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

marsbitPublished on 2026-03-12Last updated on 2026-03-12

Abstract

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.

Related Questions

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.

Related Reads

You Use Claude and Codex Every Day, but Meta Has Restricted Internal Use

In May, Meta imposed internal restrictions on its engineers regarding the use of Claude Code and Codex, two widely used AI programming tools. Despite being a major client, Meta's guidelines, still in effect, prohibit these external models from being used for specific tasks to prevent potential "escalations with partners." The core concern is "distillation"—the risk that outputs from Claude or Codex could inadvertently contaminate the training data and evaluation processes for Meta's in-house AI coding assistant, MetaCode. If MetaCode is trained or evaluated using data generated by these external models, it risks learning their capabilities rather than developing its own, blurring the line of intellectual origin. The restrictions are precise: engineers cannot use the external models to generate test questions, debug source code, or suggest test cases. AI-generated content is also barred from environments accessible to MetaCode. However, AI can still assist with peripheral tasks like workflow setup and code organization, provided all outputs are manually reviewed. This caution reflects a broader industry dilemma. While distillation is a common technique, using a competitor's model output for training raises legal and ethical questions about the ownership of derived capabilities. Contractual terms from companies like OpenAI and Anthropic explicitly forbid using their outputs to build competing products, putting enforcement power in the hands of rivals. The move is also financially motivated, as Meta seeks to reduce its hefty internal AI spending, estimated in the billions this year. Meta's policy illustrates the delicate balance companies must strike: leveraging powerful external AI tools while safeguarding the integrity and independence of their own AI development. As AI systems increasingly help build other AIs, distinguishing the origin of capabilities becomes a fundamental challenge for the entire industry.

marsbit41m ago

You Use Claude and Codex Every Day, but Meta Has Restricted Internal Use

marsbit41m ago

Why Do We Need an AI Content Perspective Today?

The article "Why Do We Need an AI Content Perspective Today?" explores the complex and often contentious integration of AI into the cultural and creative industries, particularly film and television. It begins with the cancellation of Amazon's AI-generated animation "Punky Duck," highlighting the ethical debates surrounding AI content. AI's rapid advancement is transforming video production, enabling cost-effective, full-length AI films (e.g., "RAPHAEL," "Dreams of Violets") while sparking industry resistance over issues like "synthetic actors." The core debate has shifted from whether to use AI to how to use it responsibly. The article analyzes why AI's entry into film is uniquely unsettling. It distinguishes between "cultural fast food" (short-form, fast-paced content like micro-dramas) and "cultural main courses" (traditional, long-form film/TV). AI currently excels at the former, matching its fragmented narratives, shallow emotional needs, and free-to-consumer models. However, venturing into the latter challenges the human-centric essence of storytelling—creativity, emotional depth, and the unique value of human labor and experience. While AI can generate massive volumes of content and lower costs, it risks devaluing human creativity, leading to homogenized output, and creating unfair competition through potential intellectual property infringement. Its efficiency also amplifies content safety risks, making preemptive governance crucial. To counter these risks, the article proposes establishing clear boundaries guided by a human-centered AI content perspective. It outlines four principles: 1) Amplify, rather than displace, human creative space; 2) Respect and protect human creative output; 3) Ensure human creative control and responsibility remain paramount; and 4) Guarantee transparency and traceability in AI creation. The conclusion emphasizes that humans must act as the "helmsmen" of technology, steering AI development to enhance, not replace, the core human values at the heart of cultural expression.

marsbit1h ago

Why Do We Need an AI Content Perspective Today?

marsbit1h ago

Planck Retracted? The Father of Quantum Tripped by an Algorithm

The recent discovery that two articles (published in 1940 and 1942) by Max Planck, the Nobel laureate and founder of quantum theory, are marked as "retracted" on Springer's digital platform highlights a curious clash between historical publishing practices and modern automated systems. An investigation suggests these retractions are algorithmic errors, not due to fraud or misconduct. The papers, philosophical reflections on science published in *Die Naturwissenschaften*, were likely flagged by the platform's systems. One article, a republished lecture, may have been mistaken for duplicate publication. Another, sharing a title with a prior article by a different author (a common practice for continuing debates at the time), may have triggered a similar automated check. The digital versions have even been replaced with blank pages, contrary to normal practice of preserving retracted texts. This incident underscores how contemporary digital infrastructure, built around concepts like "self-plagiarism" and strict copyright, can misclassify and obscure legitimate historical scholarly communication. It serves as a warning that digital archives are not neutral mirrors of the past but are filtered by platform rules, potentially distorting the scientific record. As AI systems increasingly rely on such databases, such erroneous metadata could propagate, affecting how future tools interpret and access historical knowledge.

marsbit1h ago

Planck Retracted? The Father of Quantum Tripped by an Algorithm

marsbit1h ago

Refunds! Claude 4.8 Sees Overnight Major 'Dumb-Down', GPT-5.6's Computational Power Reportedly 'Halved'

The AI community is currently alarmed by widespread reports of significant performance degradation in two leading models. This article details a "mass self-testing frenzy" triggered by a mysterious prompt designed to detect a hidden "Juice" value, representing a model's reasoning compute budget. On OpenAI's side, users suspect a covert, limited test of a "GPT-5.6-sol" model is underway. When using a specific XML prompt on the Codex platform, a normal "gpt-5.5 xhigh" model reportedly returns a Juice value of 768. However, some users routed to the suspected GPT-5.6 test receive a drastically reduced value of 128—a six-fold decrease. This has sparked debate on whether it signifies a major efficiency leap or a "watered-down, low-cost version" achieved by slashing reasoning depth to save computational expenses. Simultaneously, Anthropic's Claude models, particularly the flagship Opus 4.8 Max, are facing intense user backlash for a perceived "physical brain cut." Users on platforms like Reddit report a dramatic decline in the model's once-impressive reasoning, with complaints of it becoming "absurdly" weakened, performing worse than older, lighter models like Haiku. Specific criticisms include: losing long-context memory, refusing to think deeply even in high-reasoning modes, providing instant incorrect answers, and engaging in unhelpful, argumentative, or "gaslighting" behavior where it contradicts users unnecessarily. The article speculates these "stealth downgrades" might be a calculated corporate strategy. Companies could initially release models with temporarily boosted compute to create an illusion of a major breakthrough, then silently scale back parameters later to manage unsustainable inference costs. A proposed underlying cause is a tightened funding environment, potentially exacerbated by SpaceX's massive IPO soaking up market liquidity, which could delay AI company IPOs and force cost-cutting measures like model "nerfing." The core issue highlighted is the asymmetry of information: subscribers pay for a service that can be silently and fundamentally altered without notification or explanation. The viral "Juice test" resonates because it represents users' desire for transparency about what they are actually paying for.

marsbit1h ago

Refunds! Claude 4.8 Sees Overnight Major 'Dumb-Down', GPT-5.6's Computational Power Reportedly 'Halved'

marsbit1h ago

Trading

Spot
活动图片