NVDL vs NVDU: Leverage, Returns & Decay
Over the days both have traded, NVDL finished 3.8 points from its stated multiple and NVDU 4.0. Updated Nov 13, 2026.
- Return gap
- +0.3 pts1 year to Nov 13, 2026: NVDU ahead, total return
- Fee gap
- 0.72 ptsA year: NVDL 1.64%, NVDU 0.92%, per filings dated Mar 3, 2026 and Oct 28, 2026
- Overlap
- –No book read for both funds
- Net flow
- –Needs daily flows for both funds
Compare two other funds
Growth of 100
Drawn from the numbers on this page. Growth of 100, both funds, distributions reinvested
NVDL −41.1%, NVDU −40.8%: NVDU +0.3 pts ahead, 1 year, Nov 13, 2025 to Nov 13, 2026.
ShareUse this comparison
Client brief:
The two-page PDF for this pair is coming. Save the comparison to follow its changes. Free. It needs an email sign-in, once.
Fees over time
Hypothetical. $10,000 in NVDL for 10 years at 1.64% a year costs about $2,586 in fees and lost growth; NVDU at 0.92% would cost $1,495. Assumes 6% a year before fees, a sample rate, not a forecast.
How each fund moved
NVDL returned −51.4% while 2 times NVDA's move would have been −55.2%. Data from Aug 14, 2026 to Nov 13, 2026.
NVDL returned −51.4% while 2 times NVDA's move would have been −55.2%. Data from Aug 14, 2026 to Nov 13, 2026.NVDU returned −51.2% while 2 times NVDA's move would have been −55.2%. Data from Aug 14, 2026 to Nov 13, 2026.
Data from Aug 14, 2026 to Nov 13, 2026.IQ Leverage Score
How closely it delivered its daily multiple over 3 months, ranked against 498 leveraged ETFs at 2x, over three months. 100 is best.
NVDU scores higher · Did it keep up with its own daily multiple, compounded day by day, over three months?
Three months only, against the fund's own daily multiple; a different quarter can rank it very differently.
Key numbers
NVDU costs 0.72 points a year less; their one-year returns differ by 0.3 points; NVDL is 5.8 times larger.
Pair scorecard
Open the table
Points (pts) are percentage points.
| Total return, 1 year | −41.1% | −40.8% |
|---|---|---|
| What its daily multiple gave, 3 months | −50.0% | −50.0% |
| What it tracks | NVDA | NVDA |
| Against its daily multiple, 3 months | 1.4 points short | 1.2 points short |
| Expense ratio | 1.64% | 0.92% |
| Net assets, as of Nov 12, 2026 | $2.5bn | $434m |
| Holdings in common | not published |
The numbers behind it
Net assets on Nov 12, 2026: $2.5bn for GraniteShares 2x Long NVDA Daily ETF and $434m for Direxion Daily NVDA Bull 2X ETF.
Side by side
Open the table
| Fund | GraniteShares 2x Long NVDA Daily ETF · Aims to return twice the daily move of NVIDIA (NVDA) | Direxion Daily NVDA Bull 2X ETF · Aims to return twice the daily move of NVIDIA (NVDA) |
| Issuer | GraniteShares | Direxion |
| Sets out to return | +2x | +2x |
| Underlying asset | NVDA | NVDA |
| Segment | company | company |
| Fund returned, 3 months | −51.4% | −51.2% |
| Underlying returned, over that window | −27.6% | −27.6% |
| What the stated multiple implies, over that window | −55.2% | −55.2% |
| Difference from stated, over that window | +3.8 pts | +4.0 pts |
| Fund returned, 1 year | −41.1% | −40.8% |
| Difference from stated, over that window | −15.9 pts | −15.6 pts |
| Underlying volatility | 42% | 42% |
| Difference over the days both have traded | +3.8 pts | +4.0 pts |
| Expense ratio | 1.64% | 0.92% |
| First traded | Dec 13, 2022 | Sep 13, 2023 |
| Net assets, as of Nov 12, 2026 | $2.5bn | $434m |
As of Nov 13, 2026. Source: ETFIQ.
Window by window
Open the full table
| 1 month | −48.0% | −47.9% | −52.8% | −52.8% | +4.8 pts | +4.9 pts |
|---|---|---|---|---|---|---|
| 3 months | −51.4% | −51.2% | −55.2% | −55.2% | +3.8 pts | +4.0 pts |
| 6 months | −54.4% | −54.3% | −55.2% | −55.2% | +0.8 pts | +0.9 pts |
| 1 year | −41.1% | −40.8% | −25.2% | −25.2% | −15.9 pts | −15.6 pts |
| 3 years | +251.0% | +214.6% | not meaningful | not meaningful | not meaningful | not meaningful |
| Since launch (Not the same days: NVDL Dec 2022 · NVDU Sep 2023) | +1,265.8% | +251.1% | not meaningful | not meaningful | not meaningful | not meaningful |
From Sep 2023 to Nov 2026, NVDL returned more: +367.6% to +317.8%. NVDL and NVDU swung about the same: 87.8% and 87.0%. NVDL and NVDU fell about the same at their worst: −66.8% and −66.6%.
Full scorecard
Open the full table
| Total return | +367.6% | +317.8% | NVDL | Price change plus distributions, reinvested. | |
| Annualized | +63.1% | +57.4% | NVDL (slight) | Total return as a yearly rate. | |
| Price return | +320.1% | +230.9% | NVDL | Price change alone, distributions left out. | |
| Volatility | 87.8% | 87.0% | Tie | Spread of weekly returns, annualized. Lower is steadier. | |
| Downside deviation | 53.9% | 53.7% | Tie | Like volatility, counting only weeks below the Treasury bill rate. | |
| Max drawdown | −66.8% | −66.6% | Tie | Deepest fall from a high to the low after it. | |
| Upside capture | Not measured | Not measured | 100.0% | Share of its benchmark's rise caught in the weeks it rose. | |
| Downside capture | 108.3% | 108.3% | 100.0% | Tie | Share of its benchmark's fall taken in the weeks it fell. Lower is better. |
| Beta | 1.97 | 1.95 | 1.00 | Tie | How far it moves for each 1% its benchmark moves. |
| R squared | 0.99 | 0.99 | 1.00 | Not an edge | Share of its weekly moves its benchmark explains. |
| Sharpe | 0.94 | 0.90 | NVDL (slight) | Return above Treasury bills per unit of volatility. | |
| Sortino | 1.53 | 1.46 | NVDL (slight) | Return above Treasury bills per unit of downside deviation. | |
| Treynor | 0.418 | 0.401 | 0.494 | NVDL (slight) | Return above Treasury bills per unit of beta. |
| Calmar | 0.95 | 0.86 | NVDL (slight) | Annualized return divided by the max drawdown. | |
| Benchmark | NVDA | NVDA | NVDA | The ETF each fund is measured against. | |
| Alpha | −14.9% a year | −18.2% a year | 0.0% | Both significant | Yearly return beyond what its beta to the benchmark predicts. |
| Alpha t-stat | −3.59 | −3.67 | Not an edge | How sure the alpha is. Below 2, it could be chance. | |
| Alpha range | −23.1% to −6.7% | −28.0% to −8.4% | The 95% range for the yearly alpha. | ||
| Information ratio | 0.76 | 0.67 | NVDL | Return against its benchmark per unit of tracking error. | |
| Tracking error | 43.7% | 43.1% | 0.0% | Not an edge | How far weekly returns stray from the benchmark, annualized. |
More
Computed by ETFIQ from Tiingo daily prices and ETFIQ's distribution records, 164 weekly returns. Cash rate: the 13-week Treasury bill.
About NVDL
Over the three months to Nov 13, 2026, NVDL lost 51.4%, against −50.0% for twice NVDA's daily moves. Read the multiple against the whole window instead and 2 times NVDA's −27.6% implies −55.2%, which makes NVDL look 3.8 points over.
More
That is not the fund's objective, which is daily. Of that 3.8, daily compounding is +5.2 points, which any 2 times fund would show over the same path, and the fund itself is −1.4 points against 2 times each day's move, compounded. NVDL aims to return +2 times NVDA's move each day, then resets. Over one day it does that. Over longer periods the daily results compound, so the total is not +2 times the period's move: it is lower after a fall and recovery, and can be higher in a steady run. NVDA moved at 42% annualized over that window. The more it moves, the further the two numbers separate.
About NVDU
Over the three months to Nov 13, 2026, NVDU lost 51.2%, against −50.0% for twice NVDA's daily moves. Read the multiple against the whole window instead and 2 times NVDA's −27.6% implies −55.2%, which makes NVDU look 4.0 points over.
More
Of that 4.0, daily compounding is +5.2 points, which any 2 times fund would show over the same path, and the fund itself is −1.2 points against 2 times each day's move, compounded. NVDU aims to return +2 times NVDA's move each day, then resets.
Questions people ask
Which came closer to its daily multiple, NVDL or NVDU?
Over the window to Nov 13, 2026, NVDL finished 3.8 points from what its multiple implies and NVDU finished 4.0 points from its own, so NVDL came closer. Neither number predicts the next window: it depends on how much the underlying moves about.
Are NVDL and NVDU levered on the same thing?
Yes. Both are levered on NVIDIA, NVDL at +2 times and NVDU at +2 times the daily move.
Which one decays faster, NVDL or NVDU?
Decay follows how much the underlying moves about. Over this window NVDL’s moved at 42% annualized and NVDU’s at 42%, so NVDL has the rougher ride and, at the same multiple, loses more to compounding.
Can I hold NVDL or NVDU for a year?
Both reset every day, so the multiple in the name applies to a single day, and daily results compound. Over a year the total is not the multiple times the year’s move; after a fall and recovery it is lower.
Other comparisons
Sources
More
Links are sources, not endorsements. A comparison is not a recommendation.
Copy and cite
Cite this pageCitation, APA and Markdown
ETFIQ, NVDL against NVDU, data as of Nov 13, 2026. https://etfiq.com/compare/leverage/nvdl-vs-nvdu
ETFIQ. (Nov 13, 2026). NVDL against NVDU. Retrieved from https://etfiq.com/compare/leverage/nvdl-vs-nvdu
[NVDL against NVDU (ETFIQ, Nov 13, 2026)](https://etfiq.com/compare/leverage/nvdl-vs-nvdu)
Free to use with attribution for data ETFIQ computes; issuer data stays under its owner's terms.