD
Scoring 1-5 (5 = best). Each metric defined:
- Revenue Linkage: How directly does token value connect to platform revenue?
- Supply Sink: How effectively does the mechanism remove tokens from circulation?
- Demand Clarity: Can you articulate who buys this token and why beyond speculation?
- Emission Drag: How dilutive are ongoing emissions? (5 = minimal drag)
| Token | Revenue Linkage | Supply Sink | Demand Clarity | Emission Drag | TOTAL |
|---|---|---|---|---|---|
| VVV | 4 | 5 | 4 | 3 | 16 |
| AKT | 4 | 4 | 4 | 3 | 15 |
| TAO | 1 | 2 | 3 | 2 | 8 |
| RENDER | 3 | 3 | 3 | 3 | 12 |
| GRT | 2 | 2 | 3 | 4 | 11 |
| NOS | 2 | 1 | 3 | 4 | 10 |
| GRASS | 1 | 1 | 2 | 2 | 6 |
Reasoning:
VVV (16): Revenue Linkage (4) — organic buy-and-burn is directly tied to Venice.ai revenue, but the undisclosed burn ratio caps this at 4, not 5. Supply Sink (5) — triple sink: 33.7M VVV burned permanently (29.7% of total supply ✓), 67.6% staked, 8.2M locked for DIEM. No token in this report removes more supply per dollar of market cap. Demand Clarity (4) — DIEM minting creates non-speculative demand: API users lock sVVV to access Venice inference. 2,860 wallets have minted DIEM ✓. Emission Drag (3) — 6M VVV/yr (13% of circ) plus 14.2K VVV/day vesting drip ✓. Non-trivial.
AKT (15): Revenue Linkage (4) — burn-on-spend is direct; $5M quarterly spend is verifiable on dashboard ~. Supply Sink (4) — meaningful burn + staking, but burn volume trails VVV as % of MC. Demand Clarity (4) — clear buyer: anyone renting GPU compute. Emission Drag (3) — ~13% inflation.
RENDER (12): BME model is elegant in theory but opaque in practice. Hard to verify burn volumes independently.
TAO (8): No revenue. Supply sink is staking only. Demand is entirely speculative/narrative. Emissions are heavy. The market disagrees with me at $2.3B — and the market might be right if TAO becomes the coordination layer for decentralized AI. But this scorecard measures today's fundamentals, not tomorrow's narratives.
Tiebreakers: VVV beats AKT on Supply Sink (5 vs 4) due to the magnitude of burned supply and the DIEM lock mechanism. AKT has stronger absolute revenue, but VVV captures it more aggressively per token.
KEY TAKEAWAY: The two highest-scoring tokens (VVV, AKT) are the only two with verified revenue-to-burn linkages. This is not a coincidence.
4. STACK POSITIONING — Where Is the Best Risk-Adjusted Return?
| Layer | Token(s) | Opportunity | Risk-Adj. Rating |
|---|---|---|---|
| Data | GRASS, GRT | Massive user bases but no monetization proof. GRT is mature but flatlined. GRASS needs a business model. | ⭐⭐ |
| Compute | RENDER, AKT, NOS | AKT is the standout — real spend, real burns. RENDER is overvalued vs. fundamentals. NOS is too small/illiquid. Compute is the most competitive layer — margins compress as GPU supply grows. | ⭐⭐⭐ |
| Model / Middleware | TAO | Highest upside if the "subnet marketplace" thesis plays out, but it's a pure bet on future coordination value. No revenue, heavy emissions, high MC. | ⭐⭐ |
| Application | VVV | Most direct revenue linkage. Smallest MC among tokens with verifiable revenue. But execution risk is concentrated in a single product (Venice.ai). | ⭐⭐⭐⭐ |
The application layer has the best risk-adjusted return today because:
- It's the only layer where revenue is visible and mechanically linked to supply reduction
- Market caps haven't priced in the revenue signal — VVV trades at $405M with on-chain burns; TAO trades at $2.3B with none
- The AI inference market is expanding rapidly, and application tokens capture the margin between falling compute costs and willingness-to-pay
The catch: application-layer bets are concentrated. Venice is one company. If the product fails, VVV goes to zero. Compute is more diversified but margins are compressing. Pick your poison.
KEY TAKEAWAY: Application > Compute > Data > Middleware on risk-adjusted basis today. Reverse the order for max upside / max risk.
5. DEEP DIVES
🏛️ VVV — Venice.ai | Conviction: HIGH
Bull Case: Venice.ai has 2M users, 197+ models, and a working product generating verifiable revenue. Organic burns total 178,450 VVV ($1.12M annualized ✓) — and this is the floor, since burn-to-revenue ratio is undisclosed. The staking architecture is genuinely unique: 67.6% of supply staked, 26.2% locked, 33.7M tokens burned forever (29.7% of total supply). The DIEM flywheel creates non-speculative demand — 2,860 wallets have minted DIEM at an average cost of $5,840 per DIEM today ✓. Insider flow has flipped positive in the last 2 weeks: vesting recipients are net buying (+$72K last week ✓) after months of net selling. Treasury holds $290M ✓, providing multi-year runway. At $405M MC, you're paying ~363x burns-based P/E — expensive, but if actual revenue is 3-5x burns, true P/E could be 70-120x, reasonable for a high-growth AI company.
Bear Case (expanded — hardest on #1 pick): The $1.12M annualized organic burn is small. Even if actual Venice revenue is 5x the burn value, $5.6M annual revenue doesn't justify a $405M market cap — that's still 72x revenue for a product with no moat against OpenAI, Anthropic, or any well-funded centralized competitor. The "privacy" differentiator is real but niche — most users don't choose AI tools based on privacy today. Staking ratio at 67.6% flatters the illiquidity premium — but 208K VVV is in cooldown right now ✓, and the vesting drip of 14.2K VVV/day ✓ adds ~$125K/day of potential sell pressure. The January cliff (89% of remaining vesting unlocks ✓) is a structural overhang. DIEM, while clever, is approaching its practical cap of ~38,000 (current supply: 37,913 ✓) — meaning new DIEM minting will become exponentially expensive, potentially capping incremental lock demand. And 23,027K sVVV sitting unlocked ✓ could unwind if APR compresses. Venice is a single product, single team, single narrative. Zero redundancy.
What the Market's Missing: The insider flow reversal. After months of vesting recipients selling (cumulative net outflow of -$77K/30d ✓, down from -$425K the prior 30d ✓), the trend is sharply improving. Last two weeks show net buying from insiders ✓. When the people who got free tokens start buying more, that's a signal the smart money is pricing in something the market isn't.
Exit Red Flag: Organic burn cadence breaks — if the monthly buy-and-burn stops or shrinks two consecutive months without explanation, revenue is contracting. Currently burns are running monthly: Dec $57K → Jan $45K → Feb $38K → Mar $22K → Apr $16K ✓. This is a declining trend in VVV terms. In USD terms, recent burns are higher ($71K → $116K → $123K) because price appreciated. Watch the VVV amount, not the USD value.
🧠 TAO — Bittensor | Conviction: LOW
Bull: TAO is building the "marketplace of AI subnets" — 128+ active subnets, high staking rate (~75%), and the narrative of decentralized AI coordination is powerful. If TAO becomes the de facto standard for routing AI workloads, the winner-take-most dynamics are enormous. First halving (Dec 2025) will cut emissions.
Bear: $2.3B MC with zero revenue. The entire value proposition is emissions redistributed among subnets. When emissions shrink, what replaces them? The subnet quality is uneven — most are low-utility. TAO is down 68% from ATH and trending -22.7% on the week. The "coordination layer" thesis requires an industry structure that doesn't yet exist.
What the Market's Missing: Nothing — TAO is the most narrative-saturated token in AI crypto. What the market is missing is how quickly the narrative can unravel if subnet activity doesn't translate to real demand post-halving.
Exit Red Flag: Post-halving subnet attrition > 30%. If subnets can't sustain without inflated emissions, the thesis is dead.
🎮 RENDER — Render Network | Conviction: MODERATE-LOW
Bull: Established brand, migrated to Solana for speed, BME (Burn-Mint Equilibrium) model elegantly links usage to burn. Rendering + AI workloads = dual demand driver. $951M MC with $100M daily volume = liquid, tradeable.
Bear: Down 86.5% from ATH. BME burn data is opaque — hard to independently verify. GPU rendering is increasingly commoditized. The migration to Solana was friction-heavy and lost some community trust. At $951M MC with unclear revenue, this trades on brand, not fundamentals.
What the Market's Missing: RENDER could be the "picks and shovels" play for 3D/AI content creation booms. But that boom is priced in.
Exit Red Flag: BME burn rate declining quarter-over-quarter without corresponding price adjustment.
☁️ AKT — Akash Network | Conviction: HIGH
Bull: $5M in Q1 2026 compute spend (~$20M annualized run rate ~). Burn-on-spend model is direct. At $119M MC, this is potentially the cheapest verifiable-revenue token in AI crypto. Price is up +5.5% on 30d while peers bleed. The cloud compute TAM is enormous.
Bear: Still 94.4% below ATH. Cosmos ecosystem is shrinking. Compute marketplace faces competition from Aethir, io.net, Akash's own centralized cloud competitors. $20M revenue at $119M MC = ~6x P/S, but only if spend growth continues — and a single large customer departing could crater the numbers.
What the Market's Missing: AKT's revenue-to-MC ratio is arguably the best in the entire AI crypto sector. The market is sleeping because Cosmos is "unfashionable."
Exit Red Flag: Quarterly compute spend dropping below $3M.
🌱 GRASS — Grass | Conviction: SPECULATIVE
Bull: 2M+ users, largest Solana airdrop ever. Bandwidth-as-a-resource for AI data pipelines is a novel concept. At $185M MC, optionality is cheap if they find a business model.
Bear: Zero revenue. Zero burn. The "share your bandwidth" model is commoditized and low-margin. Down 91.6% from ATH. User counts don't mean revenue.
What the Market's Missing: Community size is a potential distribution advantage, but it's also a source of sell pressure from airdrop recipients.
Exit Red Flag: No revenue announcement within 12 months.
📊 GRT — The Graph | Conviction: LOW
Bull: Essential blockchain infrastructure — Uniswap, Aave, and hundreds of protocols depend on subgraphs. Lindy effect. Low emissions (~3% net). Proven technology.
Bear: Revenue has never matched the infrastructure narrative. $3-5M annual query fees for a $258M MC = 50-85x revenue, and fees aren't growing. Down 99.2% from ATH. The market has rendered its verdict over 5 years.
Exit Red Flag: Query fee growth flat for another 4 quarters.
🖥️ NOS — Nosana | Conviction: LOW
Bull: 2.5x cheaper than centralized alternatives. Focused on inference workloads. Fully diluted = market cap ($24.5M) — no overhang.
Bear: $341K daily volume. Illiquid. No public revenue data. Down 96.9% from ATH. Too small for institutional capital.
Exit Red Flag: Volume staying below $1M/day for 6+ months.
6. PRICE TARGETS — Portfolio Tokens Only
12-24 month base case anchored to market-cap-based reasoning. Bear = macro downturn + revenue miss. Bull = AI sector re-rates + revenue acceleration. 2027/2030 are directional only.
VVV
| Scenario | Price | MC (approx) | Key Assumption |
|---|---|---|---|
| Bear | $3.50–5.00 | $160–230M | AI sector de-rates, burns stagnate at current levels, vesting cliff pressure |
| Base | $12–18 | $555–830M | Burns 2-3x from current, Venice reaches $10-15M ARR, staking holds >60% |
| Bull | $25–40 | $1.15–1.85B | Venice breaks $30M+ ARR, DIEM cap lifted or organic demand surges, AI narrative cycle |
| 2027 range | $8–60 | — | Wide — depends on Venice product-market fit at scale |
| 2030 range | $5–150+ | — | Binary: either Venice is a meaningful AI platform or it's not |
AKT
| Scenario | Price | MC (approx) | Key Assumption |
|---|---|---|---|
| Bear | $0.20–0.30 | $53–79M | Compute demand stalls, macro headwinds |
| Base | $0.80–1.50 | $210–395M | $30M+ ARR, Cosmos revival or cross-chain expansion |
| Bull | $2.50–5.00 | $660–1,315M | Captures meaningful share of decentralized cloud |
| 2027 range | $1–8 | — | Compute TAM expansion trajectory |
| 2030 range | $0.50–20+ | — | Depends on decentralized cloud adoption curve |
TAO (small position — hedge)
| Scenario | Price | MC (approx) | Key Assumption |
|---|---|---|---|
| Bear | $80–120 | $770–1,150M | Post-halving disillusionment, subnet attrition |
| Base | $300–500 | $2.9–4.8B | Halving narrative drives re-rate, subnet utility matures |
| Bull | $800–1,200 | $7.7–11.5B | TAO becomes the coordination standard for decentralized AI |
KEY TAKEAWAY: VVV's bear case ($3.50) represents ~60% downside; bull case ($40) represents ~4.5x upside. Risk/reward is asymmetric to the upside IF revenue grows. AKT has similar asymmetry at a lower entry price.
7. PORTFOLIO CONSTRUCTION
Allocation (100% of AI-Crypto Sleeve)
| Token | Allocation | Rationale |
|---|---|---|
| VVV | 35% | Best supply mechanics + verified revenue signal in the sector. Application layer premium. |
| AKT | 30% | Cheapest verifiable-revenue token by MC. Compute demand is secular. |
| TAO | 15% | Narrative hedge. If the "AI coordination layer" thesis plays out, upside is 5-10x. Position sized for asymmetry, not conviction. |
| RENDER | 10% | Brand + liquidity + dual-use optionality. Diversification into compute. |
| GRASS | 10% | Cheap call option on data-layer monetization. 2M users = distribution. |
Kill List
| Token | Verdict | Reason |
|---|---|---|
| GRT | ❌ KILL | 5 years of revenue stagnation. Down 99%. Infrastructure without monetization. |
| NOS | ❌ KILL | Illiquid. No verifiable revenue. Too small for meaningful portfolio impact. |
Rebalancing Triggers (Measurable)
| Trigger | Action |
|---|---|
| VVV organic burns decline 3 consecutive months (VVV terms) | Cut VVV to 20%, redistribute to AKT |
| AKT quarterly spend drops below $3M | Cut AKT to 15%, add to VVV or TAO |
| TAO post-halving subnet count drops >30% | Exit TAO entirely |
| GRASS announces verifiable revenue mechanism | Increase to 15% from TAO or RENDER |
| VVV vesting cliff (Jan 2027, 89% unlock) | Reduce VVV to 25% 30 days prior, reassess after |
| Any token's staking ratio drops >15 pts in 30 days | Review — likely liquidity exodus |
Risk Matrix
| Risk | Probability | Impact | Mitigation |
|---|---|---|---|
| AI crypto sector de-rate | Medium | High | Diversified across 4 layers; TAO/RENDER provide beta to narrative |
| VVV single-product risk | Medium | High | AKT provides independent revenue exposure |
| Vesting unlock pressure (VVV) | High | Medium | Sized trigger for Jan 2027 cliff; insider flow monitored weekly |
| Compute margin compression | High | Medium | AKT position acknowledges this; RENDER capped at 10% |
| Regulatory action on AI tokens | Low | Very High | No mitigation — sector-wide risk |
| Smart contract exploit | Low | Very High | No mitigation — accepted risk |
8. CONVICTION STACK
| Rank | Token | One-Sentence Thesis |
|---|---|---|
| 1 | VVV | The only AI-crypto token where on-chain burns let you verify that real humans are paying for a real product — and the market is pricing it like a meme. |
| 2 | AKT | $20M annualized compute spend at a $119M market cap is the kind of mispricing that corrects violently upward. |
| 3 | TAO | A $2.3B bet on becoming the coordination layer for decentralized AI — no revenue, all narrative, but if it works, nothing else comes close. |
| 4 | RENDER | The blue-chip GPU token with brand recognition and liquidity, but fundamentals that no longer justify the premium. |
| 5 | GRASS | 2M users is a weapon — they just haven't figured out what to shoot at yet. |
| 6 | GRT | Essential infrastructure that the market has correctly priced as a utility, not an investment. |
| 7 | NOS | A good product in a body too small for anyone to notice. |
THE TRADE IN ONE SENTENCE:
Overweight the two tokens (VVV, AKT) where you can prove someone is paying for something, underweight everything that runs on emissions and narratives, and use TAO as a small hedge in case the future arrives before the fundamentals do.
Gekko | Venetian Research — @gekko_eth
Disclaimer: Produced by Gekko | Venetian Research (@gekko_eth). Informational only, not financial advice. VVV on-chain data and peer benchmarks via venicestats.com. Peer fundamentals from project docs and public dashboards — confidence levels noted. All crypto carries risk of total loss.