AI Solves 80-Year Math Problem as Nvidia Rewrites PC Chips

Here is the part that should stop you mid-scroll: a Fields Medal winner — the mathematician's equivalent of a Nobel Prize — called an AI's solution to an 80-year-old problem a genuine milestone. Not a party trick. Not a benchmark flex. A milestone.

The problem in question is the Erdős unit distance conjecture, a deceptively simple geometry puzzle first posed in 1946. Picture a bunch of dots scattered on a flat plane. Now count how many pairs of dots sit exactly one unit of distance apart. The question Paul Erdős asked was: how many such pairs can you pack into a set of points before the geometry simply won't allow any more? Mathematicians had been wrestling with the upper bound of that answer for eight decades. OpenAI's model didn't just make progress on it — it disproved the conjecture entirely.

Fields Medalist Tim Gowers publicly called it a milestone in AI mathematics. University of Toronto professor Daniel Litt went a step further, saying it was the first AI result he found exciting on its own terms, not just as a sign of things to come. That is a meaningful distinction. Mathematicians are a skeptical crowd.

So how did the model actually do it? Not through some alien form of reasoning we've never seen before. The AI drew on existing techniques spread across several subfields of mathematics and stitched them together into a complete proof. It was more like a very well-read, extremely tireless research assistant than a genius having a eureka moment in the shower. The proof has since been cleaned up and extended by human mathematicians, which tells you something important about where we actually are.

This is the latest step in a progression that has moved faster than almost anyone predicted. Three years ago, large language models were fumbling basic arithmetic. Last year, they started acing high school math competitions. Earlier this year, they were contributing to research in constrained settings, with heavy human interpretation required to turn their outputs into anything publishable. Now one has independently resolved a major open conjecture.

The most honest framing right now is that AI and human mathematicians are better together than apart. AI systems have effectively read everything — every paper, every proof, every clever trick from every subfield — and they will grind through hundreds of dead-end proof strategies without complaint. Humans, meanwhile, can still think more deeply about a single hard problem and ask the kinds of questions that point research in genuinely new directions.

But that division of labor may not hold. AI capabilities in mathematics have been compounding so quickly that serious people are now openly asking what role human mathematicians will play ten years from now. That is not a dismissive question — it is an honest one.

Paul Erdős, who posed the original problem, wrote more than 1,500 mathematical papers in his lifetime, more than any other mathematician in history. His gift was finding problems that fit in a sentence but contained worlds of hidden complexity. It is a little poetic that a problem he dreamed up became the one that marked a new era in machine intelligence.

Nvidia just decided that dominating the data center and the gaming GPU market was not enough, and this fall it is coming for Intel and AMD's living room too.

The company is launching the RTX Spark, its first chip designed to power consumer laptops and mini-PCs from the ground up. This is not a graphics card slotted into someone else's machine. It is a full system-on-a-chip — CPU cores, GPU cores, and unified memory all on one piece of silicon — the same architecture play that Apple pulled off with the M1 and completely reshaped the laptop industry. Nvidia is betting it can do the same thing, only with a sharper focus on AI workloads.

The flagship RTX Spark packs 20 CPU cores, 6,144 GPU cores, and up to 128GB of LPDDR5X unified memory. Those specs are not timid. Nvidia claims the chip can handle editing 12K video, rendering a 90GB 3D scene, and running graphically demanding games at 1440p and 100 frames per second — all on a laptop thinner than 14 millimeters, unplugged. Whether those claims hold up in independent testing is a different question, and notably, Nvidia did not share a single benchmark chart at announcement.

The memory ceiling is where things get genuinely interesting for AI. With 128GB of unified memory, an RTX Spark machine can run 120-billion-parameter AI models locally — on your laptop, not in a cloud data center. For context, that is the class of model that most people currently access through a web browser and a monthly subscription. Running one locally means no latency, no data leaving your machine, and no usage limits.

Microsoft is paying close attention. At its Build conference, the company is showing off new security and containment features designed specifically for personal AI agents running on RTX Spark hardware. The vision Nvidia is selling is a PC where you talk to your computer instead of clicking through menus — where an AI agent handles repetitive tasks, automates workflows, and controls apps on your behalf.

There is a real catch worth flagging. Like Apple's M-series and Qualcomm's Snapdragon X chips, the RTX Spark is built on Arm architecture, not the x86 architecture that Intel and AMD have used for decades. That means older Windows software needs to run through an emulation layer, which can hurt performance. Microsoft has spent years improving its Prism emulator for exactly this scenario, and the situation is meaningfully better than it was when Arm-based Windows laptops first arrived, but it is still a real compatibility question that buyers will need to think through.

Nvidia is also planning lower-cost versions of the chip with as little as 16GB of RAM, which suggests it wants to seed the market broadly rather than only chase premium buyers.

For Intel and AMD, this is the most credible new threat to their PC chip dominance in years. Apple already proved that a vertically integrated chip strategy can flip the laptop market. Now Nvidia — a company with arguably the strongest AI hardware brand on the planet — is running the same play on Windows.