What percentage of NVIDIA’s new equity commitments to independent AI developers and GPU-cloud providers from November 13, 2026 through June 30, 2027 will go to its three largest recipients?
Latest forecast
Generated Nov 13, 2026, 12:29 PM · High effort
Write-up
Executive Summary
The median forecast is 96.5% (50th percentile), with a range from 67% at the 10th percentile to 100% at the 90th percentile. The central expectation is therefore that NVIDIA’s three largest recipient groups will receive nearly all of its qualifying, publicly quantified new equity commitments during November 13, 2026–June 30, 2027. A broader spread remains plausible.
This measures concentration within the companion total-commitments question’s ledger, referred to as Q1, not across NVIDIA’s entire investment activity. Commitments are combined by corporate group—an ultimate parent and its controlled subsidiaries—using ownership at each commitment date.
The forecast is conditional on a positive Q1 total that can be allocated to recipient groups. A zero or unresolvable denominator means annulment, not zero concentration.
Key drivers
- The demonstrated pipeline is smaller than the headline deal list. Anthropic and Nscale filings describe agreements entered before the forecast window. Their later closings alone would not establish new commitments. Reporting also indicated that Thinking Machines had contracted earlier and NVIDIA was not expected to participate in OpenAI’s current round. Figure and Firmus remained prospective opportunities, not established in-window allocations.
- Strategic investments have differed greatly in size. Selected pre-window official disclosures included NVIDIA’s $30 billion allocation to OpenAI and its $2 billion CoreWeave common-stock investment. Such differences support a pattern in which a few large recipients dominate many smaller allocations. These are historical reference points, not amounts included in the forecast-window ledger.
- Tighter approvals could increase concentration. Following Colette Kress’s appointment as interim CEO, Reuters reported that she and her staff had challenged risky customer-financing proposals. NVIDIA’s official announcement nevertheless emphasized continuity, not an investment halt. Greater selectivity could leave fewer qualifying recipients; if only one, two or three groups receive positive amounts, the top-three share is necessarily 100%.
- Disclosure matters as much as participation. NVIDIA’s November 9 Korean-startup announcement named investments and planned investments without individual amounts. Such announcements do not automatically create quantified ledger entries. Later filings could reveal additional allocations, including smaller ones that reduce the top-three share, provided the commitments fall within the window and satisfy Q1.
Scenarios
The reasoning gives greatest weight to selective or sparse investing. Broader participation and later disclosure account for the lower end of the forecast distribution.
| Scenario | Weight in the reasoning and why | Effect on concentration |
|---|---|---|
| Selective strategic investing | Greatest weight: historical differences in investment size support a few large strategic recipients alongside smaller allocations, while tighter approvals favor selection. | Large recipients account for most of the total, keeping concentration high. |
| Sparse, cautious investing | Substantial weight: the timing audit removes several apparent future deals, and the leadership transition could slow new approvals. | A small recipient count pushes concentration toward the ceiling; no more than three positive groups produces 100%. |
| Broader investing without a dominant recipient | A meaningful but smaller role: the evidence does not establish a ceiling on future recipient count or guarantee another unusually large investment. | Numerous medium-sized allocations leave more of the total outside the largest three, lowering concentration. |
| A genuinely new mega-commitment | A small role: NVIDIA has made very large strategic allocations, but familiar old announcements cannot simply be carried into this window. | A new dominant allocation pushes concentration toward the ceiling. |
| Extensive later quantification | The least weight, but retained because the evidence cutoff permits later filings to reveal amounts missing from initial announcements. | Many additional, smaller qualifying allocations lower the measured top-three share. |
Implications
- Focus diligence on the largest relationships and their terms. If the forecast is well-calibrated, a few recipient groups will explain most of the publicly quantified qualifying allocation. Counting investment announcements would be a poor substitute for examining NVIDIA-specific amounts.
- Do not interpret restraint as diversification. A reduction in total commitments could coexist with higher concentration if fewer recipients receive qualifying allocations. Conversely, a broader spread of medium-sized investments could lower concentration without requiring a larger dominant deal.
- Keep this measure separate from total strategic exposure. High concentration in the resolved ledger would not establish concentration across undisclosed investments or other commercial relationships. Nor does this percentage forecast the total dollars committed.
Signs to watch
| Timing or development | What to watch and how it would affect the forecast |
|---|---|
| Around November 16, 2026: expected Nscale funding | The prospectus links this expected closing to a September agreement. Closing alone should not move the forecast. Evidence of a genuinely additional in-window obligation would warrant a rerun: a dominant new allocation would push concentration higher, while a wider set of comparable allocations would push it lower. |
| Unscheduled capital-allocation announcements | Confirmation of selective approvals that remove smaller recipients would favor higher concentration. Withdrawal from unusually large investments while smaller commitments continue would favor lower concentration. An explicit policy change would settle a major uncertainty and make the forecast worth rerunning. |
| Signed terms from prospective recipients, including Figure and Firmus | Official NVIDIA-specific amounts and commitment dates would matter more than reports of fundraising discussions or whole-round totals. Several comparable qualifying allocations would move concentration lower; one dominant new allocation would move it higher. Figure’s eligibility also depends on Q1’s recipient rules. Material signed terms warrant a rerun. |
| June 30, 2027: commitment window closes | The end of the window narrows uncertainty about new commitments, although supporting evidence may arrive later. A positive ledger with no more than three recipient groups fixes the share at the ceiling; a broader set of similarly sized recipients supports a lower result. |
| Through December 31, 2027: public-evidence cutoff | Filings that reveal additional qualifying amounts outside the largest three would lower concentration. Additional amounts allocated to the largest recipients would raise it, holding other allocations unchanged. Material disclosures warrant a rerun; finalizing the shared ledger should narrow uncertainty. Resolution follows the cutoff once Q1 is finalized. |
Research and methodology
The most influential sources were:
- Official transaction disclosures: OpenAI’s February 27 financing announcement and CoreWeave’s January 26 investment announcement supplied NVIDIA-specific historical amounts and illustrated differences in investment size.
- Recipient securities filings: Nscale’s November 12 amended prospectus and Anthropic’s November 10 amended prospectus were central to distinguishing agreements from subsequent funding or closing events.
- Leadership and approval evidence: NVIDIA’s leadership-transition filing established the change, while Reuters’ November 12 reporting informed the possibility of tighter investment selection. Reported past behavior was not treated as an announced investment moratorium.
- Pipeline reporting: The Information’s November 12 update helped distinguish active possibilities from earlier agreements and unsupported assumptions about NVIDIA’s participation.
- Disclosure practice: NVIDIA’s November 9 Korean-startup announcement illustrated why named investments and publicly quantified allocations are different datasets.
The research assembled a selected historical reference dataset, separating official amounts from reported amounts and checking for old commitments, conversions and deferred funding. It was not a complete historical Q1-compatible ledger.
A scenario-based simulation then varied the number of recipient groups and the differences in their total allocations. Each simulated amount represented a group’s aggregate over the entire window, so repeat investments did not create extra recipients. The largest three group totals were divided by the total for all qualifying groups. The assumptions and scenario weights were judgmental rather than statistically fitted to a complete historical series; uncertainty about the investing pattern matters more than simulation precision.
Limitations
The full finalized Q1 specification was unavailable. Its detailed rules on instruments, conditional commitments, recipient eligibility and source priority remain binding; the forecast does not invent replacements.
The historical reference dataset was selective, and underlying executed agreements or precise signing dates could not be obtained for every reported transaction. Undisclosed NVIDIA allocations cannot be estimated into the resolved ledger. The missing Q1 text, underlying agreements and later recipient filings are the most important evidence gaps.
Forecast history
Hover a data point to trace its series, or click to view the forecast generated at that time.
Signed forecast receipt
Signed Nov 13, 2026, 12:29 PM with ed25519 key preseen-prod-ed25519-20260523 and externally timestamped Nov 13, 2026, 12:29 PM.
sha256:c6c0332288adf5...40de044ee3Want forecasts like this?
Preseen researches your questions and answers them with calibrated, source-backed forecasts.