Nvidia (NVDA) is expected to unveil a new AI-optimized CPU at its GTC Technology Conference keynote today, likely the Vera CPU tailored for agentic AI workloads, following a major deployment deal with Meta Platforms to scale Grace CPUs and upcoming Vera servers in 2027.

Nvidia is pivoting from a pure GPU monopoly to a full-stack AI infrastructure platform by entering the CPU market for agentic systems that require low-latency sequential execution rather than massive parallel compute, potentially unlocking a new revenue stream and justifying premium valuations.

A recent study identified one single habit that doubled Americans’ retirement savings and moved retirement from dream, to reality. Read more here.

Nvidia (NASDAQ:NVDA) is hosting its annual GTC Technology Conference this week in San Jose, a four-day gathering that draws thousands of developers, engineers, and tech leaders to dive deep into AI, accelerated computing, and graphics innovations.

The event kicks off today at 2 p.m. EST with CEO Jensen Huang’s highly anticipated keynote address. Huang has teased the unveiling of “a chip that will surprise the world,” building massive buzz around what could be a game-changing reveal.

Whether it’s a next-gen GPU, a breakthrough in inference tech, or something entirely new, the announcement carries heavy weight: In an era where Nvidia’s stock has lost momentum amid AI hype fatigue, this moment could reignite investor enthusiasm and propel shares out of their prolonged slump.

Read: Data Shows One Habit Doubles American’s Savings And Boosts Retirement

Most Americans drastically underestimate how much they need to retire and overestimate how prepared they are. But data shows that people with one habit have more than double the savings of those who don’t.

Speculation has run wild in the weeks leading up to GTC. Some analysts expect details on the Rubin GPU architecture (successor to Blackwell) or even integration of Groq’s language processing units for faster inference. Others whisper about ARM-based chips for Windows laptops or advances in physical AI and robotics. Yet the theory gaining the most traction -- and the one I consider most likely -- is a major push into AI-optimized CPUs tailored for the emerging agentic AI era.

This isn’t wild conjecture. Just last month, Nvidia inked a massive multiyear deal with Meta Platforms (NASDAQ:META) to deploy millions of its Grace CPUs in standalone configurations, alongside Blackwell and upcoming Rubin GPUs. The partnership explicitly includes a roadmap for Vera CPU-only servers in 2027, signaling Meta’s bet on discrete CPUs for scaling AI workloads.

CNBC and other reports highlight how agentic AI -- autonomous systems that plan, reason, and act sequentially -- makes traditional GPUs less ideal as the primary engine. CPUs are becoming the bottleneck for orchestration, data processing, and inference tasks that demand low-latency, sequential execution rather than massive parallel compute.

I believe the surprise chip is either the long-teased Vera CPU itself or a new AI-native CPU variant optimized for these agentic workflows. Huang has repeatedly emphasized the shift toward “agentic systems” in recent interviews, and the timing after the Meta deal feels deliberate. Revealing a production-ready, high-performance CPU would position Nvidia to capture an entirely new slice of the data-center market, one that extends far beyond its GPU dominance.

The implications for Nvidia would be profound. GPUs have driven explosive growth, but the agentic era could spark a “CPU supercycle” as hyperscalers like Meta, Google, and Amazon (NASDAQ:AMZN) seek balanced infrastructure for increasingly complex AI agents. A successful launch could diversify revenue streams, reduce dependency on GPU cycles, and cement Nvidia as the full-stack AI infrastructure leader.

Early adoption by Meta alone validates the technology, potentially accelerating enterprise uptake and boosting long-term margins through software ecosystems like CUDA extensions for CPUs.

Not everyone is convinced the reveal will move the needle. UBS analyst Timothy Arcuri recently noted that “it is hard to see Nvidia being able to provide thesis-altering commentary that creates a breakout for the stock,” even with major announcements. The bar is sky-high after years of blockbuster GPU launches and investors have grown accustomed to trillion-dollar valuations built on AI hype. UBS and peers argue that incremental CPU details might excite developers but fail to deliver the jaw-dropping roadmap needed to justify fresh capital inflows.

I respectfully disagree on the long-term impact. An agentic-era AI CPU isn’t just another iteration -- it represents a strategic pivot into a market where demand is exploding. Agentic AI workloads are projected to reshape everything from enterprise automation to consumer apps, requiring CPUs that Nvidia can now supply at scale. If the surprise chip demonstrates superior performance-per-watt or seamless integration with existing GPU clusters (as hinted in the Meta deployment), it could fundamentally expand Nvidia’s addressable market.

This shifts the investment thesis from “GPU monopoly” to “AI platform dominance,” potentially sustaining premium valuations for years. Short-term volatility is possible, but the structural tailwinds feel real.

Nvidia shares have been range-bound for more than six months, hovering near recent levels as investors tire of relentless AI hype cycles and digest sky-high expectations. Even a compelling AI CPU announcement could provide fresh momentum for growth by validating the next phase of AI infrastructure. Yet the risk of a “buy the rumor, sell the news” reaction looms large -- especially since CPU ambitions have been discussed for months and the market has already priced in some optimism. Recent buying interest suggests conviction remains, but disappointment over missing “surprise” elements could trigger a quick pullback.

Ultimately, today’s keynote won’t rewrite Nvidia’s story overnight. But if Huang delivers on the agentic CPU front, it could mark the beginning of a broader growth chapter beyond GPUs.

Most Americans drastically underestimate how much they need to retire and overestimate how prepared they are. But data shows that people with one habit have more than double the savings of those who don’t.

And no, it’s got nothing to do with increasing your income, savings, clipping coupons, or even cutting back on your lifestyle. It’s much more straightforward (and powerful) than any of that. Frankly, it’s shocking more people don’t adopt the habit given how easy it is.