Thematic Research · Bitcoin × Artificial Intelligence · 2026–2030
The AI economy's Bitcoin TAM is option value: $1.5–5 trillion of 2030 forecasts against roughly $50 million of settled reality.
Version 1.1 · All claims independently verified against primary sources, 10 June 2026 · Download PDF
Executive summary
The "AI pays in Bitcoin" thesis is real but pre-revenue. Forecasts put agentic commerce at $1.5–5 trillion of annual value by 2030[1][2]; the leading live payment rail for AI agents has settled roughly $50 million — cumulatively — through April 2026[7][17]. That rail is not Bitcoin: Coinbase's x402 protocol settles in stablecoins and was institutionalized under the Linux Foundation in April 2026 with Stripe, Cloudflare, Shopify, and Solana as founding members[5][6]. Bitcoin's native agent stack — L402, Lightning agent tools, identity-free pay-per-call — shipped in February 2026 and is technically complete, with no published volume[3][4].
Two structural developments define the current state of play. First, the convergence path opened: Tether's USDT went live on Lightning via Taproot Assets in March 2026, explicitly marketed for AI-agent payments — meaning "stablecoins win" and "Bitcoin rails win" are no longer mutually exclusive outcomes[13][14]. Second, the one channel where AI and Bitcoin already exchange contracted dollars is infrastructure, not payments: more than $70 billion of AI and high-performance-computing contracts have been announced across the public Bitcoin mining sector, and the resulting capacity migration has pushed network hashrate into the longest sustained drawdown of the industrial mining era[10][11]. The AI-treasury channel, by contrast, is confirmed empty: no AI company holds Bitcoin per filings, and the measurable corporate flow runs the other way — miners selling Bitcoin to fund AI conversions[11][15].
The Fund tracks five signposts on this theme (Exhibit 5). One is partially triggered. This note presents the analysis; it does not constitute investment advice or a recommendation regarding any security or digital asset.
Three transmission channels
C1 — Payments. AI agents settle machine-to-machine commerce on Bitcoin or Lightning, creating fee demand, channel liquidity, and monetary-premium effects.
C2 — Treasury. AI companies or sovereigns accumulate Bitcoin as the machine economy's reserve asset.
C3 — Infrastructure. Bitcoin miners convert power, land, and cooling into AI compute — value accrues to converting equities and is neutral-to-negative for Bitcoin's own network fundamentals.
As of June 2026: C1 is pre-revenue, C2 is empty, and C3 carries all of the contracted dollars.
The forecasts — and why they overstate the settlement opportunity
Data: Juniper Research forecasts $1.5 trillion of agentic commerce spend in 2030, from a 2025–26 base it characterizes as pilot deployments only (secondary coverage attributes a ~$8 billion 2026 baseline — an implied ~187× ramp in four years)[1]. McKinsey's October 2025 study sizes agentic commerce at $3–5 trillion globally by 2030, including up to $1 trillion of orchestrated US retail revenue[2].
Assessment: McKinsey's operative verb is "orchestrate," not "settle." Commerce an agent arranges will mostly clear on cards and ACH behind the agent interface. The settlement-layer opportunity is the relevant pool for crypto rails, and at crypto take-rates of 0.1–1.0% (versus 200–300bps for cards), even $3 trillion of agentic volume implies a $3–30 billion annual settlement-revenue pool — material, but two orders of magnitude below the headline figures commonly quoted.
Exhibit 1 · Third-party estimates and methodology assessment| Source | Figure | Method | Assessment |
|---|---|---|---|
| Juniper Research (Apr 2026)[1] | $1.5T in 2030 | Top-down adoption ramp from ~$8B 2026 base | ~187× implied four-year ramp; no rail-share split; volume ≠ settlement revenue |
| McKinsey (Oct 2025)[2] | $3–5T by 2030 | Top-down share of global commerce | "Orchestrated" ≠ settled — most volume clears on legacy rails behind the agent UX |
| Chainalysis (Jun 2026)[7] | >100M txns | Bottom-up on-chain count | ~Half artificial per Artemis[9]; counts ≠ value (~$50M cumulative[17]) |
The live rails: a stablecoin incumbent, a Bitcoin challenger, and a bridge
x402 — the incumbent
Data: x402 (Coinbase-developed, HTTP-402-based, USDC settlement) went from near zero in mid-2025 to over 100 million cumulative transactions by Q1 2026; by April 2026, coverage cites 165 million-plus transactions, roughly $50 million of cumulative volume, and 480,000-plus transacting agents[7][17]. Daily run-rate as of March 2026: ~131,000 transactions generating ~$28,000 of volume, at an average payment of ~$0.20[8]. Artemis Analytics attributes roughly half of observed transactions to artificial activity — self-dealing and wash patterns[9]; early growth was inflated by memecoin minting[7]. Payment quality is nonetheless improving: $1-plus payments rose from 49% to 95% of volume in early 2026, and tester-to-payer conversion improved four-fold in six months[7][17]. The x402 Foundation now sits under the Linux Foundation (April 2026) with settlement on Base, Ethereum, Solana, and Stellar — and no Bitcoin or Lightning support announced[6].
L402 / Lightning — the challenger
Data: Lightning Labs open-sourced its Lightning agent toolkit in February 2026: seven composable skills covering node operation, key-isolated remote signing, scoped spending credentials, paying and hosting L402-gated APIs, and node queries over MCP; the lnget client pays a Lightning invoice on any HTTP 402 challenge with no account, API key, or identity, and pairs with the Aperture reverse proxy to close a full agent-to-agent commerce loop[3][4]. Coinbase launched its x402-based Agentic Wallets the same day[17]. Assessment: the Bitcoin-native stack is technically complete and its identity-free settlement is a genuine differentiator — agents do not have government IDs or bank accounts — but sixteen weeks post-launch there is no published L402 volume data, a gap we re-confirmed by direct search in June 2026. Public Lightning channel-capacity data for 2025–26 is internally contradictory across sources (Exhibit 2) and is flagged for primary-source reconciliation.
Taproot Assets — the bridge
Data: USDT went live on Lightning via Taproot Assets around March 2026 on a production-grade protocol (auditable issuer supply, multi-channel payments), per Tether CEO confirmation — though with no published supply or volume figures, no major exchange distribution, and a thin wallet ecosystem[13]. Tether and Bitfinex explicitly market the rail for AI-agent payments via L402[14]. Assessment: this is the development that merges the scenarios — stablecoin-denominated agent payments settling over Bitcoin infrastructure. It is live, funded, promoted, and pre-volume. Notably, the two agent-payment ecosystems remain unmerged: the Bitcoin camp's standard is L402 while x402 excludes Bitcoin entirely.
Exhibit 2 · Key metrics| Metric | Value | As of | Status |
|---|---|---|---|
| Agentic commerce forecast, 2030 (Juniper)[1] | $1.5T (from ~$8B 2026 base) | Apr 2026 | Verified (base single-sourced) |
| Agentic commerce forecast, 2030 (McKinsey)[2] | $3–5T | Oct 2025 | Verified |
| x402 cumulative transactions / volume / agents[17] | 165M+ / ~$50M / 480k+ | Apr 2026 | Verified |
| x402 daily throughput[8] | ~131k txns · ~$28k · ~$0.20 avg | Mar 2026 | Verified |
| x402 artificial-activity share (Artemis)[9] | ~50% of transactions | Mar 2026 | Verified |
| L402 / Lightning agent-payment volume[3][4] | No published data | Jun 2026 | Confirmed gap |
| Lightning public channel capacity[16] | 4,100–5,637 BTC reported, late '25–'26 | Jun 2026 | Contested |
| USDT on Lightning (Taproot Assets)[13] | Live; no volume/exchange data | Mar 2026 | Verified (single-source date) |
| Public-miner contracted AI/HPC revenue[11] | >$70B cumulative | Q1 2026 | Verified |
| Bitcoin network hashrate, 30-day MA[10] | 964 EH/s · −13.2% vs peak · 187-day drawdown | 18 May 2026 | Verified, corroborated |
| AI/HPC equity value per MW[10] | ~3× since summer 2025 | 18 May 2026 | Verified |
| AI-company Bitcoin treasury holdings[15] | Zero confirmed | Jun 2026 | Verified (absence) |
The treasury channel is empty — and currently flows backward
Data: A June 2026 filings sweep found no AI company holding Bitcoin in corporate treasury. The closest adjacency is legacy: SpaceX — owner of xAI since February 2026 — holds 18,712 BTC (~$1.29 billion) per its IPO filing, purchased in 2021 with no AI rationale stated[15]. The US Strategic Bitcoin Reserve remains undisclosed in size, framed by officials entirely as custody and store-of-value — no official has invoked AI. The one measurable AI-related corporate Bitcoin flow is miners selling treasury Bitcoin to fund data-center conversions[11]. Assessment: we assign this channel no value in the base case and monitor for a first mover.
The infrastructure channel carries the contracted dollars
Data: More than $70 billion of cumulative AI/HPC contracts have been announced across public Bitcoin miners, including CoreWeave–Core Scientific at $10.2 billion over 12 years (a contractual relationship — the acquisition attempt failed in an October 2025 shareholder vote), TeraWulf at $12.8 billion of contracted HPC revenue, and Hut 8's $7 billion, 15-year River Bend lease, where 2026 coverage identifies Fluidstack and Anthropic as counterparties[11][12]. The eleven largest US public miners shed ~7 EH/s of hashrate in a single quarter as power moved to hyperscalers on 10–15-year leases; the per-megawatt equity value of AI capacity has roughly tripled since summer 2025; and network hashrate sits 13.2% below its November 2025 peak in a 187-day drawdown — the longest of the industrial mining era, with a further ~9% difficulty decline estimated for mid-June 2026[10].
Assessment: this is the only AI×Bitcoin channel with contracted dollars, and it cuts both ways — substantial value creation at converting operators against erosion of Bitcoin's security-budget inputs. Converting miners are progressively becoming AI-infrastructure businesses whose economics decouple from Bitcoin. The Anthropic/River Bend datapoint is the first direct AI-lab footprint on former Bitcoin-mining infrastructure; it is compute demand, not Bitcoin demand.
Exhibit 3 · US public miner hashrate, Q4 2025 → Q1 2026 (VanEck, company disclosures)| Company | Q4 2025 (EH/s) | Q1 2026 (EH/s) | Change | Stated AI/HPC posture |
|---|---|---|---|---|
| IREN | 43.0 | 35.8 | −17% | Multi-year mining exit committed |
| Keel (KEEL) | 19.5 | 12.0 | −38% | Furthest along; mining capex halted |
| Cipher (CIFR) | 15.6 | 11.1 | −29% | Full exit by July 2027 (outside date) |
| TeraWulf (WULF) | 6.8 | 4.0 | −41% | Exit by ~April 2028 halving |
| Core Scientific (CORZ) | 10.9 | 9.7 | −11% | Near-pure-play AI by early 2027 |
| Riot (RIOT) | 34.0 | 42.3 | +8.3 EH/s | Partial AI pivot; still growing hashrate |
| Bitdeer (BTDR) | 43.2 | 50.2 | +7.0 EH/s | Partial AI pivot; still growing hashrate |
| MARA | 51.9 | 55.5 | +3.6 EH/s | Partial AI pivot; still growing hashrate |
| Top-11 total | 293.9 | 287.1 | −6.8 EH/s | Structural capacity migration[10] |
Scenario analysis
Probabilities are TON618 Capital analyst judgment as of 10 June 2026, not market-implied. Scenarios are analytical constructs for assessing the theme; they are not forecasts of, or guides to, any portfolio action.
Exhibit 4 · 2030 scenarios, probability-weighted| Scenario | Prob. | Key assumptions | 2030 outcome for "AI on Bitcoin" |
|---|---|---|---|
| Bull | 15% | Agentic volume ≥$3T; identity-free settlement proves decisive; Taproot-Assets stablecoins scale on Lightning | Bitcoin rails carry 5–10% of agentic settlement ($150–300B); $1.5–3B settlement-revenue pool. Caveat: if the bull case arrives via stablecoins-on-Bitcoin-rails, value accrues to Bitcoin infrastructure while the unit of account stays USD |
| Base | 55% | Agentic volume <$1T by 2030; stablecoins take ≥90% of crypto-settled agent payments on x402-ecosystem chains; miner conversion continues | Bitcoin payments leg stays niche (<$50B); AI×Bitcoin value accrues through the infrastructure channel |
| Bear | 30% | Agentic payment demand stays negligible (extrapolating current run-rates); rail ambiguity resolves negatively; hashrate drawdown extends | Payments ≈ 0; treasury stays at zero; infrastructure conversion proceeds but fully decouples miners from Bitcoin; security-budget narrative erodes without an offsetting demand channel |
Weights unchanged from v1.0: this cycle's verification moved evidence roughly symmetrically (pro-Bitcoin-rails: USDT live on Lightning with explicit agent positioning; pro-stablecoin-incumbency: Linux Foundation institutionalization, zero treasury demand, no L402 volume).
Risk factors
Exhibit 5 · Principal risks to the theme| Risk | Likelihood | Magnitude | Note |
|---|---|---|---|
| Stablecoin rails win agent payments outright | High | High | Linux Foundation governance; Stripe/Shopify distribution[6]. Partial offset: Taproot-Assets bridge[13] |
| Agentic-commerce volumes disappoint vs. forecasts | Med-High | High | ~$50M cumulative settled vs. $1.5–5T forecasts[8][17] |
| Security-budget narrative damage from miner exodus | Medium | Medium | Record drawdown corroborated; further difficulty decline due June 2026[10] |
| Regulatory scrutiny of identity-free agent payments | Medium | Med-High | L402's differentiator is also its AML surface |
| Adoption metrics inflated by artificial activity | High | Medium | ~50% of x402 transactions artificial[9]; mitigant: payment quality improving[7] |
| Execution and counterparty concentration at converting miners | Medium | High | Fluidstack appears across multiple miner contracts[11][12]; failed CoreWeave acquisition shows deal risk[12] |
What we are watching
Exhibit 6 · Signposts, as of 10 June 2026| # | Signpost | Status |
|---|---|---|
| S1 | Published L402/Lightning agent-payment volume exceeding $1M/day | Not observed; data gap re-confirmed Jun 2026 |
| S2 | A major AI lab integrates Lightning/Bitcoin settlement in production | Not observed (River Bend involvement is compute, not payments) |
| S3 | x402 adds Bitcoin/Lightning settlement, or Taproot-Assets stablecoins reach scale on Lightning | Partially triggered — USDT live Mar 2026, pre-volume; x402 still excludes Bitcoin |
| S4 | Lightning public capacity inflects upward for two consecutive quarters | Indeterminate — capacity data contested; primary reconciliation in progress |
| S5 | Wash-adjusted crypto-settled agentic volume crosses $10B annualized | Not observed; ~$50M cumulative through Apr 2026 |
Sources
- Juniper Research, "Agentic Commerce Set to Generate $1.5 Trillion Globally by 2030," press release, 7 Apr 2026.
- McKinsey & Company, "The agentic commerce opportunity," Oct 2025; Digital Commerce 360 coverage, 20 Oct 2025.
- Lightning Labs, "The Agents Are Here and They Want to Transact," 11 Feb 2026.
- Lightning Labs, "L402 for Agents," 11 Mar 2026.
- Cloudflare, "x402: a new standard for internet-native payments," 23 Sep 2025.
- x402 Foundation establishment under the Linux Foundation, 2 Apr 2026 (founding members incl. Stripe, Cloudflare, Shopify, Solana); x402.org ecosystem documentation.
- Chainalysis, "x402 and agentic payments adoption," 3 Jun 2026.
- CoinDesk, "Coinbase-Backed AI Payments Protocol Wants to Fix Micropayments, but Demand Is Just Not There Yet," 11 Mar 2026.
- Artemis Analytics, x402 wash-activity analysis (30-day filter), 11 Mar 2026.
- VanEck, "Bitcoin ChainCheck," data as of 18 May 2026; corroboration: CoinDesk/CoinShares, "Bitcoin Hashrate Posts First Quarterly Drop in Six Years," 30 Mar 2026.
- CoinDesk, "Bitcoin Miners Are Becoming AI Companies — and Selling Their BTC to Fund the Transition" (CoinShares data), 27 Mar 2026.
- Core Scientific Form 8-K, Q1 FY2026 (SEC EDGAR); Sherwood News on the CoreWeave shareholder vote, Oct 2025; TeraWulf Form 8-K, Q1 2026 (SEC EDGAR).
- Tether, USDT on Lightning announcement, 30 Jan 2025; go-live coverage citing CEO confirmation, Mar 2026; Lightning Labs, Taproot Assets v0.7, 16 Dec 2025.
- Bitfinex, "Why Bitcoin and Stablecoins on Lightning Will Power the Next Phase of AI Agent Payments," 6 Mar 2026.
- CoinDesk, "SpaceX Holds 18,712 Bitcoin at Fair Value of $1.29B, IPO Filing Shows," 20 May 2026; SEC EDGAR filings sweep (Nvidia FY2026 et al.), Jun 2026.
- CryptoSlate (Aug 2025), news.bitcoin.com, and BYDFi (May 2026) — Lightning capacity series (figures conflict; flagged contested).
- Coinbase, Agentic Wallets launch, 11 Feb 2026; ecosystem adoption trackers, Apr 2026.
Use of AI & feedback. Artificial intelligence was used in the creation of this report; all methodology and data integrity have been reviewed and approved by a professional. Direct feedback to CIO Keyth Beck, keyth@ton618capital.com. Version 1.1 · analyst: TON618 Thematic Research.
TON618 Capital · Thematic Research · Published 10 June 2026. For information purposes only. This document is not an offer to sell or a solicitation of an offer to buy any security, fund interest, or digital asset, and it does not constitute investment advice or a recommendation to buy, sell, or hold any instrument. Nothing herein describes or implies any current or intended portfolio action. Statements of opinion ("we believe," scenario probabilities) are the views of TON618 Capital as of the publication date and are subject to change without notice; scenarios are analytical constructs, not predictions. Figures are drawn from third-party sources believed reliable but not independently guaranteed; verification status is noted per claim in the underlying research file. The Fund and its principals may hold positions in instruments economically related to the subjects discussed. Past performance is not indicative of future results. Digital assets are highly volatile and may result in total loss of capital.