Author: Jeff Park Source: X, @dgt10011 Translation: Shan Ouba, Jinse Finance
Since I entered the crypto industry, I have always held two independent yet closely related goals in mind: first, to reshape the underlying principles of sound currency based on a censorship-resistant value storage medium (Bitcoin); the other equally important mission is to leverage technology to build a more efficient and accurate capital market adapted to the digital age. And I believe that Solana can help advance the second mission. Before we delve into the discussion, let's talk about a revolutionary technology company that is now a household name but was little known in its early days—Nvidia.
In 2013, Nvidia began developing its first supercomputer specifically designed for AI, the DGX-1, initiating a comprehensive revolution from the underlying chip architecture and launching Pascal architecture chips using FinFET technology.
This chip combines the three core ideal characteristics of transistors: high precision, high energy efficiency, and controllability. Although this was the most important hardware upgrade in the transistor field since the 1970s, it was unknown to the public more than a decade ago. Jensen Huang's ability to predict that the industry's core bottleneck had shifted from simple computing power to data transmission between chips was inspired by the NVLink interconnect technology in 2014. Even after the DGX-1 was officially launched in 2016, it took Nvidia another ten years to grow into one of the world's most legendary and top-valued companies. Many people mistakenly believe that Nvidia's success stems from its superior hardware circuitry compared to competitors, but the true core of its success lies in its comprehensive software system. As Moore's Law gradually became obsolete, computer scientists used numerical algorithms to accelerate matrix operations, giving GPUs an unprecedented new instruction set. Stephen Witt, in his book *Thinking Machines*, calculated that hardware transistors only contributed 2.5 times the performance improvement, with the remaining nearly 400 times speedup coming entirely from a mathematical toolkit called CUDA. Sharing NVIDIA's journey serves two purposes: First, the path from scientific discovery to commercialization is long and arduous; second, the simultaneous pursuit of high precision, high efficiency, and controllability at both the software and hardware levels is the core appeal of contemporary breakthrough technologies. NVIDIA and Solana share a profound commonality in this regard, far exceeding what is superficially apparent. In my view, it's no coincidence that Solana founder Toly comes from Qualcomm's GPU R&D system. The project's initial core bet was on a Sealevel parallel execution environment: transactions in a multi-shard state can be processed simultaneously and in parallel on multi-core processors. Essentially, this logic transplants the core ideas of CUDA to the blockchain ledger system. The complex timing coordination logic required to achieve efficient parallelism gave rise to the well-known original consensus mechanism—Proof of History (PoH). Although Solana's early architecture was adapted to GPU-to-TPU hardware, NVIDIA's deeper lessons for the industry are worth learning: its long-term core competitive advantage lies in its continuously iterating mathematical toolkit. The most direct evidence is that its underlying consensus mechanism has been completely rewritten; the Alpenglow upgrade will completely eliminate PoH and enable a new, faster consensus path. Currently, the market is generally bearish on the Solana mainnet, arguing that general-purpose public chains lack real-world application value, and judging its market fit based on block space utilization is flawed. However, a highly relevant story from 2009 illustrates this: Geoff Hinton led graduate students (including Sutskever) in conducting machine learning experiments using NVIDIA graphics cards, but their request for free graphics cards was refused. At that time, AI research was not included in NVIDIA's plans for parallel computing applications. Ultimately, however, AI became the most explosive application scenario for parallel computing—parallel computing itself is a general-purpose underlying technology. This early technology initially gained popularity through the academic community's bulk purchase of GeForce graphics cards and multi-card interconnection, but Jensen Huang had already noticed this trend as early as 2004, when the entire industry was unaware. How did he seize this opportunity? This precisely explains how all cutting-edge research without short-term commercial returns obtains funding: relying on a small, core group of enthusiasts. NVIDIA's first dual-purpose chip combining graphics and parallel computing was highly controversial internally because it increased the production cost of GeForce graphics cards, exceeding that of competing products like Radeon—this is what the industry calls the "CUDA tax." The brilliance of CUDA lies in making gamers bear the huge cost of chip development. Correspondingly, Solana's "CUDA tax" undoubtedly comes from the core, loyal user group within the trading community—a group that integrates gaming, entertainment, and finance. An objective fact: all early cutting-edge technologies need a group of early developers and users who pursue ultimate performance and delve into general-purpose scenarios; this is also a hallmark of a product-market fit. If you are willing to observe, this trend is already clearly visible today. I have ample evidence to suggest that a major breakthrough in cryptography is in its early stages. Solana became the world's first listed and simultaneously launched on-chain trading market. Within 24 hours of its listing, SPCX saw $52 million in transactions across 51 liquidity pools on Solana. Solana is gradually becoming the underlying AI settlement platform, covering the decentralized computing power inference market, DePIN data networks, and autonomous on-chain intelligent agents. Simultaneously, it has joined the ranks of leading stablecoin ecosystems, consistently ranking among the top in monthly on-chain transaction volume. All of this stems from its underlying architecture, which from its inception adhered to a rigorous design principle of pursuing ultimate general-purpose performance. Of course, this is only early data, and no definitive conclusions have been reached. The core questions to be addressed are: can these niche scenarios achieve a mature commercial closed loop, building a sustainable and shared network economic model that is not limited to a single purpose? Can related applications break free from reliance on short-term incentive subsidies and form a self-reinforcing long-term positive cycle? Huang Renxun's friends have long commented that his core ability is not a talent for predicting the future, but a kind of "resonance perception": based on what is seen in the present, combined with continuous and repeated communication with customers and employees, he uses logical deduction to predict the long-term trend of the industry. I believe that, whether from the perspective of development accumulation or fundamental thinking, the Solana community is fully prepared and has the comprehensive capabilities to drive the realization of the next generation of industry opportunities. I am full of anticipation and honor to participate in this journey, and I will do my best to contribute to this promising future of the industry.
JinseFinance