The crypto market continues its expansion into September 2026. This month’s focus is on projects showing strong on-chain traction, growing fee generation, and alignment with dominant narratives such as DeFi efficiency, institutional infrastructure, and high-performance Layer-1 ecosystems.
Methodology: This list focuses on altcoins with strong fundamentals, favorable market trends, and solid market cap growth. These projects have displayed strong developer ecosystems and the potential to outperform the broader market during the bull run.
| Token | Core Theme | Why It Could Be Undervalued |
|---|---|---|
| Pump.fun (PUMP) | Meme-coin infrastructure | Pump.fun generates substantial protocol revenue, while its programmed buy-and-burn mechanism creates a direct link between platform activity and PUMP supply reduction. |
| Hyperliquid (HYPE) | On-chain trading infrastructure | Hyperliquid combines a high-volume perpetual exchange, HyperEVM and permissionless HIP-3 markets with unusually strong token value capture through HYPE purchases. |
| Solana (SOL) | High-performance Layer 1 | Solana continues improving throughput and latency while expanding beyond memecoins into payments, trading, tokenization and financial applications. |
| Kite (KITE) | AI agent payments | Kite has moved from an AI narrative to a live Layer-1 built specifically for autonomous agents, identity and stablecoin-denominated payments. |
| Ethereum (ETH) | Smart-contract settlement layer | Ethereum’s L1 is becoming cheaper and more scalable while the upcoming Glamsterdam upgrade targets another major increase in execution efficiency and capacity. |
| Morpho (MORPHO) | On-chain lending & credit | Morpho is evolving from crypto lending infrastructure into a broader credit network, with fixed-rate, fixed-term Morpho Midnight now live. |
| Allora (ALLO) | Decentralized AI intelligence | Allora’s production network is expanding beyond simple price forecasts into richer machine-learning outputs and real financial applications. |
| Lighter (LIT) | Decentralized perpetual trading | Lighter combines exchange-grade execution with verifiable order matching, while protocol fees are already being used for LIT buybacks. |
| BNB (BNB) | Exchange-backed Layer 1 ecosystem | BNB Chain continues reducing latency and increasing capacity while BNB remains tied to one of crypto’s largest application and trading ecosystems. |
| Venice Token (VVV) | Tokenized AI compute | VVV combines staking and AI-compute utility with programmatic burns, while its annual emissions are being reduced starting September 1. |
PUMP is the native token of Pump.fun, a Solana-based token-launch and trading platform that lets users create and trade tokens without requiring traditional exchange listings or complicated liquidity setup.
The platform has expanded from its original bonding-curve launchpad into a broader trading ecosystem that includes PumpSwap and other trading products. This makes PUMP an exposure to the economics of one of the largest consumer applications built around on-chain token creation and speculation.
The biggest PUMP catalyst entering September 2026 is its buy-and-burn mechanism. Pump.fun currently directs approximately 50% of platform revenue toward market purchases of PUMP, with purchased tokens then permanently burned. As of August 30, Pump.fun reported roughly $445 million of cumulative PUMP purchases and burns, removing more than 163 billion PUMP, or over 16% of the original 1 trillion-token supply.
The mechanism is backed by an operating business rather than token emissions alone. Pump.fun’s own dashboard showed roughly $432 million in annualized protocol revenue based on its recent 90-day average entering September.
PUMP could be undervalued if the market continues to treat it mainly as another meme-related token rather than as an asset attached to a highly monetized on-chain trading business.
Its thesis is unusually straightforward: more activity on Pump.fun can mean more protocol revenue, and a portion of that revenue is currently converted into PUMP purchases and permanent burns. This creates considerably clearer token value capture than many application tokens whose underlying platforms generate fees without creating direct token demand.
The risk is that Pump.fun’s revenue is highly dependent on speculative token activity. If meme-coin creation and trading cool sharply, revenue and therefore the scale of PUMP purchases could fall. PUMP also does not represent equity or a contractual claim on Pump.fun revenue, so investors should distinguish the buy-and-burn mechanism from conventional shareholder distributions.
HYPE is the native token of Hyperliquid, a purpose-built blockchain for on-chain financial markets.
HyperCore powers Hyperliquid’s order-book-based perpetual futures and spot markets, while HyperEVM allows developers to deploy EVM-compatible applications around the network’s trading infrastructure.
HIP-3 allows builders to launch their own perpetual-futures markets. Deployers must stake HYPE, creating a direct token requirement for teams introducing new markets. Hyperliquid is also expanding into outcome contracts and prediction-style markets, broadening its product range beyond spot and perpetual trading.
Protocol activity can also support HYPE through the Assistance Fund, which uses designated trading fees to acquire HYPE. Token burns and fee-linked demand give HYPE a clearer connection to platform activity than many governance-only tokens.
Hyperliquid is evolving from a decentralized perpetual-futures exchange into a broader on-chain financial platform.
Permissionless markets, spot assets, HyperEVM applications and outcome contracts could allow more financial activity to settle within the same ecosystem. This may increase demand for HYPE through staking, gas, market deployment and fee-related token purchases.
The investment case still depends heavily on sustained trading volume. A prolonged decline in market activity, stronger decentralized-exchange competition or technical problems could weaken demand for HYPE.
SOL is the native token of Solana, a high-performance Layer-1 blockchain used for trading, DeFi, payments, stablecoins, consumer applications and tokenized assets.
SOL is used for transaction fees and staking, connecting the token to network security and blockspace demand.
Solana’s ecosystem is expanding beyond memecoin trading. By June 2026, the value of real-world assets on Solana had crossed $3 billion, while the network continued gaining stablecoin, payment and tokenized-equity activity. MoneyGram also joined the Solana Developer Platform as an infrastructure partner and validator.
Development continues across Agave, Firedancer and Alpenglow. Alpenglow is intended to substantially reduce confirmation times, while validator-client improvements could increase performance and network resilience.
SOL could be undervalued if investors continue to see Solana mainly as a memecoin chain.
Payments, stablecoins, tokenized stocks, real-world assets and institutional infrastructure can create more durable network demand than short-lived speculative trading. Higher application activity can increase demand for blockspace, while staking removes part of the liquid SOL supply.
A recovery in consumer crypto and on-chain trading could provide an additional catalyst, although Solana must continue demonstrating network reliability as its capacity and application complexity increase.
KITE is the native token of Kite, an EVM-compatible Layer-1 blockchain designed specifically for autonomous AI agents, payments and machine-to-machine economic activity.
Kite combines its blockchain with Agent Passport, an authorization layer that gives agents verifiable identities and programmable spending permissions. The goal is to let AI agents transact autonomously while users retain control over budgets, permissions and where funds can be spent.
Kite’s biggest development in 2026 is that its agent-payment thesis has moved onto a functioning mainnet. Kite Mainnet is live with KITE as its native asset, EVM compatibility and infrastructure for stablecoin-denominated payments. Agent Passport is also live as the system through which users can define what an autonomous agent is allowed to do financially.
KITE has direct functions inside this architecture. It is used for staking and network participation, while Kite’s token model allows commissions generated by AI services to be converted from stablecoin revenue into KITE before being distributed within the network.
KITE could be undervalued if autonomous-agent payments develop into a major crypto use case and Kite establishes itself as infrastructure beneath that economy.
AI agents need more than intelligence to operate independently. They need identity, authorization, payments and rules defining what they can spend. Kite is attempting to build these functions at the protocol level rather than adding them later to a general-purpose blockchain.
The key risk is adoption. A purpose-built AI payments chain is valuable only if agents, developers and services actually use it. KITE’s bull case therefore depends on real payment volume and AI-service activity eventually becoming large enough to create meaningful staking and transaction demand for the token.
ETH is the native asset of Ethereum, the largest general-purpose smart-contract settlement network and the base layer for much of DeFi, stablecoins, tokenized assets and Layer-2 infrastructure.
ETH is used to pay transaction fees and secure Ethereum through proof-of-stake. It also serves as collateral and a reserve asset across a large portion of the Ethereum ecosystem.
Ethereum’s next major catalyst is Glamsterdam, currently planned for Q4 2026. The upgrade is designed to improve Layer-1 scalability through changes including Block-Level Access Lists, which prepare Ethereum for greater parallel transaction processing and higher gas limits. It also includes changes aimed at reducing costs for basic ETH transfers.
Preparations are already underway. The Ethereum Foundation launched the Platåberget testnet in August specifically to test the Glamsterdam changes before they progress to longer-lived testnets.
ETH could be undervalued if investors continue to price Ethereum around the old assumption that its base layer is permanently expensive and incapable of scaling.
That assumption is becoming increasingly outdated. Pectra and Fusaka have already changed Ethereum’s cost and scaling profile, and ethereum.org noted in May 2026 that ordinary mainnet transactions had become dramatically cheaper than during the 2021–2023 period. Glamsterdam aims to extend that trend further.
The central investment question remains value capture. Ethereum’s L2-centric scaling strategy can move execution away from L1, potentially reducing direct fee pressure. ETH’s bull case is strongest if Ethereum simultaneously remains the preferred settlement, collateral and security layer for the growing economy built above it.
MORPHO is the governance token associated with Morpho, an open lending infrastructure network that allows applications, institutions and users to lend and borrow assets through noncustodial markets.
Morpho has progressively moved from optimizing existing DeFi lending pools to building its own lending infrastructure, vaults and credit markets. Its integrations increasingly allow fintech and institutional products to use Morpho underneath the interface without requiring end users to interact directly with DeFi.
Morpho Midnight went live in July 2026, giving the ecosystem fixed-rate, fixed-term on-chain credit alongside a new Markets application. Instead of forcing borrowers and lenders into continuously variable rates, Midnight allows the market to express rate, term and risk more directly.
This follows a period of substantial institutional expansion. Morpho has been selected for integrations by companies including Coinbase and Robinhood, and the Morpho Association raised $175 million in June 2026 with participation from investors including Paradigm, a16z crypto and Ribbit, alongside strategic participants.
MORPHO could be undervalued if the market continues to view Morpho mainly as another DeFi lending protocol rather than infrastructure for a much broader on-chain credit market.
Midnight materially expands that addressable market. Fixed rates and fixed maturities are standard components of traditional credit markets, making Morpho’s infrastructure easier to map onto products used by institutions and fintechs.
The key limitation is token value capture. Protocol adoption does not automatically mean proportional MORPHO demand. The investment thesis therefore depends on governance and token economics becoming increasingly valuable as Morpho becomes embedded in more financial products—not simply on deposits and loans rising.
ALLO is the native token of Allora Network, a decentralized intelligence network in which multiple machine-learning models compete and collaborate to generate adaptive predictions for applications.
Instead of forcing a developer to rely on a single model, Allora evaluates contributors according to their forecasting performance and combines their outputs into network-level intelligence. ALLO coordinates incentives between model workers, reputers, validators and users of that intelligence.
Allora’s August 2026 mainnet upgrade significantly broadened the type of intelligence the network can produce. Mainnet v0.17 added classification and multi-output topics rather than restricting every network inference to a single numeric prediction.
Allora also launched its first mainnet volatility topics in August, producing short-horizon volatility forecasts for BTC, ETH, XRP and SOL. This expands the network beyond directional price prediction into outputs that can potentially be used for risk management, derivatives and automated portfolio decisions.
ALLO could be undervalued if decentralized intelligence develops into an infrastructure layer used by DeFi applications and autonomous agents rather than remaining a standalone AI narrative.
The important distinction is that Allora is not trying to build another chatbot. It is creating machine-readable intelligence that applications can consume automatically. More complex mainnet outputs make the network usable for a broader range of prediction and decision-making problems.
ALLO’s smaller valuation relative to established AI and infrastructure tokens can offer greater upside, but also greater execution risk. Its bull case depends on applications paying for and acting on Allora’s intelligence at scale, creating sustainable demand for the participants that ALLO incentivizes.
LIT is the native infrastructure token of Lighter, a decentralized trading platform built as a custom zero-knowledge rollup on Ethereum.
Lighter is designed to combine centralized-exchange-style performance with verifiable execution. Order matching and liquidations can be cryptographically verified, while the exchange offers zero-fee trading for standard retail accounts and more specialized infrastructure for professional participants.
LIT now has more explicit value capture from the exchange. Lighter uses trading-fee revenue to buy LIT through market buybacks, while staking LIT unlocks access to parts of the ecosystem such as the Lighter Liquidity Pool. One LIT staked currently allows up to 10 USDC of LLP deposits.
The protocol is already generating measurable revenue. DeFiLlama data through August 2026 showed Lighter using millions of dollars in 2026 protocol earnings for LIT purchases, giving the token an economic link to actual exchange activity rather than relying entirely on governance utility.
LIT could be undervalued if decentralized perpetual exchanges continue taking market share from centralized venues and Lighter establishes itself as one of Hyperliquid’s strongest competitors.
Its proposition combines two characteristics that traders typically have to compromise between: high-performance execution and on-chain verifiability. The addition of staking utility and fee-funded token purchases gives LIT a clearer connection to the exchange’s growth.
The competitive risk is substantial. Hyperliquid already has enormous liquidity and network effects, while other perp DEXs compete for the same traders. Lighter therefore needs to convert technical performance into sticky volume, and investors also need to account for the difference between its circulating supply and fully diluted supply.
BNB is the native asset of the BNB Chain ecosystem and is used for transaction fees, staking and network participation across BNB Smart Chain.
Its utility is also connected to the wider Binance ecosystem, giving BNB exposure to both an active Layer-1 network and one of crypto’s largest trading and application ecosystems.
BNB Chain has spent 2026 aggressively improving network performance. Its Fermi upgrade reduced block times to roughly 0.45 seconds, while the Pasteur hard fork went live on BSC mainnet on August 25, 2026.
Pasteur improves bridge and validator security while increasing how much transaction activity can fit inside each block. BNB Chain reported testnet benchmarks increasing from roughly 1,237 to 2,324 transactions per second through the relevant block-capacity changes.
BNB could be undervalued if the market continues to treat it primarily as an exchange token while BNB Chain develops into faster infrastructure for trading, payments and consumer applications.
Unlike many Layer-1 assets, BNB already has an established user base, liquidity network and application ecosystem. Faster block production and higher capacity can make that existing distribution more valuable rather than requiring the network to build demand from zero.
Its main risk is concentration. BNB’s perception and ecosystem remain closely connected to Binance, so regulatory, competitive or operational developments surrounding that broader ecosystem can influence BNB even when the underlying blockchain is performing well.
VVV is the native token of Venice, a private AI platform that provides text, image, code and model inference through consumer products and an API.
VVV is designed to turn access to AI compute into an on-chain economic resource. Holders can stake VVV and lock staked VVV to mint DIEM, a transferable token representing recurring Venice API capacity. One staked DIEM currently provides $1 per day of API credit.
VVV has one of the most timely catalysts on this list because a tokenomics change takes effect on September 1, 2026. Venice is reducing annual VVV emissions from 3 million to 2.5 million VVV on September 1, followed by another reduction to 2 million annually on October 1.
At the same time, Venice has been expanding its programmatic burn mechanisms. New subscriptions already trigger VVV purchases and burns, and Venice introduced an additional mechanism under which $5 for every $100 of qualifying API credits purchased is directed toward buying and burning VVV.
VVV could be undervalued if investors begin valuing it less like a conventional AI token and more like an asset attached to consumable AI-compute capacity.
Its token economics now combine several forces: VVV can be staked, staked tokens can be locked to create DIEM compute capacity, paid product activity can trigger market purchases and burns, and annual emissions are being reduced. That creates a more direct relationship between Venice adoption and VVV’s supply-demand structure than exists for many AI tokens.
The bull case ultimately depends on AI demand. Burns and lower emissions improve tokenomics, but they cannot replace product adoption. VVV becomes substantially more compelling if Venice’s API usage, paid subscriptions and demand for tokenized compute grow fast enough that real AI consumption (not token incentives) is the primary driver of the ecosystem.
