Kyle Samani Is Back: This Time, We're Outperforming CEX in Efficiency!

marsbitPublicado em 2026-03-12Última atualização em 2026-03-12

Resumo

Kyle Samani, former co-founder of Multicoin Capital, has returned to advocate for PropAMM, a breakthrough in market microstructure innovation on Solana. He argues that PropAMM could surpass traditional centralized exchanges (CEXs) in efficiency by hosting market maker algorithms directly on the blockchain, eliminating latency from bidirectional communication between exchanges and market makers. In CEXs, market makers co-locate servers and constantly exchange data with the exchange, leading to delays. In contrast, PropAMM runs algorithms on the same physical hardware as the Solana blockchain, allowing price updates to occur almost instantaneously within the same silicon chip. This reduces latency significantly compared to traditional systems. Currently, PropAMM already offers narrower spreads for SOL-USDC spot trading than major CEXs. Samani expects it to dominate on-chain trading this year, including spots, perpetuals, and prediction markets. Challenges remain in ensuring best execution for takers due to private algorithms and non-deterministic routing, but solutions from aggregators like Jupiter and dFlow are anticipated. Upcoming Solana upgrades—such as higher compute limits, reduced slot times, and lower network latency—will further enhance PropAMM performance, solidifying its advantage over CEXs.

Source:Kyle Samani

Compiled | Odaily Planet Daily (@OdailyChina); Translator | Azuma (@azuma_eth)

Editor's Note: The man who knows best how to promote Solana, former co-founder of Multicoin Capital Kyle Samani, who had loudly exited the crypto world not long ago, is back!

Last night, Kyle Samani posted a long Thread on his personal X account. In the post, Kyle Samani once again demonstrated his highly persuasive "promotion" (not derogatory here) rhetoric, using "efficiency"—a weak point in the decentralized narrative—as a breakthrough. He detailed how the PropAMM currently being promoted in the Solana ecosystem will catch up with or even surpass traditional centralized models in efficiency, arguing that PropAMM is one of the most important innovations in market microstructure in recent years, or even decades.

  • Related articles: 《The Man Who Knows Best How to Promote SOL Exits the Crypto World》; 《Is There More to Kyle Samani's Exit?》.

Below is the original content by Kyle Samani, compiled by Odaily Planet Daily.

PropAMM is one of the most important innovations in market microstructure in recent years, and perhaps even one of the most significant in decades.

To help everyone understand this conclusion, let's first look at how market makers (MM) quote prices on traditional centralized exchanges (CEX).

Market makers typically physically co-locate with the exchange. Each market maker runs an algorithm on a server and connects to another server running the exchange's system via a network cable of uniform length (e.g., 50 meters).

A massive stream of data is constantly sent back and forth between market makers and the exchange. Every time a market maker sends an order to the exchange—whether it's a limit order, cancellation, or market order—the exchange must broadcast this information to all other market makers; then, other market makers resend their own orders based on the new information; this cycle repeats indefinitely.

Here is a simple diagram.

Now let's look at how propAMM works on the Solana mainnet.

The beauty of propAMM on Solana is that the blockchain itself directly "hosts" the market maker algorithms. This means the system no longer needs to send billions of messages back and forth between market makers and the exchange; the market-making algorithms will run directly on the same physical machine as the exchange.

The new diagram is as follows. (Yes, only the Solana blockchain is needed!)

There has long been a common view in the cryptocurrency industry that decentralized systems must be slower (have higher latency) than centralized systems because they require communication between global nodes.

But if you think about it differently, on-chain hosted algorithms may actually have lower latency than traditional centralized exchanges in finance.

Why is that? The reason is that the latency required for propAMM to update prices only involves electrons moving within the same physical piece of silicon. For example, if the last market order causes a change in the SOL-USD price, this information is immediately visible to all propAMMs and used to price the next market order. Everything happens within the same piece of silicon; no bidirectional communication between servers is needed.

It's worth noting that propAMM does require frequent oracle updates, but this is not a problem and does not change the overall fact I described above.

The most critical point remains: when the exchange—in this case, the Solana blockchain—directly hosts the propAMM algorithms, the market makers' pricing changes in real-time within the same physical piece of silicon.

propAMM has already become the dominant mechanism for SOL-USDC spot quotes on Solana, with narrower spreads than all major CEXs. I expect this market structure to become the dominant model for on-chain trading this year, including spot, perpetual contracts (perps), and even prediction markets.

The biggest challenge for propAMM is that there is currently no way to ensure that the taker always gets the best execution, because:

  • All propAMM algorithms are not public (which is reasonable, as traditional market-making algorithms are also proprietary);
  • When routing trades between multiple propAMMs, the result is non-deterministic.

However, this problem can be solved. I expect all relevant aggregator teams to launch solutions this year, such as Jupiter and dFlow for spot, and Phoenix for contracts.

Current propAMM is still not fully optimized and is subject to various limitations of the Solana blockchain itself. This year, Solana will roll out a series of major upgrades that will significantly improve propAMM's performance, including:

  1. Higher CU (compute unit) limits per transaction and larger transaction sizes;
  2. Higher CU limits per block;
  3. Alpenglow: reducing slot time from 400ms to 100–150ms;
  4. DoubleZero: reducing global network latency;
  5. Application-controlled execution;
  6. Multiple concurrent leaders.

If propAMM on the Solana mainnet can already offer narrower quotes than all CEXs without these upgrades, imagine how powerful they will become as these upgrades are gradually implemented.

Perguntas relacionadas

QWhat is PropAMM and why does Kyle Samani consider it a major innovation in market microstructure?

APropAMM is a mechanism on Solana where market-making algorithms are hosted directly on the blockchain, eliminating the need for constant bidirectional communication between market makers and exchanges. Kyle Samani considers it one of the most important innovations in market microstructure in recent years, or even decades, because it significantly improves efficiency by running algorithms on the same physical hardware as the exchange, reducing latency and narrowing spreads compared to traditional CEXs.

QHow does PropAMM on Solana achieve lower latency compared to traditional centralized exchanges (CEXs)?

APropAMM achieves lower latency by hosting market-making algorithms directly on the Solana blockchain, which runs on the same physical hardware. This means price updates and order processing occur within the same silicon chip, eliminating the delay caused by network communication between separate servers in traditional CEX setups. Electronic movements within the chip are faster than inter-server data transfers.

QWhat current evidence does Kyle Samani provide to show PropAMM's effectiveness on Solana?

AKyle Samani points out that PropAMM has become the dominant mechanism for SOL-USDC spot quotes on Solana, with narrower spreads than all major centralized exchanges (CEXs). This demonstrates its superior performance in market efficiency and execution quality compared to traditional systems.

QWhat are the main challenges facing PropAMM, and how does Samani expect them to be addressed?

AThe main challenges are ensuring best execution for takers, as PropAMM algorithms are not public and routing between multiple PropAMMs is non-deterministic. Samani expects solutions to be developed by aggregator teams like Jupiter and dFlow for spot trading, and Phoenix for perpetual contracts, within the year to optimize routing and execution.

QWhat upcoming Solana upgrades does Samani mention that will further enhance PropAMM performance?

ASamani mentions several Solana upgrades: higher CU (Compute Unit) limits per transaction and larger transaction sizes, higher CU limits per block, Alpenglow (reducing slot time from 400ms to 100-150ms), DoubleZero (reducing global network latency), application-controlled execution, and multiple concurrent leaders. These will collectively boost PropAMM efficiency and performance.

Leituras Relacionadas

From South Korea to the United States: Blue-Collar Jobs Are Becoming Increasingly Popular, Thanks to AI

AI is reshaping the labor market's value proposition. The traditional four-year college degree is losing its appeal as a guaranteed career path, while skilled blue-collar trades like electricians, welders, and plumbers are experiencing historic demand and wage premiums. This shift is driven by dual pressures: AI's displacement of certain white-collar roles and a booming need for physical infrastructure and data center construction. Data confirms the trend. In the U.S., vocational school revenue surged, and a significant portion of recent layoffs are AI-related. Surveys show a majority of Gen Z adults plan to pursue blue-collar work, citing better job security against AI automation. Vocational education interest has exploded recently. Experts cite a psychological shift as younger generations seek tangible, AI-resistant careers and avoid high student debt. In many cases, salaries for skilled trades now match or exceed those requiring a bachelor's degree. In South Korea, semiconductor vocational high schools boast near-total employment, with graduates securing high-paying roles at companies like Samsung. The shortage is structural, exacerbated by a retiring baby boomer workforce and massive infrastructure projects. Companies like JPMorgan Chase, Meta, and Lowe's are investing heavily in training programs. However, overcoming historical stigma and a "perception gap" around trade careers remains a key challenge to closing the talent gap.

marsbitHá 44m

From South Korea to the United States: Blue-Collar Jobs Are Becoming Increasingly Popular, Thanks to AI

marsbitHá 44m

Qualcomm: AI Hype Subsides, When Will Smartphones Emerge from the Gloom?

Qualcomm reported its Q3 FY2026 results (ending June 2026), with revenue of $9.95B, down 4% YoY but above expectations. Gross margin declined to 53.1%, pressured by rising costs across manufacturing and memory. Key business segments showed mixed performance: Handset revenue fell 19.6% YoY to $5.09B, dragged by an 11% decline in non-Apple Android shipments and weaker high-end mix. Conversely, Automotive revenue surged 61% to $1.59B, and IoT grew 9% to $1.83B. Core operating profit dropped 41% YoY due to margin compression and higher expenses. Management's Q4 FY2026 guidance projects revenue of $9.7B-$10.5B, in line with consensus, but Non-GAAP EPS guidance of $2.05-$2.25 fell short of expectations. Amidst persistent weakness in its core handset market, Qualcomm is pursuing growth in AI, focusing on Edge AI (smartphones, PCs, automotive) and Data Center AI. Its data center strategy includes four pillars: AI accelerators (e.g., AI200), commercial CPUs (Dragonfly C1000), custom silicon, and connectivity solutions. While these initiatives initially boosted its stock, concerns over AI capital expenditure sustainability have since erased those gains. The company targets $5B in data center revenue for FY2027 and $15B for FY2029. The report concludes that with the traditional handset business still under pressure, the data center opportunity is currently viewed as a longer-term option, and a more conservative valuation based on core operations may be warranted until AI contributions materialize.

marsbitHá 48m

Qualcomm: AI Hype Subsides, When Will Smartphones Emerge from the Gloom?

marsbitHá 48m

From TPU to Self-Evolving Agents: How Jeff Dean Predicts the Next Step in AI

At the 2026 YC Startup School, Jeff Dean outlined his vision for AI's next phase, shifting focus from simply scaling models to building intelligent, autonomous systems. He believes AI's progress is no longer just about creating smarter models, but about integrating them into systems capable of long-term, iterative work, automated experimentation, and continuous learning. This evolution moves the competition from "who has the bigger model" to "who can best organize intelligence." Dean suggests AI capabilities are now comparable to a junior engineer, enabling the automation of complex workflows. However, the true challenge and opportunity lie in managing these AI "workers" at scale. He emphasizes the importance of **context engineering**—structuring tools, memory, and feedback loops—over raw model power. For startups, this means building deep expertise in niche domains where general models currently fail (near 0-1% success rates), leveraging proprietary data, specialized tools, and domain-specific evaluators. A recurring theme is re-examining fundamental constraints. Dean's past work, like moving Google's search index to memory or creating the TPU, stemmed from questioning outdated assumptions about hardware and cost. He sees similar inflection points today, particularly in **specialized inference hardware** to drastically reduce latency and energy consumption for real-time Agent operation. Notably, he points out that in modern AI systems, the dominant cost is often not computation but **data movement**. Reliable, long-running Agents require robust system design, borrowing concepts from distributed computing like checkpointing, state management, and parallel exploration to handle failures and maintain progress over days or weeks. As AI automates execution, the scarcest human skills will shift to **defining clear specifications**, **judging what problems are worth solving** (taste), and designing effective feedback loops. Ultimately, Dean's framework prioritizes understanding the problem deeply, identifying the true bottlenecks, and systematically building closed-loop systems where AI can not only perform tasks but also improve AI itself.

marsbitHá 48m

From TPU to Self-Evolving Agents: How Jeff Dean Predicts the Next Step in AI

marsbitHá 48m

Trading

Spot
活动图片