In September 2026, Cardano founder Charles Hoskinson made a widely consequential assessment for the crypto industry during an broadcast: AI programming tools have destroyed the usefulness of developer counts as a growth metric for blockchain ecosystems. In his words, people without technical knowledge and experience can now use AI tools to "look like a developer," so the number of developers is no longer "a scarce and limited resource."
The immediate backdrop to this statement is that Midnight, a privacy-focused sidechain project within the Cardano ecosystem, cut part of its developer relations team and shifted its focus to "builders," as Hoskinson described. But the deeper issue points to valuation inertia across the entire industry—a habit that’s been in place for nearly a decade: measuring a chain’s "growth" using code submission volume, developer activity, and the number of hackers participating in hackathons. When AI can generate working software for anyone who doesn’t understand code logic, these metrics’ underlying informational value is being systematically diluted.
For ADA holders and traders following the Cardano ecosystem, this raises a more pragmatic question: if developer count is no longer an effective growth signal, which metrics should be included in the observation framework?
The Failure Mechanism of Developer Metrics: Scarcity Disappears and Verification Costs Surge
To understand Hoskinson’s judgment, you first need to know why developer counts were "effective" over the past decade.
In the early days of the blockchain industry, engineers who could write in Haskell or Rust were scarce resources. If a chain had more core developers, it typically meant faster protocol iteration, richer toolchain development, and stronger ecosystem pull. That logic depended on a key premise: human programming capability was constrained and difficult to replicate quickly.
The widespread adoption of AI programming assistants fundamentally changes that premise. As Hoskinson described, in current code bounties and hackathon events, participants have already been submitting AI-generated code as if it were original work. As a result, teams must spend time verifying the author’s identity instead of using resources to attract genuinely real projects. This means the developer-count metric faces not only "inflation," but a structural rise in verification costs: when everyone who submits code could be an "agent" of AI, the cost of distinguishing real development ability from tool-assisted output becomes extremely high.
One detail worth noting is that while Hoskinson criticized developer metrics, Midnight—the fund within his ecosystem—continued to run three hackathons, with a total prize pool of $26,500, including a Korea event that started on September 1, 2026. Hoskinson’s explanation was that he and the foundation hold different views. This kind of "split between words and actions" is itself a data point worth recording. It suggests that within the Cardano ecosystem, consensus on growth metrics has not yet formed, and inertia behind the old metrics still persists.
What Signals Have On-Chain Metrics Already Given?
Hoskinson’s "declaration" creates an opportunity to re-examine Cardano’s on-chain data. Based on verifiable information, the ADA ecosystem’s on-chain activity shows a set of concerning signals.
Fluctuations in active addresses. According to DeFiLlama data, Cardano’s weekly active addresses fell from 133,966 to 83,899 between August 31 and September 5, 2026, a decline of 37.4%. Over the same period, weekly transaction counts dropped from 174,025 to 108,997, also by 37.4%. A near 40% drop in activity in a single week—whether or not seasonal factors are involved—points to one fact: Cardano has not established a stable organic demand baseline on-chain.
Structural bottlenecks in stablecoin supply. As of September 7, 2026, Cardano network DeFi total value locked (TVL) is about $65.33 million, of which stablecoins account for $64.53 million. Stablecoins are nearly equivalent to DeFi liquidity itself. That means Cardano’s DeFi ecosystem relies almost entirely on stablecoins as both a pricing and trading medium, with little liquidity buffering from native assets or derivatives. A deeper problem is the absolute size of stablecoins: the $64.53 million stablecoin supply during the same period represents less than 0.5% of Solana stablecoin circulation (about $15 billion), and far below Avalanche’s $1.4 billion level.
A divergence between transaction fees and on-chain revenue. July 2026 data shows that Cardano’s DeFi protocol application-layer fees fell by 67.1% over a 30-day period. Meanwhile, on-chain Gas fees only fell by 35.7% over the same timeframe. Fees at the application layer are dropping faster than base transaction costs, which means users are leaving DeFi applications rather than merely reducing trading frequency. When stablecoin supply is insufficient, lending, market making, and leveraged trading can’t scale. Fee income shrinking is a natural result of liquidity shortages.
Taken together, these data point to a conclusion: Cardano’s on-chain economy has not yet built a self-sustaining flywheel. Stablecoins are DeFi’s operating capital. Without deep pools of USD liquidity, any "ecosystem growth" narrative lacks underlying support.
An Alternative Framework: From "How Many People Write Code" to "How Many People Pay for Value"
If developer counts are no longer reliable, how should ADA ecosystem growth be observed? Based on Hoskinson’s own framing and what the on-chain data suggests, the following four dimensions form a verifiable alternative framework.
First, the absolute level of stablecoin supply and its growth slope. This is the most critical leading indicator for Cardano DeFi. Changes in stablecoin market value directly determine available liquidity for lending protocols, DEX depth, and the willingness of market makers to participate. As of early September 2026, the $64.53 million level is a significant structural weakness for a smart contract platform ranked in the top 30 by market cap. Observing whether stablecoin supply is trending upward matters more than focusing on short-term fluctuations in ADA price.
Second, paid users—not just "user counts." Hoskinson has been explicit that he values projects with paying users and external investors. This standard can be translated into on-chain observables: the number of wallets generating real fee revenue in DeFi protocols; the number of addresses completing purchases in NFT marketplaces rather than merely minting; and the net inflow asset size across cross-chain bridges. These indicators are harder for AI or bots to fake because they directly correlate with capital flows.
Third, on-chain transaction density and capital turnover. Transaction counts themselves may be boosted by low-value interactions. A more meaningful approach is to track average value per transaction, the share of DEX volume relative to TVL, and the turnover speed of stablecoins across protocols. If stablecoin supply grows but turnover stalls, it suggests liquidity remains in a "hold" state rather than entering a "use" state.
Fourth, alternative validation signals for developer quality. If code submission volume is no longer trustworthy, you can look at indirect indicators: the number and quality of audit reports, the proportion of mainnet-deployed contracts that have undergone formal verification, and the number of cases where ecosystem projects obtained external funding. Due diligence by external investment institutions, to some extent, performs the function of validating developer quality.
ADA’s Current Market Conditions and Comparison with Ecosystem Indicators
As of September 14, 2026, according to Gate market data, ADA’s trading price is around $0.20931, with a market cap of approximately $7.643 billion. 24-hour trading volume is about $3.8255 million, and market sentiment is neutral. Price changes are -6.18% over the past 7 days, +17.21% over the past 30 days, and -76.69% over the past year.
Source: Gate market data
There is a notable divergence between these price figures and the on-chain indicators. ADA recorded positive returns over the past 30 days. The price rebound is more likely driven by broad-market beta, short-term capital rotation, or position adjustments in a low-liquidity environment—not by an improvement in ecosystem fundamentals.
On a longer timescale, ADA’s current price is still about 93% below its historical high of $3.09 (September 2021). Its market cap ranking has fallen from the top ten in 2021 to 27th today. This relative decline aligns with the long-term stagnation seen in on-chain metrics such as stablecoin supply and DeFi TVL.
Conclusion
Hoskinson’s assessment that "AI destroys developer metrics" isn’t valuable because it announces the death of a single metric. Its real value lies in exposing how fragile the valuation anchor the crypto industry has leaned on for years is. For the ADA ecosystem, the real challenge is not finding a new number to replace "developer count," but answering a more fundamental question: does Cardano’s on-chain economy have the structural conditions needed to generate sustainable fee revenue?
Stablecoin supply depth, the presence of paying users, and liquidity turnover efficiency will not be distorted simply because AI programming tools have become common. These metrics directly tie to capital flows and real demand. Until there are substantive improvements in these indicators within the ADA ecosystem, any narrative about "growth" should be treated with skepticism.
FAQ
Q1: Hoskinson said the developer-count metric is no longer valid. So how many developers are building on Cardano today?
Hoskinson did not provide a specific estimate for developer numbers. His core argument is that "the number itself is no longer credible." Instead of focusing on an un-verifiable figure, you should track the number of audited contracts deployed to the mainnet and real ecosystem project case studies that received external funding.
Q2: ADA’s current price is decoupled from on-chain indicators. Which should be prioritized?
On-chain indicators better reflect the ecosystem’s real usage. ADA’s price is influenced by broader market sentiment, ETF flows, and derivatives markets. Short-term price volatility doesn’t necessarily track ecosystem fundamentals. Stablecoin supply and paying-user data are more stable observation anchors.
Q3: Cardano’s DeFi TVL is only a little over $60 million. Does that scale mean the ecosystem has already failed?
A small TVL is a factual observation, but "failure" is a value judgment. A more constructive interpretation is that Cardano’s DeFi ecosystem is still in an early stage, and the structural shortage in stablecoin supply is the primary bottleneck limiting scale. Whether this bottleneck gets resolved is the most worth tracking variable over the next few quarters.
Q4: Midnight cut its developer relations team and shifted to "builders." What does that mean?
It signals a strategy adjustment within the Cardano ecosystem: shifting from "recruiting more people to write code" to "supporting projects that have already generated revenue or users." The logic behind the shift is that after AI lowers the coding barrier, what’s scarce is no longer code-production capacity, but productization capability and user acquisition capability.
