Grayscale has replaced its bitcoin miners ETF with GCPU, a fund that now tracks companies in high-performance computing and artificial intelligence. The theme change has not prevented a sharp drop, roughly 37% over three months according to Tickeron. This analysis looks at the issuer’s thesis, the portfolio composition, the diverging performance figures across sources, and the identified risks.
๐ Key takeaways
- GCPU replaces the MNRS fund and tracks the Indxx High Performance Computing Index, now focused on AI compute.
- The top ten holdings account for 68.01% of the portfolio, with a heavy presence of former bitcoin miners.
- The share price fell 22.6% over 30 days according to Tickeron, to about $21.28.
- The fund does not invest directly in digital assets, but its behaviour remains sensitive to that market.
A fund repositioned around AI compute
Introduced on September 22, 2026 by Grayscale, GCPU trades on NYSE Arca. It replaces the Grayscale Bitcoin Miners ETF (ticker MNRS) and now tracks the Indxx High Performance Computing Index, formerly the Indxx Bitcoin Miners Index. The fund was launched on January 30, 2025. Grayscale Advisors serves as adviser and Foreside Fund Services as distributor. The issuer, founded in 2013, presents itself as the largest investment platform focused on digital assets, based on assets under management as of June 30, 2026.
- Annual expense ratio: 0.59%
- Distribution frequency: semi-annual
- Bloomberg IOPV ticker: GCPU.IV; index ticker: IBMIIN
- CUSIP: 38963H206; ISIN: US38963H2067
According to Grayscale, the rules-based index, rebalanced quarterly, targets companies active in high-performance computing, AI cloud, accelerated computing infrastructure, data hosting, GPU services, and related hardware and software. Roughly half of the target portfolio consists of companies built around GPU cloud and AI hosting. The other half includes firms, notably bitcoin miners, that have announced or executed a shift of their power capacity toward AI. Constituents are selected and weighted by market capitalisation and theme-related revenues.

A thesis built on scarce capacity
Grayscale’s thesis rests on a gap between demand for AI compute and available physical capacity. Occupancy rates in North American data centers are reportedly at historic lows, with available capacity representing only about six months of demand. Building a new data center can take two to five years. CBRE, cited as of August 27, 2026, describes this trend, while McKinsey (document dated August 1, 2025) and Giga Energy (guide dated March 12, 2026) document these timelines.
| Indicator | Value | Cited source |
|---|---|---|
| Data center construction time | 2 to 5 years | McKinsey, Giga Energy |
| Available capacity in North America | about 6 months of demand | CBRE |
| Global AI investment in 2026 | over $1 trillion | Goldman Sachs |
| AI investment as share of GDP | about 1% (vs. 2% to 5% in earlier build-outs) | Issuer |
In a note dated August 7, 2026, Goldman Sachs estimates that global AI investment will exceed one trillion dollars in 2026, with most of it directed at physical infrastructure. For Grayscale, this gap is a durable and investable dynamic. In its launch announcement, Steve Vanourny, head of the index at Grayscale, put it this way:
“Grayscale has spent more than a decade giving investors access to rare and transformative technologies, such as digital assets, and we believe compute belongs in that same category. GCPU is a direct extension of that thesis, giving investors access to the companies closing the gap between AI compute demand and the physical infrastructure needed to meet it.”
Steve Vanourny, Head of Index, Grayscale
Portfolio composition: high concentration
As of November 13, 2026, the top ten positions totalled 68.01% of the fund. Robinhood shows 67.92% as of November 12, 2026, a gap that reflects the date and calculation method. CoreWeave leads with 8.74%, followed by DigitalOcean (8.44%) and Hut 8 (7.71%).
| Rank | Company | Weight | Close ($) |
|---|---|---|---|
| 1 | CoreWeave | 8.74% | 53.13 |
| 2 | DigitalOcean | 8.44% | 86.32 |
| 3 | Hut 8 | 7.71% | 57.57 |
| 4 | Nebius Group | 7.01% | 115.88 |
| 5 | SBI Holdings | 6.72% | 2,946.50 |
| 6 | Hewlett Packard Enterprise | 6.35% | 48.29 |
| 7 | Core Scientific | 6.21% | 14.52 |
| 8 | TeraWulf | 5.85% | 12.62 |
| 9 | Riot Platforms | 5.64% | 17.52 |
| 10 | IREN | 5.34% | 18.17 |
Tickeron estimates the fund holds about 30 securities and remains heavily oriented toward bitcoin mining infrastructure despite its new mandate. Sector exposure stays concentrated in information technology, with an additional weight in financial services linked to mining stocks. Those names, including Hut 8, IREN, Core Scientific, TeraWulf and Applied Digital, have often shown a beta above that of bitcoin.
Performance: figures that diverge by source
Tickeron reports a decline of about 22.6% over 30 days, from a close near $27.50 to roughly $21.28. Over three months, the drop is estimated at 37%, starting from a close near $33.95. The first part of that period covered the former mining index before the September conversion; the second covered the new AI compute strategy. Coinbase shows -18.54% over one month and -44.35% over one year.
| Source | Price ($) | Date | Change |
|---|---|---|---|
| Grayscale (NAV) | 21.87 | Nov 13, 2026 | -4.54% on the day |
| Grayscale (market price) | 21.83 | Nov 13, 2026 | -4.75% on the day |
| Robinhood | 21.63 | Nov 16, 2026 | Range 21.05 to 22.44 |
| Tickeron | 21.28 | End of period | -22.6% over 30 days |
| Light Horse (NAV) | 23.85 | Nov 11, 2026 | Not specified |
As of November 13, 2026, Grayscale reported $7,654,603 in assets under management for 350,000 shares outstanding. The market price traded at a 0.18% discount to NAV, and the median 30-day bid-ask spread stood at 0.32%. The last distribution, of $0.162885 paid on January 7, 2026, was not broken down between income, short- or long-term gains, and return of capital.
Tickeron attributes the decline to volatility in digital asset markets, which weighed on repurposed miners, and to the sell-off in capital-intensive growth names such as Nebius, CoreWeave and DigitalOcean. The source also points to a rotation of institutional and retail capital toward more liquid, less volatile segments. In its view, the fund’s concentration, small size and moves in a few large positions amplify its volatility. Light Horse Securities, for its part, lists a NAV of $23.85 and leverage of 1, yet its description still carries the MNRS name and the former mining theme, which contradicts GCPU’s current status.
Real bitcoin exposure and risks
The fund documents state that it does not invest directly in digital assets or ICOs, and does not use derivatives to give holders such exposure. It may, however, hold companies that have operated in mining, blockchain or digital infrastructure. Their inclusion depends on their high-performance computing or AI cloud activities, not solely on their crypto business. The fund therefore does not directly track digital asset price movements.
The documentation flags non-diversification risk, potentially high turnover, possible exposure to emerging markets, intense competition and rapid obsolescence in AI infrastructure. It also cites dependence on customer spending, supply chain disruptions, power shortages, export controls and overbuilding of capacity. Regulatory risks tied to AI, data protection and energy complete the list, along with the limited operating track record of some former miners entering the AI market.
- Direction of digital asset markets
- AI capital spending and data center demand
- Conversion of mining operations into contracted recurring revenue
- Interest rates and access to capital for energy-intensive projects
- Regulation, portfolio concentration and fund inflows or outflows
Tickeron identifies these monitoring factors and concludes that together they keep the fund’s risk level high.

Conclusion: a plausible thesis, a fragile vehicle
GCPU rests on a credible structural thesis: if compute capacity stays scarce for several years, data center and GPU operators could retain lasting pricing power. In a bullish scenario, effective conversion of former miners into contracted revenue and sustained data center demand could reduce the fund’s sensitivity to bitcoin.
In a bearish scenario, overbuilding, tighter rates, slower AI spending or another period of weakness in digital assets could prolong the decline. Concentration of about 68% in ten names and the fund’s small size add to these risks. Investors should therefore monitor the actual portfolio composition and fund flows rather than the stated theme alone.
Sources
- Tickeron, GCPU analysis
- Grayscale, GCPU official page
- Coinbase, GCPU quote page
- Robinhood, GCPU quote page
- Light Horse Securities, GCPU quote page
- Grayscale, GCPU launch announcement
Cet article est publiรฉ ร titre informatif et รฉducatif. Il ne constitue en aucun cas un conseil en investissement. Faites vos propres recherches (DYOR) avant toute dรฉcision.


