DRAM $40 Call Expiring September 18th 2026

(Roundhill Memory ETF, $40 Call Expiring Sept 18th 2026)

Created: April 26th 2026

1️⃣ Trade Snapshot

• Instrument: Roundhill Memory ETF (ticker: DRAM). The first-ever pure-play memory semiconductor ETF, launched April 2, 2026.
• Trade type: Two-phase momentum play on the memory supercycle. Phase 1 (current): ride the supercycle upward with a single call option. Phase 2 (future): when cycle indicators flash, buy puts on the same ETF to profit from the downturn.
• This is NOT a value or undervaluation play. This is a deliberate, minimum-sized momentum trade that acknowledges we are in the mid-to-late innings of the memory supercycle. The ETF launch itself is a late-cycle sentiment indicator. This position represents a very small portion of the portfolio, designed purely to capture short-term gains on the AI memory trade.
• Current price (April 26, 2026): ~$37.10
• Position structure: 1 contract $40 Call expiring September 18, 2026. Nothing more.
• Exit rule: When DRAM reaches the $40-$50 range, sell the call.

What is this ETF? DRAM provides pure-play exposure to the global memory semiconductor industry. It holds the three companies that manufacture virtually all of the world's DRAM, HBM, and NAND: SK Hynix (~25.8%), Micron (~24.6%), and Samsung (~24.1%). These three make up ~74.5% of the fund. The remaining ~25.5% includes Kioxia (~6%), SanDisk (~5%), and other memory companies. The ETF uses a modified market-cap weighting with a 25% cap per company and may hold derivatives (swaps/forwards) for exposure to non-U.S. listed stocks like SK Hynix and Samsung.

Why an ETF instead of individual stocks? SK Hynix (57% HBM market share) and Samsung are Korean-listed and not accessible on Wealthsimple. DRAM gives you exposure to the full memory complex through a single U.S.-listed instrument with liquid options (~11,000 contracts/day).

How Phase 1 makes money:

• The $40 strike requires only an ~8% move from current price. A modest bar for an ETF that moved 42% in its first 3 weeks.
• UBS projects DRAM contract prices rising another 58-63% in Q2 2026, a direct tailwind.
• MU earnings (July 1), Samsung earnings (July), and SK Hynix earnings (July) are all catalysts within the window.
• You can sell the call at any time for a profit.

How Phase 2 works (future, not active now):

• After exiting the call in the $40-$50 range, hold cash and monitor the memory supercycle indicator checklist.
• When 3+ Tier 1 indicators flash (DDR5 spot prices declining, inventories rising above 5 weeks, CapEx announcements accelerating, China capacity reaching critical mass), buy puts on DRAM to profit from the downturn.
• The same ETF works for both directions. No need to switch instruments.

2️⃣ What is the Memory Supercycle?

Memory semiconductors (DRAM, NAND, HBM) are one of the most cyclical parts of the tech industry. Demand surges, prices spike, manufacturers invest in new capacity, that capacity comes online 2-3 years later, supply overwhelms demand, prices crash. This cycle has repeated at least 4 times in the last 30 years, with corrections of 30-50%+ each time.

The current cycle (2024-present) is driven by AI data centers requiring HBM (High Bandwidth Memory) for GPUs. HBM uses 3-4x the wafer capacity of standard DRAM, so shifting production to HBM simultaneously constrains commodity DRAM supply, creating shortages across the board.

Where we are now (April 2026):

• DRAM prices up ~171% YoY
• DDR5 spot prices quadrupled since September 2025
• Supplier inventories at 3.3 weeks (extreme tightness, matching 2018 supercycle lows)
• HBM capacity 100% sold out through 2026 under binding contracts
• Microsoft and Google signed unprecedented 3-year contracts with 10-30% advance deposits

Why we believe the cycle continues for at least 6 more months:

• UBS projects DRAM supply-demand gap persists through Q4 2027
• Q2 2026 contract prices projected to rise another 58-63%
• HBM4 demand for NVIDIA's Vera Rubin platform is pulling capacity forward
• No major new supply (fabs) comes online until 2027-2028 at earliest
• The 3-year contracts provide a revenue floor that didn't exist in prior cycles

Why we are cautious (mid-to-late cycle signals):

• The DRAM ETF itself launching is a classic late-cycle sentiment indicator (BTIG flagged this)
• Reports of double-ordering and inflated backlogs
• Gross margins approaching 55-60% historical peak zone
• China's CXMT approaching 800K wafers/quarter by 2027
• All three major manufacturers investing aggressively in new capacity ($25B MU, $29B SK Hynix, 110T won Samsung)

3️⃣ Key Dates & Numbers

4️⃣ Risk Management

• Exit in the $40-$50 range. When DRAM reaches this range, sell the call. This is a predefined exit. No second-guessing.
• Stop loss on call: 50% of entry premium. If the call loses half its value, exit.
• Do not add to this position. 1 contract is the full size. This is deliberately the smallest possible position because we are entering late in the cycle.
• If holding the call into August and the position is profitable, consider selling before September expiry to avoid accelerating theta decay in the final 30 days.
• Monitor the supercycle indicators monthly. If Tier 1 indicators start flashing (DDR5 spot prices declining, inventories rising), exit immediately regardless of price. The cycle turning is more important than hitting the target range.

This is a momentum trade, not a conviction hold. The sizing (1 contract only) reflects that we are knowingly entering mid-to-late cycle. The maximum loss is the call premium, which represents a very small portion of the total portfolio. This trade exists purely to capture short-term upside on the AI memory theme with defined, limited risk.

5️⃣ ETF Structure & Holdings

Key structural note: The ETF uses swaps and forwards for non-U.S. listed stocks (SK Hynix, Samsung). This adds counterparty risk that direct stock ownership doesn't have, but it's the only way to get these holdings into a U.S.-listed fund.

Rebalancing: At least quarterly, using active management with a 25% cap per company.

6️⃣ Phase 2 Preview (Not Active Yet)

Phase 2 will only activate after Phase 1 is complete (call sold in the $40-$50 range or stopped out).

When the supercycle indicators flash, the plan is to buy puts on DRAM to profit from the memory cycle downturn. A separate indicator checklist has been developed that tracks leading, coincident, and lagging signals for the memory cycle peak. The key triggers are:

• DDR5 spot prices declining for 3+ consecutive weeks
• Supplier inventories rising above 5 weeks
• Management language shifting from "sold out" to "balanced"
• Gross margins flattening or guiding down
• Any hyperscaler deferring HBM deliveries (strongest signal)

Phase 2 strike, expiry, and sizing will be determined when conditions are met. This document will be updated at that time.