Ethereum co-founder Vitalik Buterin shared a new proposal over the weekend that would radically overhaul the system that powers its smart contracts.

Buterin’s suggestion, which he posted on Ethereum’s primary developer forum, involves replacing the Ethereum Virtual Machine, the software engine that powers programs on the network, with RISC-V, a popular open-source framework that offers built-in encryption and other benefits. .

The EVM is a key piece of Ethereum’s underlying design and has been seen as one of the main elements that helped the network succeed in a crowded field of other blockchains. Many non-Ethereum networks have used the EVM to build their own chains, as has a growing ecosystem of layer-2 networks built atop Ethereum, including Coinbase’s Base chain.

The EVM has long played an essential role in Ethereum’s development. Other chains that use it can seamlessly connect with apps on Ethereum, and developers on EVM-based networks can transition more smoothly to building applications directly within the Ethereum ecosystem.

Buterin argued that transitioning Ethereum to a RISC-V architecture will “greatly improve the efficiency of the Ethereum execution layer, resolving one of the primary scaling bottlenecks, and can also greatly improve the execution layer’s simplicity.” (The execution layer is the part of the network that reads smart contracts.)

The RISC-V architecture, which has seen limited adoption in other blockchain ecosystems, like Polkadot, could offer «efficiency gains over 100x» for certain kinds of applications, according to Buterin. These improvements could reduce the network’s costs — long seen as a major barrier to adoption.

Among the primary benefits of RISC-V is its native support for certain kinds of encryption. Transitioning to the new architecture could, in Buterin’s view, be a simpler alternative to the community’s current plan, which involves rebuilding the EVM around zero-knowledge cryptography.

Buterin’s proposal is something developers would tackle over the long term, comparable to projects like the Beam Chain, which is looking to revamp Ethereum’s consensus layer.

The RISC-V comes at a time of broader soul-searching for the Ethereum community. Recently, transaction volumes have declined, and Ethereum’s token has lagged behind the broader market.

Earlier this year, the Ethereum Foundation, the primary non-profit that supports the development of the broader Ethereum ecosystem, underwent a leadership transition in an attempt to remedy the impression among community members that the ecosystem lacked a clear roadmap and was losing its lead compared to competitors.

Read more: Top Ethereum Researcher’s Dramatic Proposal Draws Standing-Room-Only Crowd in Bangkok

The AI revolution in trading should be a game-changer, but instead, it’s become a quick money grab. Everywhere you turn, yet another ChatGPT wrapper is being marketed as the next big thing for crypto traders. The promises? “AI-powered insights,” “next-gen trading signals,” “perfect agentic trading.” The reality? Overhyped, overpriced, and underperforming vaporware that doesn’t scratch the surface of what’s truly needed.

Saad Naja is a speaker at the AI Summit during Consensus 2025, Toronto, May 14-16.

AI should be designed to augment the trader experience, not sideline it. Companies like Spectral Labs and Creator.Bid are innovating with AI agents but risk heading toward vaporware status if they fail to deliver real utility beyond surface-level GPT wrappers. They have an overreliance on Large Language Models (LLMs) like ChatGPT without offering any unique utility, prioritizing AI buzzwords over substance and AI architecture transparency.

AI Agents Should Augment Trading

Combining AI and trading is a transformative leap, for humans to make trading gains more effectively with powerful foresight, investing less time, but not to replace humans from the trading equation entirely. Traders don’t need another emotionless agent with unfettered agency. They need tools that help them trade better, faster, and more confidently in environments that simulate real market volatility before going trading in the real markets.

Too many GPT wrappers rush to market with fluffy, half-baked agents that prey on fear, confusion, and FOMO. With barely-trained Large Language Models (LLMs) and little transparency, some of these AI trading “solutions” reinforce set and forget bad habits.

Trading isn’t just about hyper speed or automation, it’s about thoughtful decision-making. It’s about balancing science with intuition, data with emotion. In this first wave of agent design, what’s missing is the art of the trader’s journey: their skill progression, unique strategy development, and fast evolution through interactive mentorship and simulations.

Just Fancy Calculators

The real innovation lies in developing a meta-model that blends predictive trading LLMs, real-time APIs, sentiment analysis, and on-chain data, while filtering through the chaos of Crypto Twitter.

Emotion and sentiment do move markets. If your AI Trader agent can’t detect when a community flips bullish or bearish, or front-run that signal, it’s a non-starter.

GPT Wrappers rejecting emotion-driven market moves offer lower-risk, lower-reward gains within portfolio optimization. A better agent reads nuance, tone, and psycholinguistics, just as skilled traders do.

And while 20 years of high-quality trading data spanning multiple cycles, markets and instruments is a great start, true mastery comes through engagement and progression loops that stick. The best agents learn from data, people and thrive with coaching.

Better to Lose Pretend Money

Financial systems intimidate most people. Many never start, or blow up fast. Simulated environments help fix that. The thrill of winning, the pain of losing, and the joy of bouncing back are what build resilience and shift gears from sterile chat and voice interfaces.

AI Trader agents should teach this, back-test and simulate trading comeback strategies in virtual trading environments, not just of successful trades but comebacks from the unforeseen events. Think of it like learning to drive: real growth comes from time on the road and close calls, not just reading your state’s handbook.

Simulations can show traders how to spot candlestick patterns, manage risk, adapt to volatility, or respond to new tariff headlines, without losing their heads in the process. By learning through agents, traders can refine strategies and own their positions, win or lose.

Before My Bags, Win My Trust

AI Agents’ life-like responses are fast improving to being indistinguishable from human responses through conversational and contextual depth (closing the “Uncanny Valley” gap). But for traders to accept and trust these agents, they need to feel real, be interactive, intelligent, and relatable.

Agents with personality, ones that vibe like real traders, whether cautious portfolio managers or cautious portfolio optimizers can become trusted copilots. The key to this trust is control. Traders must have the right to refuse or approve the AI Agent’s calls.

On-demand chat access is another lever, alongside visibility of trading gains and comebacks built on the sweat and tears of real traders. The best agents won’t just execute trades, they’ll explain why. They’ll evolve with the trader. They’ll earn access to manage funds only after proving themselves, like interns earning a seat on the trading desk.

Fun, slick AAA aesthetics and progression will keep traders coming back in shared experiences opposed to solo missions. Through tokenization and co-learning models, AI agents could become not just tools, but co-owned assets — solving crypto’s trader liquidity problem along the way.

First-to-market players must be viewed with healthy skepticism. If Trader AI Agents are going to make a real impact, they must move beyond sterile chat interfaces and become dynamic, educational, and emotionally intelligent.

Until then, GPT wrappers remain what they are slick distractions dressed up as innovation, extracting more value from users than they deliver, as the AI token market correction indicated.

The convergence of AI and crypto should empower traders. With the right incentives and a trader-first mindset, AI Agents could unlock unprecedented learnings and earnings. Not by replacing the trader but by evolving them.

Despite its growing footprint as a major corporate holder of bitcoin (BTC), Strategy’s large-scale purchases of the cryptocurrency appear to have little, if any, influence on its price, according to a research paper by TD Cowen.

The findings published Monday challenge a popular theory among skeptics — that Strategy’s aggressive buying spree is helping prop up bitcoin’s value, and that without its continued demand, prices would falter. But based on the data, that argument doesn’t hold much weight, the analysts said.

A Big Buyer, But a Small Slice of the Market

Strategy recently issued another 1.8 million shares under its at-the-market (ATM) offering, raising an additional $842 million in net proceeds. The funds were used to purchase 6,556 bitcoins, boosting the firm’s bitcoin yield this quarter by 1% to 12.1%. However, when measured against the broader bitcoin market, these purchases are just a drop in the bucket.

According to the TD Cowen analysis, Strategy’s bitcoin buys have typically accounted for just 3.3% of weekly trading volume on average. Over the past 27 weeks, the company’s total activity amounted to 8.4% of volume — but this figure was skewed by a handful of weeks where its buying briefly surged past 20%. In eight of those weeks, Strategy didn’t buy any bitcoin at all.

“Our conclusion is that in most periods, it doesn’t appear plausible that Strategy’s purchases could have had a sustained, material impact on the price of bitcoin,” TD Cowen analysts wrote.

Correlation? Not Much.

The analysis further tested the relationship between Strategy’s bitcoin purchases and market prices — and found it to be statistically weak. The correlation coefficient between Strategy’s weekly bitcoin buy volume and BTC price at week’s end came in at just 25%. When comparing purchases to weekly price changes, the correlation rose only slightly to 28%.

Given a correlation coefficient close to 0 suggests no or weak correlation, these results indicate little to no link between Strategy’s actions and short-term market movements — let alone any kind of sustained price influence, the paper said.

What About Outpacing Miners?

Another common critique is that Strategy frequently purchases more bitcoin than is mined in a given period, implying it’s creating upward price pressure. While technically true, the analysis shows this argument misunderstands how the bitcoin market works.

Over the past six months, secondary bitcoin trading has outpaced mining volume by nearly 20 times. Even removing Strategy’s purchases from the equation, secondary market activity still exceeds new supply by 17 times. In that environment, miners and buyers alike are price takers — not setters.

“As we have seen, its purchases represent a very small percentage of total bitcoin trading volume; thus the idea that it is somehow having a profound or even notable impact on bitcoin price action seems incongruous, to us,” TD Cowen said.

Building Value, Not Hype

While Strategy’s influence on the bitcoin market may be overstated, the value it’s generated for shareholders is harder to ignore.

Last week’s purchases created an estimated incremental gain of 5,281 bitcoins, bringing quarter-to-date gains to nearly $600 million. Since the beginning of 2023, Strategy has increased its bitcoin holdings by 306%, while only expanding its fully diluted share count by 94% — a strong showing for a company using bitcoin as a strategic treasury asset.

With $1.53 billion in remaining ATM capacity and board approval for a larger share authorization, Strategy is well-positioned to continue this strategy — without disrupting the very market it’s betting on.

“We expect Strategy will continue to drive positive BTC Yield for the foreseeable future. While BTC Yield will likely fall to the extent bitcoin continues to rise in price, the dollar value of incremental gains from Strategy’s Treasury Operations could remain highly advantageous to shareholders,” the analysts wrote.

Disclaimer: Parts of this article were generated with the assistance from AI tools and reviewed by our editorial team to ensure accuracy and adherence to our standards. For more information, see CoinDesk’s full AI Policy.