Solana’s Anatoly Yakovenko Says Agentic Coding Is Changing Everything

Speaking at TechCrunch Disrupt, Yakovenko described how “agentic coding” — AI systems that can reason, plan, and write code with minimal human input — has quietly redefined how he works. 

“AI has been a great force multiplier for somebody who’s an expert,” he said. “Now I can just watch Claude churning through its thing and I can almost smell when it’s going off the rails.”

He wasn’t joking about the obsession. “If people are in a meeting with me and I’m not paying attention,” he added, “it’s because I’m watching Claude.”

Yakovenko’s comments highlight a growing shift among developers who are moving from automation to autonomy, letting AI not just complete tasks, but decide how to do them. 

It’s the next leap in human–machine collaboration, one that’s already redefining how codebases evolve, bugs get fixed, and infrastructure runs itself.

Wall Street’s big money is already moving — quietly building positions in a handful of stocks before the next rally.

Our analysts tracked the flows and found 10 companies leading the charge.

Some are household names. Others are under-the-radar innovators about to break out.

Together, they form the Post-Rate-Cut Playbook smart investors are following right now.

See Wall Street’s Top 10 Stocks for Fall — free before it moves behind the paywall >>

Agentic AI isn’t just a buzzword. Across the broader IT world, it’s beginning to function like an invisible team of specialists. 

Instead of merely suggesting code snippets, these AI systems reason through entire workflows: spinning up servers, allocating compute, patching vulnerabilities, and even testing deployments in real time.

A recent CIO analysis described how enterprise teams are already using such agents to autonomously resolve issues, streamline software testing, and even detect and repair outages before human engineers notice. 

In some cases, the technology has cut problem-resolution times from hours to seconds.

The takeaway: what Yakovenko sees inside Solana’s developer tools mirrors what’s happening in Fortune 500 IT stacks. The same agentic logic running crypto infrastructure is also starting to govern data centers, cloud orchestration, and financial systems.

Solana’s comeback has been powered as much by narrative as by numbers. Co-founder Anatoly Yakovenko’s open enthusiasm for AI-driven coding tools, particularly agentic systems that can write, test, and optimize code autonomously, has become a symbol of how Solana wants to scale—fast, flexible, and human-augmented.

That ethos has translated into a broader developer revival. On-chain activity has climbed, new projects are launching on compressed NFT and DeFi rails, and Solana’s network throughput continues to outpace rivals. With trading platforms and consumer apps now stacking atop its high-speed architecture, the ecosystem feels less like a speculative playground and more like a production environment for next-generation financial infrastructure.

At the same time, the network’s cultural capital has rebounded. From gaming to tokenized loyalty programs, Solana’s developer base is again expanding into use cases that reach real users. And behind the scenes, the system’s efficiency upgrades—lower fees, faster block times, and stable uptime—are reinforcing confidence among both retail and institutional participants.

In short, Solana has regained its story: a fast, developer-led alternative to Ethereum that now intersects with the frontier of AI and automation.

If you own ZERO of the Next Magnificent Seven stocks.

Original Mag Seven turned $7,000 into $1.18 million.

But these seven AI stocks could do it in 6 years (not 20).

Now, the man who called Nvidia in 2005 is revealing details on all seven for FREE.

Find Out More Now Before It's Too Late.

That optimism is finding support in the numbers. On-chain data show older Solana wallets beginning to move coins back into circulation, a pattern that once signaled looming sell pressure. But this time, there’s a new kind of buyer waiting.

Institutional products linked to Solana have quietly absorbed the excess supply. ETFs such as Bitwise’s Solana Staking Fund (BSOL) and Grayscale’s GSOL have taken in roughly $117 million in recent inflows, offsetting distribution from early holders. That handoff—long-term wallets selling while ETFs accumulate—has reshaped the market’s base layer of demand.

The result is a more stable price structure. Even as volatility hit broader crypto markets, SOL has held near the $180–$200 range. Derivatives open interest has expanded, ETF flows have deepened, and liquidity has grown without the sharp drawdowns that typically follow whale activity.

For the first time, Solana’s retail, developer, and institutional layers are aligned. Innovation drives attention, regulated products absorb capital, and the network’s fundamentals tighten around a more durable floor. The next move higher will depend less on hype and more on whether that new ecosystem can sustain its balance between creation and consolidation.

Agentic AI could soon blur the line between blockchain development and enterprise IT. In crypto, it means networks that self-heal and optimize on-chain performance. In corporate tech, it means infrastructures that adjust capacity, patch security holes, and manage incidents without human triage.

Both worlds are converging on the same idea: intelligent automation that doesn’t wait for a ticket.

Yakovenko’s fascination with Claude may sound like a quirk, but it’s really a glimpse of a broader shift, one where developers spend less time typing and more time supervising autonomous systems.

As he put it, the best engineers in this new world may not be the fastest coders, but the ones who can “smell when it’s going off the rails.”

Yakovenko isn’t just watching AI write code; he’s watching a new software paradigm emerge.

The agentic systems transforming Solana’s developer culture are part of the same revolution now taking root across global IT, one where machines don’t just execute instructions but decide what needs to be done.

The line between coder and code is beginning to fade.