7.2 Billion Tokens Added in Three Months: STRK's Real Supply Pressure Isn't Just from Unlocks

Published on 2026-10-12Last updated on 2026-10-12

Abstract

In the 90 days up to October 7th, the STRK supply increased by approximately 720 million tokens. Of these, 381 million came from monthly unlocks for contributors and investors, 300 million from the team's reserve wallet, and around 39 million from staking inflation. During the same period, there were no mechanisms such as burns or buybacks to offset the newly added supply.

Trading Snapshot

• Key Data|Net supply increase of approximately 720 million tokens in the past 90 days, equivalent to 9.70% of the 7.42 billion circulating supply.

• Supply Sources|Monthly unlocks of 381 million tokens, team reserve wallet releases of 300 million tokens, staking issuance of approximately 39 million tokens.

• Subsequent Schedule|Monthly unlocks of up to 127 million tokens each on October 15, November 15, and December 15, with the fixed series continuing until March 2027.

• Risk/Invalidation Conditions|STRK fees are currently not burned, nor is there any buyback; team reserve releases and staking issuance will cause actual supply changes to exceed the fixed unlock schedule.

Approximately 9.70% Supply Increase in 90 Days

Within the 90 days ending October 7, approximately 720 million STRK tokens entered the market, equivalent to 9.70% of the 7.42 billion circulating supply. Among these, 381 million came from monthly unlocks for contributors and investors, 300 million were released from the team reserve wallet, and approximately 39 million were newly minted as staking rewards.

Monthly unlocks are known calendar events, while team reserve releases and staking minting will alter the actual supply trajectory. Discrepancies remain between monitoring data and traceable on-chain flows. Therefore, the original article estimates the net supply increase for the next 90 days to be approximately 8.48%, while maintaining a note on differences in monitoring methodologies.

Staking Rewards Increase the Total STRK Supply

Starknet pays staking rewards with newly minted STRK. During the statistical period of the original article, the STRK contract on Ethereum minted approximately 1.3 million tokens every three days and sent them to the staking reward contract, occurring 30 times in 90 days for a total of 39 million new tokens. The total STRK supply thus increased from approximately 10.135 billion to 10.174 billion.

Monthly unlocks themselves do not increase the total supply but only move locked tokens into a freely tradable state; staking rewards directly increase the total supply. Over the past 90 days, the scale of monthly unlocks was about ten times that of staking issuance, but both increase the supply the market must absorb.

Fees Have Utility, but No Burning or Buyback

STRK is already the fee token for the Starknet network, providing it with a real use case. Network fees are currently not burned, nor are they used to buy back STRK, meaning there is no fixed contraction mechanism corresponding to newly added or released tokens.

Future supply is determined by three components: fixed unlocks continuing until March 15, 2027; team reserve releases without a publicly fixed schedule; and new minting that varies with staking scale. The former can be calculated in advance, while the latter two depend on wallet actions and changes in governance parameters.

Continuous Monthly Unlocks Over the Next Three Months

Up to 127 million STRK tokens are expected to unlock on October 15, November 15, and December 15, each equivalent to approximately 1.7% of the circulating supply at the time of the original article. The team reserve wallet still holds approximately 950 million STRK, with the last two releases occurring on July 28 and September 10.

Starknet v0.14.4 went live on October 6, adjusting gas weights; another proposal suggests linking the base fee to the STRK price. None of these changes include token burning. Therefore, supply pressure stems simultaneously from fixed unlocks, reserve releases, and staking issuance.

Related Reads

Airwallex Founder "Slaps His Own Face": From Ranting Against Stablecoins to Building His Own Chain

Jack Zhang, co-founder and CEO of Airwallex, publicly dismissed stablecoins in June 2025, saying they could not lower costs and that crypto had solved no real-world problems in 15 years. Just 17 months later, he co-founded Alta, a blockchain company that raised $80 million to build an L1 designed specifically for licensed financial institutions. The turnaround reflects a broader industry shift: Stripe acquired Bridge for $1.1 billion and incubated Tempo, Circle launched Arc, and stablecoin market cap doubled past $300 billion as U.S. and Hong Kong regulation brought institutional certainty. Alta, co-founded with Loong Wang and Catherine Porter, uses recursive zero-knowledge cryptography to combine compliance with privacy, embedding programmable compliance and identity directly into the protocol—features general-purpose chains cannot offer banks. Airwallex serves as Alta's distribution channel, connecting the chain to its network across 200 countries and 90 currencies. Zhang's stated logic is consistent: he never endorsed crypto broadly, but now sees stablecoins as the on-chain form of fiat currency and a new settlement layer. The real battle is over who controls stablecoin settlement infrastructure, with Stripe, Circle, Fireblocks, and Canton Network all racing to build chain-level rails. Zhang's shift is less a conversion than a commercial necessity: as stablecoin settlement becomes standard for enterprise payments, payment companies that do not own the settlement layer risk becoming application-layer interfaces running on someone else's chain.

marsbit1h ago

Airwallex Founder "Slaps His Own Face": From Ranting Against Stablecoins to Building His Own Chain

marsbit1h ago

The Art of the Trade: How They Made Money Arbitraging USDE Stock

Starkiller Capital outlines a arbitrage trade on StablecoinX (Nasdaq: USDE), a crypto treasury company holding about 20% of ENA tokens. At Friday's close, USDE traded at $6.43 per share, with each share backed by 125.48 ENA worth $20.10—meaning investors pay 32 cents per $1 of ENA, or $155M for $485M in tokens. The firm's approach: buy USDE stock while shorting an equal dollar amount of ENA, so if both fall but the stock falls less, short gains offset stock losses. Starkiller first entered in August at 0.18x mNAV, took profits as the discount narrowed to 0.60x, and has rebuilt a larger position at 0.32x. The key new catalyst: Ethena's foundation waived all token lockups effective October 5, letting the company sell ENA to fund an approved share buyback. At current prices, every $1 of ENA sold repurchases shares representing $3.13 of token assets, strongly accretionary to NAV per share. The discount widened recently not because of company fundamentals but because crypto market beta fell—ENA dropped 29% and比特币 7% in October—and because USDE holders exiting ENA exposure sold the illiquid stock; short interest rose 78% to ~15% of float. Risks include: no redemption mechanism forcing convergence (GBTC traded at a discount for two years), the foundation controls the exit path and must approve token sales, a resale S-1 covers over half the A-shares, mNAV has ranged from 0.13x to 0.61x, warrants at $11.50 and $15 strike could dilute upside, and a crypto crash could push mNAV back to lows. Catalysts to watch: the Q3 10-Q filing (due by Nov 16) for buyback details. Starkiller argues the stock should eventually trade around 0.7x mNAV, implying roughly 56% upside on the long leg just from a return to the October 5 level.

marsbit2h ago

The Art of the Trade: How They Made Money Arbitraging USDE Stock

marsbit2h ago

Embedded chip vendors are starting to sell modules directly?

On November 5, Texas Instruments (TI) announced the TICS100MOD, a multi-protocol industrial communication System-on-Module (SoM) supporting EtherCAT, PROFINET, and EtherNet/IP, with volume availability expected by the end of 2026. The module integrates TI's AM2612 real-time MCU—featuring dual Arm Cortex-R5F cores and PRU-ICSS—along with two DP83826AI 100Mbps Ethernet PHYs, protocol stacks, SDK, and development tools in a 30mm × 18mm package. It connects to an external host MCU or MPU via SPI or QSPI, handling industrial Ethernet protocol processing while the host runs device control and applications. The article notes that while MPU-based SoMs are common, MCU-based SoMs are relatively rare, and TI's MCU is powerful and complex enough to justify a module. TI is not the first chip vendor to offer its own SoM products: Microchip has offered SoMs such as the ATSAMA5D27-SOM1 and SAM9X60D1G-SOM, some suitable for surface-mount reflow soldering, while AMD offers the Kria K26 and K24 SoM series. In wireless connectivity, Espressif and Silicon Labs similarly provide chips alongside modules. In China, SoMs are typically called core boards, with vendors including Chuanglong Technology, MYIR Electronics, Forlinx, ZLG, and Alientek, covering TI, NXP, ST, Renesas, and domestic platforms such as Rockchip and Allwinner. TI continues to partner with third-party SoM vendors—including Chuanglong, MYIR, and Alientek around the AM62L processor, and internationally with Toradex, Variscite, Ezurio, and PHYTEC. TI's new TICS100MOD focuses specifically on industrial Ethernet communication, while customers needing complex computing platforms are still served through third-party partners.

marsbit2h ago

Embedded chip vendors are starting to sell modules directly?

marsbit2h ago

Morgan Stanley's "Memory Grim Reaper" Shawn Kim Warns Again: The Real Bomb Is the "2028 Profit Cliff," Broadly Cuts Memory Stock Price Targets

Morgan Stanley analyst Shawn Kim, dubbed Korea's "semiconductor grim reaper," has issued a new warning on the memory sector, targeting a "2028 profit cliff." In a November 9 report, he cut price targets for Samsung Electronics and SK Hynix, naming Samsung as his new top pick over SK Hynix. The core debate has shifted from near-term fundamentals to the "2028 profit cliff," driven by four pressures: (1) faster-than-expected supply response as new capacity ramps up around late 2027, with Chinese suppliers adding pricing pressure; (2) delayed 2028 data center construction—about 5GW of US projects pushed from 2028 to 2029, cutting global data center capacity growth from ~21% to 9% in 2028; (3) BOM cost constraints, with memory now ~60% of AI server costs, forcing buyers to downgrade configurations, while NAND spot prices fell ~20% in two weeks; and (4) narrowing HBM premiums as competition intensifies. Kim slashed SK Hynix's target price by 38% to 1.6 million won and Samsung's by 19% to 309,000 won, driven by lower long-term margin assumptions—SK Hynix's 2029-2035 DRAM operating margin cut to ~32% (near the 31% historical average) and Samsung's to ~39%, with HBM premiums assumed to fade to zero by 2033. Near-term 2026-2028 EPS estimates were raised on buyback-driven share reduction. Despite the warning, Kim is not fully bearish. The SOX forward P/E has halved from 26.3x to 13.4x since June, roughly matching past downturn troughs, while consensus EPS rose 24%. He argues valuation has done its job; the key variable now is earnings. Of 11 SOX downturns since 2000, nine were valuation-driven, with deep earnings cuts only in 2000-2002 and 2007-2008. AI supply chain EPS revision breadth is narrowing but still positive. Samsung is favored for its size, capital strength, and expected HBM share overtake in 2027 (~39% vs. SK Hynix's ~37%), plus diversified businesses. SK Hynix retains its HBM leadership and long-term agreement protection, though its post-2028 excess profit period is seen as shorter.

marsbit2h ago

Morgan Stanley's "Memory Grim Reaper" Shawn Kim Warns Again: The Real Bomb Is the "2028 Profit Cliff," Broadly Cuts Memory Stock Price Targets

marsbit2h ago

Trading

Spot
活动图片