HelixNetwork
Whitepaper · Draft V0.1

The DNA of AI

// Aggregate compute, empower creation — where supply and demand complete each other.

TOKEN  $HLX CHAIN  Base INCUBATOR  Helix Labs DOMAIN  helixnetwork.app
DRAFT · FULL · CH. 1–11
Contents

Eleven chapters, one helix.

Starting from the industry opportunity, this whitepaper unfolds Helix's network structure, three pillars, growth logic and technical foundation — closing with roadmap, governance and risk disclosure.

  1. 01Abstract & Vision
  2. 02Market & Opportunity
  3. 03Protocol Overview
  4. 04Pillar I — Helix Compute
  5. 05Pillar II — Helix Labs
  6. 06Pillar III — HLX Token Utility
  7. 07The Growth Flywheel
  8. 08Architecture & Chain Choice
  9. 09Roadmap
  10. 10Governance & Decentralization
  11. 11Risk Disclosure & Legal
01

Abstract & Vision

Helix Network is an ecosystem network connecting AI compute with software creation. It aggregates large-model and GPU compute scattered across the world above, incubates the developers and products of the Vibe Coding era below, and lets compute supply and application creation drive each other within a single network — forming a self-reinforcing growth loop.

Over the past three years, AI's capabilities have expanded at a staggering pace — but the dividends of that capability have not been distributed equally. On one side is compute: highly concentrated in the hands of a few cloud providers who define its price and availability, with wait times for advanced GPUs often measured in quarters. On the other side are creators: models grow ever more capable, yet those who could actually create value with them remain shut out by compute cost, distribution channels and reward mechanisms. Capability is inflating, but the barrier hasn't truly disappeared — it has merely shifted from "can you write code" to "can you afford the compute, can your product be seen, can the creator be rewarded."

It is precisely this gap that Helix Network sets out to bridge.

Three things happening in the AI industry

Compute is becoming the resource that determines who gets to participate in AI. Training and running AI demands enormous compute, and access to it is highly concentrated — a handful of vendors hold the pricing power and the advanced capacity. Even as total supply eases, price gaps, fragmentation and access barriers still keep many small and mid-sized developers from using it efficiently. The world is not short of idle GPUs; what's missing is a coordination layer that aggregates them into open, reachable supply.

The barrier to creating software is collapsing. Vibe Coding shifts the bottleneck of creation from "can you write code" to "can you describe an idea clearly." This means the number of creators will grow exponentially — but they need more than tools: they need cheap compute, channels to be seen, and rewards that match their contribution. For the first time, the supply of compute and the demand of creation can be placed in one system to complete each other.

The ownership of value is being redefined. In the centralized model, AI's value accrues to the balance sheets of a few companies; users contribute data, demand and compute consumption, yet own no part of the network. An open network, collectively owned by its participants, can return value to those who genuinely contribute compute and create applications. This is a structural difference the centralized model cannot replicate.

These three curves are unfolding at once. Helix Network stands at their intersection.

Two strands, one helix

The structure of this network can be understood through its name. The DNA double helix is made of two complementary strands: wound around each other, mutually supporting, carrying information as they grow upward. Helix Network also has two strands — Compute and Creation.

STRAND A · COMPUTE

The Compute Strand

Aggregates global large-model and GPU compute — starting as an aggregator, evolving toward a decentralized compute market and self-built compute centers. It powers creation.

STRAND B · CREATION

The Creation Strand

Through Helix Labs, incubates the developers and products of the Vibe Coding era, turning compute into real applications and real demand. It gives compute its meaning.

These two strands are not parallel business lines but a loop of cause and effect: the cheaper compute becomes, the more creation flourishes; the more creation flourishes, the higher the demand for and value of compute. Every product incubated consumes compute, brings new users, and draws more creators into the network; and ever-growing demand, in turn, makes aggregating more compute and driving down cost all the more worthwhile. Two strands, wound together, growing upward — this is Helix's growth loop.

// CORE BELIEF Those who use this network should also be those who own it. Value should not flow only to the center, but back to every participant who contributes compute and creates applications.

Within this loop, HLX is the coordination mechanism that keeps the cycle turning — it carries settlement, incentives and governance, aligning compute suppliers, application creators and network users under one shared set of incentives. Its specific role is detailed in Chapter 6; this version of the whitepaper focuses on the logic of the ecosystem itself, not the token's economic parameters.

How Helix differs from a "pure compute network"

Decentralized compute is not a brand-new field. A number of projects already aggregate GPUs and rent out compute. But most of them solve only the supply side — bringing compute in, selling it out — while demand still depends on the natural arrival of an external market. When external demand softens, such networks often survive only on subsidies.

What makes Helix different is that it internalizes the demand side. Helix Labs is not an auxiliary incubator but the network's "demand engine": every Vibe Coding product it incubates is a consumer of compute, an entry point to the ecosystem, and a source of network expansion. Compute supply and creation demand feed each other within one network, rather than each fending for itself in an external market. This "self-generated demand" spiral is Helix's most fundamental difference — and its moat.

This network is built on Base (Coinbase's Ethereum L2) — a chain whose goal is to "bring billions of users onchain," with a mature consumer-app ecosystem and a low-friction onboarding experience. It offers natural distribution soil for Helix's developer ecosystem. The specific architectural choices are covered in Chapter 8.

// IN ONE LINE Aggregate compute above, incubate creation below, and let supply and demand drive each other and grow within one network — this is Helix Network.

The chapters that follow unfold, in order, the market context Helix operates in (Ch. 2), the network's overall structure (Ch. 3), how the three pillars work (Ch. 4–6), the flywheel that drives growth (Ch. 7), the technical architecture and chain choice (Ch. 8), and the roadmap, governance and risk disclosure (Ch. 9–11).

02

Market & Opportunity

The previous chapter described three trends unfolding in the AI industry. This chapter focuses on the two most directly relevant — the supply of compute and the demand of creation — and uses data to see where they stand: compute supply is moving from "absolute shortage" to "structural imbalance," while creation demand is exploding because of Vibe Coding. A network connecting the two stands right at their intersection.

Compute: shortage easing, imbalance deepening

The acute "can't-get-a-GPU" shortage of 2023 had eased by 2026. As the Blackwell architecture shipped at scale and competing capacity from AMD and Google TPU came online, AI compute supply entered a kind of "functional balance" in 2026 — H100 cloud rental prices fell from roughly $8 per hour in early 2023 to a range of $1.8–3.5 in 2026, with spot pricing as low as $1.2.

But "no longer unobtainable" does not mean "fair and efficient." The real problem has shifted from total shortage to three structural imbalances:

First, absurd price gaps. The same H100 can cost three to five times as much per GPU-hour depending on the provider. AWS, for instance, offers only 8-GPU instances; normalized per GPU, that exceeds $7.5 per hour — roughly 3× the market median and 5× the cheapest provider. Pick the wrong provider, and a developer's bill doubles.

Second, pricing power remains concentrated. Top cloud providers lock up most of the chipmakers' advanced capacity with multi-billion-dollar forward orders, squeezing mid-market and enterprise customers out of standard channels. Bottlenecks deep in the supply chain — advanced packaging capacity booked through 2027, rising high-bandwidth memory costs — mean this concentration won't disappear soon. The rental price of the latest-generation GPUs has even moved against the usual trend, rising rather than falling.

Third, fragmentation makes compute hard to use efficiently. Idle GPUs are scattered across data centers, mining farms and small server rooms; different large-model APIs each have their own interface, billing and rate limits. A developer seeking the best cost for a task often has to compare, switch and reconcile across a dozen providers and models. The world is not short of compute; it lacks the coordination layer to aggregate it into unified, reachable supply.

// THE KEY SHIFT As shortage eases and competition intensifies, the industry's battleground is shifting from "who has compute" to "whose compute is easier to use" — developer experience, reliability and smart routing are replacing the pure price war. This is the opening for an aggregation layer.

Creation: barrier collapsing, creators growing exponentially

If compute is the supply-side story, then Vibe Coding is the structural shift happening on the demand side.

Building software once required mastering programming languages, frameworks and engineering practice — which shut most people with ideas out. When AI can turn natural-language description directly into a working product, the bottleneck of creation shifts from "can you write code" to "can you express an idea clearly." This is not a tooling upgrade but an order-of-magnitude leap in the creator base — from tens of millions of programmers to billions of "people with ideas."

But the explosion in creators immediately runs into three walls: compute cost (every generation, every AI app's operation consumes compute), distribution (once built, how does a product get seen), and reward (how does created value flow back to the creator). Today's market leaves each creator to face these three alone.

The landscape: supply side crowded, demand side still empty

Decentralizing compute and aggregating GPUs is not a new idea. The field already has a range of players with different emphases: some focus on large-scale GPU clusters for machine-learning training, some offer cloud-like general compute markets, some extend from graphics rendering into AI inference, and some attempt to decentralize the production of "intelligence" itself. Together they prove one thing — decentralized compute is a real market with paying demand, not merely a concept.

But these projects share a structural trait: almost all of them focus on the supply side. Their work is to bring compute in, aggregate it, and sell it out; demand depends on the natural arrival of an external market. This creates a fragility — when external demand softens, the network often relies on token subsidies to retain suppliers, and once subsidies taper, supply leaves with them.

This is the gap Helix sees: almost no project treats "creating demand" as part of the network itself. Compute aggregation is a red ocean, but the loop of "compute supply × creation demand" remains a blue ocean few have cultivated.

Helix's entry point

Helix does not intend to add one more player to the "whose GPU is cheaper" price war. Its entry point is to connect the two curves above — using an aggregation layer to solve compute's fragmentation and experience problems (picking up "the key shift"), while using an incubator to internalize the exploding creation demand as the network's own growth engine (filling the demand-side gap).

Mature supply-side players have compute but lack their own demand; pure application incubators can create demand but don't control compute. Helix is designed to let both happen at once within one network, driving each other — the specific structure of this loop unfolds from the next chapter on.

Note: The compute prices and supply-demand data cited in this chapter come from public industry sources in Q1–Q2 2026 and reflect market conditions at the time of writing. The compute market is highly volatile and specific figures will change over time; this chapter aims to present structural trends, not real-time quotes.

03

Protocol Overview

Helix Network is built on three pillars: Helix Compute aggregates compute, Helix Labs incubates creation, and HLX coordinates and incentivizes. They are not three separate products but three parts of one loop — supply, demand, and the layer that aligns them.

Chapter 1 described the vision; Chapter 2 made the case for the opportunity. This chapter gives the network's overall structure as a map for the chapters that follow. We first look at what each pillar is, then at how they interlock into a self-reinforcing system.

The three pillars

PILLAR 01 · SUPPLY

Helix Compute

Compute Strand · Supply

The unified layer aggregating global large-model and GPU compute. Starting as an LLM aggregator with smart routing and unified billing; evolving toward a decentralized compute market and self-built compute centers. It is the network's energy source.

PILLAR 02 · DEMAND

Helix Labs

Creation Strand · Demand

An incubator and accelerator for the Vibe Coding era. It turns ideas into products and brings products to users. Every app it incubates is a consumer of compute, forming the network's own sustainable demand.

PILLAR 03 · COORDINATION

HLX

Coordination · Alignment

The network's settlement, incentive and governance mechanism. It aligns compute suppliers, application creators and network users under one shared set of incentives. Detailed in Chapter 6.

The division of labor mirrors the three elements any healthy market needs: someone provides supply (Compute), someone creates demand (Labs), and a mechanism aligns the interests of both (HLX). The difference: in a traditional market these three often belong to separate parties at odds with one another; in Helix, they are designed as one organism that grows together.

How the loop turns

Put the three pillars into one diagram, and the way the network runs becomes clear:

// THE HELIX LOOP
01
Compute aggregates Scattered compute becomes unified, low-cost, usable supply
02
Labs incubates Cheap, reliable compute lowers the barrier, incubating more products
03
Products consume Each product consumes compute and brings users — real demand
04
Demand feeds supply Growing demand makes aggregating more compute worthwhile
The loop accelerates Stronger supply → richer creation → larger demand → back to start

The loop's key is that it needs no external demand to start or sustain it. The products Helix Labs incubates are themselves the demand; and the compute they consume makes Helix Compute's scale and cost advantage all the more meaningful. Supply and demand are mutual causes — two strands wound together, growing upward. This is the literal meaning of "helix."

Who's in the network, and how they benefit

Whether a network can run depends on whether every kind of participant has a clear reason to enter. Helix is designed so that four roles each find their place:

AI developersCall aggregated compute through one interface at better cost and experience, without comparing and switching across a dozen providers.
Vibe Coding creatorsUse Helix Labs' tools, compute subsidies and acceleration to turn ideas into products, gaining distribution and rewards within the ecosystem.
Compute suppliersConnect idle GPUs to the network, serving the real, sustained compute demand within it, and earn rewards.
Participants & holdersTake part in coordination and governance, sharing in the value the network's growth creates — those who use the network also own it.

The path of evolution

This loop won't be built overnight. The compute side and creation side each mature at their own pace — compute starts from aggregation and grows toward self-built capacity; creation starts from incubating a first batch of products and settles into a repeatable capability. How the two strands advance in stages, and what each delivers, forms the network's overall roadmap, unfolded in full in Chapter 9.

The next three chapters go deep into each pillar in turn: how compute is aggregated (Ch. 4), how creation is incubated (Ch. 5), and how HLX coordinates the whole network (Ch. 6).

04

Pillar I — Helix Compute

Helix Compute is the network's energy source. It addresses the compute predicament described in Chapter 2: fragmented supply, wild price gaps, fractured experience. Its answer comes in three steps — aggregate first, then open up, finally build — growing from a "mover" of compute into a "supplier" of compute.

These three phases are not substitutes but layers: each builds on the scale and data of the one before. We take them in turn.

PHASE 1 · NOW

Aggregate: one entry to compute

Starting as an LLM and compute aggregator. Through one unified interface and account, developers reach mainstream closed- and open-source models and compute providers — no more integrating, price-comparing and reconciling with each one separately.

PHASE 2 · NEXT

Open: a decentralized compute market

Opening to the supply side, letting third-party GPU holders connect idle compute to serve real demand within the network and earn rewards. The aggregation layer thus expands from "connecting existing providers" to "an owned supply network."

PHASE 3 · FUTURE

Build: owned compute centers

Once demand and cash flow are proven, building GPU clusters and compute centers as the network's owned capacity — backstopping high-priority, low-latency and privacy-sensitive workloads. From moving compute to owning it.

Phase one: what the aggregator solves

For today's AI developer, "using compute" is itself a hassle. Different models have different interfaces, billing and rate limits; different compute providers vary several-fold in price and unevenly in reliability. To finish a task at the best cost and performance, one often has to weigh, switch and maintain multiple bills across platforms by hand.

Helix Compute's first phase converges all of this into one entry point: one interface, one account, backed by the entire aggregated compute market. The developer describes the need; the network finds the optimal execution path behind the scenes. The value of this lands exactly on Chapter 2's judgment — when compute is no longer scarce and competition turns to "who's easier to use," a unified entry and smart scheduling are themselves the core advantage.

Smart routing: the network's brain

The value of aggregation lies not in "how many you connect" but in "whether it picks the right one automatically." Helix Compute's smart routing dynamically matches each task to the optimal path across available models and compute, weighing dimensions including:

COSTCost — the most economical option that still meets requirements
LATENCYLatency — low-latency nodes prioritized for time-sensitive tasks
QUALITYQuality — matching model capability to task complexity, neither wasteful nor compromising
PRIVACYPrivacy — sensitive workloads routed to compliant execution environments

For developers, this means that behind a single call, the whole market is bidding and matching for the task; what they get is the optimal balance of cost and experience, without managing that complexity themselves. This is what "making compute easier to use" concretely means.

From aggregation to supply: why steps two and three

As only an aggregator, Helix would hit a ceiling: its supply would depend entirely on external providers, its cost and reliability subject to others, with little durable moat. So aggregation is the starting point, not the destination.

Phase two opens the supply side, connecting idle GPUs worldwide. Its significance is not just more supply but giving Helix its "own" compute network — suppliers serve the real demand continuously generated by Helix Labs within the network (the value of the loop, see Chapter 3), rather than depending on the moods of an external market.

Phase three, building compute centers, is a natural extension once demand is fully proven. When in-network compute consumption is stable and predictable, owned capacity can backstop the most critical workloads at lower marginal cost — and put pricing power truly in the network's own hands.

An honest note: dependence on centralized compute

To be candid: for a long time to come, the most advanced GPUs will remain in the hands of centralized chipmakers and cloud providers. Any decentralized network claiming to have "fully escaped centralized compute" is, at this stage, not being honest.

Helix's stance is pragmatic. In the aggregation phase, we stand on centralized supply by design — not as a compromise, but so the network can offer real, usable service from day one. As phases two and three advance, the network's dependence on any single centralized source declines step by step — but "full decentralization" is a direction, not a marketing slogan. We would rather describe this path honestly than promise an endgame we cannot deliver at this stage.

Compute is the network's energy source, but energy must be turned into value. The next chapter enters the second pillar — Helix Labs — to see how the demand of creation is incubated.

05

Pillar II — Helix Labs

If Helix Compute is the network's energy source, Helix Labs is where that energy turns into value. It is an incubator and accelerator for the Vibe Coding era — but its real role is not an auxiliary business of the network; it is the network's "demand engine." This is what fundamentally sets Helix apart from every pure compute network.

Chapter 2 pointed to a widely overlooked gap: decentralized compute projects sit almost entirely on the supply side, with demand left to an external market. Helix Labs exists to fill that gap — it lets the network create its own demand rather than wait for demand to arrive.

From idea to launch: a complete pipeline

Vibe Coding lowers the barrier of creation to "being able to describe an idea clearly," but from an idea to a product that users can actually use and that keeps running, there is still a road to travel. Helix Labs paves that road into a complete pipeline:

01

Tools & templates

A Vibe Coding toolchain and starter templates that let creators turn ideas into working prototypes fast. These tools consume Helix Compute directly.

02

Compute subsidies

Compute support for early projects, so creators aren't deterred by compute cost at the most fragile starting stage.

03

Acceleration & resources

Accelerator programs, a mentor network and ecosystem resources to help promising projects move from prototype to mature product.

04

Distribution & cold start

Using in-ecosystem traffic and Base's distribution channels to help products get seen and clear the hardest hurdle — the cold start.

Every stage of this pipeline lowers one of the three walls creators face alone — compute cost, distribution, reward (see Chapter 2). Helix Labs turns problems creators once had to solve themselves into infrastructure the network provides.

Why it's a "demand engine," not an incubator

An ordinary incubator aims to produce successful companies; the relationship between it and its startups is one of investment. Helix Labs is different — every product it incubates is also part of the network:

A compute consumerEvery product, while running, consumes Helix Compute — forming the network's real, sustained demand.
An ecosystem entryEvery product brings its own users, who become new touchpoints for the entire network.
A source of growthSuccessful products draw in more creators, compounding the incubation capability itself.

In other words, what Helix Labs incubates is not a pile of isolated projects, but the network's own demand. This is a fundamental shift in perspective: in a pure compute network, the incubator (if any) is a cost center; in Helix, Helix Labs is the engine that makes the whole loop turn. The demand it generates feeds the compute side, and the compute side's low cost makes incubation more efficient in return — the very core force of the loop in Chapter 3.

Distribution on the shoulders of Base

The cold start is the hardest gate for any new product. A key reason Helix Labs chose to build on Base is distribution.

Base is one of the most active Ethereum L2s today. Its goal is to "bring billions of users onchain," and toward that it has built a mature consumer-app ecosystem, a low-friction wallet onboarding experience, and distribution channels with Coinbase. For products incubated by Helix Labs, this means they are born inside a distribution network aimed at mainstream users — able to draw on Base's consumer-app ecosystem, onchain social protocols and traffic entries, instead of building user reach from scratch. Placing incubated products inside an ecosystem already "bringing users onchain" is the most pragmatic solution to the cold start.

The moat: a spiral of self-generated demand

Back to the most fundamental difference. Compute aggregation is a fiercely competitive red ocean — any team with resources can connect GPUs and build an aggregator. But the loop of "compute supply × creation demand" is hard to simply copy, because it requires a team that can both aggregate and schedule compute well, and continuously incubate real demand, and make the two interlock within one network.

This is exactly where Helix's moat lies: it doesn't depend on the charity of external demand, but grows demand from within. While other networks still worry about "how to get users," Helix Labs has made the creation of demand an innate capability of the network. Two strands — compute and creation — wound together, feeding each other, rising as one: this is the true meaning of the name Helix.

Supply and demand are both in place. The next chapter looks at the third pillar that aligns them and lets the whole loop run in coordination — HLX.

06

Pillar III — HLX Token Utility

The first two pillars form supply and demand. But for supply and demand to keep interlocking within one network, a coordination mechanism is needed to align every participant's interests — this is HLX. It is the network's settlement unit, incentive bond and governance credential — the "base pair" that lets the loop run.

// SCOPE OF THIS CHAPTER This chapter describes only HLX's utility logic — what it does and why the network needs it. Specific economic parameters such as supply, allocation and vesting will be fully disclosed in a dedicated Tokenomics document later, and are out of scope for this version of the whitepaper.

Why this network needs a token

A direct question: compute aggregation, product incubation — couldn't these run on fiat or stablecoins? Why is HLX needed?

The answer is "alignment." Helix's loop has four kinds of participants — compute suppliers, AI developers, Vibe Coding creators, network users — each pursuing their own interests. For a network to run long-term, these scattered interests must point the same way: the more the network prospers, the higher every participant's return. Fiat can't do this, because it can't bind "the network's growth" to "participants' gains." A token can — being both the medium of exchange within the network and the carrier of the network's value, making those who use the network and those who own it the same people.

The four utilities of HLX

HLX plays four interrelated roles in the network:

01 · SETTLEMENT

Payment & settlement

HLX is the settlement unit within the network. Developers use it to buy compute; suppliers earn it as reward. All value flowing through the network clears at this layer.

02 · STAKING

Staking & priority

Participants can stake HLX for better compute quotas, routing priority or network rights. Staking also lets suppliers express long-term commitment, strengthening the network's stability.

03 · INCENTIVE

Incentive & distribution

HLX is the carrier of network incentives. Developers by usage and contribution, creators by their product's use and value, receive HLX incentives accordingly — the more you contribute, the more you earn.

04 · GOVERNANCE

Governance & voting

Holding HLX means a voice over the network's key parameters — compute pricing, incubation direction, treasury spending — which will gradually be decided collectively by holders (see Chapter 10).

How value flows back through the network

These four utilities are not isolated; together they form a cycle of value. The network's real usage — developers buying compute, products consuming compute, creators earning rewards — generates real network activity; and the value that activity creates flows back, by design, to the network and its participants, rather than accruing to some center.

USEReal compute consumption & product operation
ACTIVITYSettlement & value flow within the network
FLOW BACKValue returns to the network & contributors
GROWTHStronger incentives draw more participation

One basic stance of Helix bears emphasis here: HLX's value should be anchored in the network's real usage, not propped up by incentive emissions alone. This is the very point of the first five chapters — only a network with real compute demand (Compute) and a self-generated demand engine (Helix Labs) can give a token's value the support of real activity, rather than a cycle detached from use and sustained purely by subsidy. How the value-return mechanism is specifically designed will be quantified in the Tokenomics document.

// CORE POSITIONING HLX is not a speculative instrument layered on top of the network, but its settlement unit, coordination mechanism and ownership credential — aligning supply, demand and users under one set of incentives. Utility first, numbers later.

Utility first, numbers later

We deliberately discuss only HLX's utility in this version, not its economic parameters. This is a prudent choice: before chain deployment, ecosystem progress and specific mechanism design are all settled, publishing exact figures for supply, allocation and unlocks would be neither rigorous nor responsible to the community.

We would rather build the network well and explain the token's utility logic clearly first, then disclose the full economic model in a dedicated, well-considered Tokenomics document. A token's value ultimately comes from how useful the network it coordinates is — not from a set of pretty numbers.

With that, all three pillars are unfolded. The next chapter brings them together to see how the three jointly drive a self-reinforcing growth flywheel.

07

The Growth Flywheel

Each of the three pillars has value on its own; but Helix's real power lies in the loop they form together. Compute, creation and coordination are not three parallel lines but a self-reinforcing flywheel — every turn leaves the network stronger than the last. This chapter puts the three back into one picture.

The six stages of the flywheel

Helix's growth flywheel breaks down into an end-to-end loop:

HELIX LOOP SELF-REINFORCING 01 02 03 04 05 06
01
Aggregate computeCompute turns scattered compute into unified, low-cost supply
02
Lower the barrierCheap, reliable compute further lowers the cost and barrier of creation
03
Incubate productsHelix Labs incubates more ideas into real products
04
Generate consumptionEach product consumes compute and brings users — real demand
05
Feed supplyGrowing demand makes aggregating more compute and cutting cost worthwhile
06
The loop acceleratesStronger supply, lower cost — back to start, restarting at greater scale

The elegance of this loop is that it has no weak link. The output of each stage is the input of the next; every turn scales up compute, creation and demand together. This is not linear growth but compounding — the longer the flywheel turns, the greater its momentum, and the harder it is for latecomers to catch up.

Two strands, feeding each other

Reduce this flywheel to Helix's double-helix image, and it is really two strands feeding each other:

COMPUTE STRAND

The richer and cheaper compute →
the lower the barrier to creation

×
CREATION STRAND

The more creation flourishes, the more demand →
the more meaningful aggregating compute

Progress on the supply side lowers the barrier to creation; flourishing on the demand side raises the value of supply. The two strands don't grow separately and add up — they multiply: every bit of progress on one strand is amplified by the other. This multiplicative relationship is what fundamentally separates Helix from a simple bolt-together of "compute + incubation": they are not two businesses, but one interlocking system.

The hardest question: how the first turn gets going

Every flywheel faces the same challenge — it is hardest to push when at rest. What draws creators to a network with no compute scale yet? And without the demand creators bring, what draws compute suppliers? This is the classic chicken-and-egg dilemma, and the reason most two-sided networks die at the starting line.

Helix's cold-start strategy is to not depend on either side arriving naturally, but to actively ignite:

Borrow firstThe compute side starts by aggregating existing providers — the network has usable compute on day one, without waiting for suppliers to join (see Chapter 4, Phase 1).
Then create demandHelix Labs actively incubates a first batch of products, injecting the network's initial, externally-independent real demand — rather than passively waiting for developers to show up.
Leverage the ecosystemBuilding on Base's mature user and distribution ecosystem, the first products reach users faster, shortening the time to complete the flywheel's first turn (see Chapters 5 and 8).

In other words, Helix doesn't wait for "compute scale" and "creation demand" to grow naturally before the flywheel turns — it starts from aggregating existing compute plus self-generating a first batch of demand, actively pushing the flywheel through its hardest first turn. Once the loop is established, self-reinforcing compounding takes over growth. This is the decisive advantage of the "self-generated demand" loop at the cold-start stage: others must wait for the market, while Helix can ignite its own.

The flywheel explains how Helix grows. The next chapter goes more concrete — the technical architecture supporting all this, and why Base was chosen as the network's foundation.

08

Architecture & Chain Choice

This chapter gives Helix's technical panorama — but our focus is not low-level implementation detail; it's helping you understand how the network collaborates in layers, and why Base was chosen as its foundation. A good architecture should convey its skeleton in a single diagram.

A four-layer architecture

Helix's stack can be understood as four layers, top to bottom, from user touchpoint to compute supply:

LAYER 01 · ACCESS
Access layerThe unified entry for developers and creators: API, SDK, and Helix Labs' tools and console. Here users describe their needs without worrying about the complexity behind them.
LAYER 02 · ROUTING
Routing layerThe network's "brain": matching each task dynamically to the optimal model and compute by cost, latency, quality and privacy (see Chapter 4).
LAYER 03 · SETTLEMENT
Settlement layerBuilt on Base: handling HLX payment, staking, incentives and governance. All clearing and alignment of value happens here.
LAYER 04 · SUPPLY
Supply layerThe actual source of compute: aggregated third-party models and providers, GPUs connected to the network, and future self-built compute centers (see the three phases in Chapter 4).

The division is clear: the access layer governs "is it easy to use," the routing layer "is the choice right," the settlement layer "how value flows and aligns," and the supply layer "where compute comes from." Of these, the one directly tied to the blockchain is the settlement layer — it carries all of HLX's onchain logic; while compute scheduling and execution happen efficiently off-chain, with the chain handling only the clearing and coordination of value. This "onchain settlement, off-chain execution" split ensures efficiency while keeping value flows trustworthy and verifiable.

Why Base

Which chain to build the settlement layer on is a key decision. Helix chose Base — Coinbase's Ethereum Layer 2. This choice is less a technical decision than an ecosystem one.

DistributionBase is one of the most active Ethereum L2s today, aiming to "bring billions of users onchain," with a mature consumer-app ecosystem. For products incubated by Helix Labs, this is natural distribution soil.
ExperienceNative USDC and Smart Wallet compress onboarding to under a minute — crucial for a network that wants to attract mainstream creators, not just crypto-native users.
CostAs an Ethereum L2, Base offers low gas and high throughput while inheriting mainnet security — a settlement layer both cheap and trustworthy.
ChannelsSynergy with Coinbase's distribution and compliance channels, plus cross-chain bridges already opened to other ecosystems, so Helix isn't isolated on a single chain.

For a network that puts the "creator ecosystem" at its core, Base offers not just a chain but a whole environment already "bringing users onchain." Choosing Base is essentially choosing to start inside a distribution network aimed at mainstream users.

An honest balance: on decentralization

Choosing Base also means answering a question head-on: Base is led by Coinbase, and its "decentralization narrative" is weaker than Arbitrum's or Optimism's. For part of the crypto community, this is a trade-off worth weighing.

Our judgment is pragmatic. In the network's early stage, distribution and user experience decide a project's life or death more than the chain's decentralization purity. Base's advantages on both are significant, and the Ethereum security it inherits is enough to make the settlement layer trustworthy. As for "decentralization," Helix believes it should show up more in the network's own governance structure — who decides compute pricing, incubation direction, treasury spending — than merely in which chain sits underneath. Helix's path to decentralization is unfolded in the governance section of the next chapter.

// ARCHITECTURE PHILOSOPHY Onchain for settlement and coordination, off-chain for scheduling and execution; choose the most usable ecosystem underneath, and let decentralization land on the network's own governance. Pragmatic, not dogmatic.

On security, the settlement layer's smart contracts will undergo third-party audit before mainnet deployment, and changes to key parameters will be made openly through the governance process. The specifics of these safeguards will be disclosed as the network progresses.

Architecture and foundation are clear. The final three chapters give the network's evolution roadmap (Ch. 9), the governance path toward community co-governance (Ch. 10), and an honest disclosure of risks (Ch. 11).

09

Roadmap

A vision must land on verifiable steps. Helix's evolution unfolds along a clear path — from aggregating existing compute, to opening supply and incubating creation, and finally to owned compute and community co-governance. Each phase has clear deliverables and measurable milestones.

To be clear, this is a directional roadmap, not a deadline-bound promise. Specific timing will adjust with technical progress, market conditions and community consensus; we care more about what each phase delivers than about which date it lands on.

PHASE 1 · AGGREGATE

Unified compute entry goes live

Launch the LLM and compute aggregator: unified interface, smart routing, unified billing. From day one, developers call aggregated compute through one entry. Brand, community and early ecosystem are built in parallel.

Aggregator MVPSmart Routing v1Developer docs
PHASE 2 · CREATE

Helix Labs launches · HLX arrives

The Helix Labs incubator goes live, actively incubating a first batch of Vibe Coding products to inject the network's own demand. HLX is generated (TGE) and connected to settlement and incentives; the flywheel begins to turn. The settlement layer deploys on Base.

Helix Labs v1First cohortHLX TGESettlement live
PHASE 3 · OPEN

Decentralized compute market

Opening to the supply side: third-party GPU holders connect, serve real demand and earn rewards. Staking goes live, and the network expands from "connecting providers" to "an owned supply network."

Supplier onboardingStaking liveCompute verification
PHASE 4 · OWN

Owned compute centers

Once demand and cash flow are proven, build GPU clusters and compute centers as the network's owned capacity, backstopping high-priority and privacy-sensitive workloads. From moving compute to owning it.

Self-built pilotCritical-load backstop
PHASE 5 · GOVERN

Toward community co-governance

Governance is gradually handed to the community: key parameters such as compute pricing, incubation direction and treasury spending are decided collectively by HLX holders. Helix moves from foundation-led to network self-governance.

Governance frameworkParameter votingTreasury co-management

These five phases advance in layers: each builds on the validated foundation of the one before, rather than all at once. The flywheel begins turning in Phase 2 and accelerates at ever-greater scale in the phases that follow.

10

Governance & Path to Decentralization

Chapter 8 made a point: decentralization should show up more in the network's governance structure than merely in which chain sits underneath. This chapter explains how Helix gradually hands governance to the community — and why this is a step-by-step path rather than an all-at-once one.

What governance decides

The point of governance is to give those who truly use and contribute to the network a voice in its direction. In Helix, the following key parameters will be brought into community governance over time:

Compute pricingPricing rules and fee structures for aggregation and supply
Incubation directionWhich directions and projects Helix Labs' resources favor
Treasury spendingThe direction and scale of network treasury use
Incentive parametersThe rules and adjustments of incentive distribution

Step by step, not all at once

We must be honest: a network fully decentralized on day one is often not an ideal but an irresponsibility. An early network needs fast iteration and decisive calls; handing everything to onchain voting too soon would mire it in inefficiency and paralysis at its most fragile stage. Many projects use the slogan of "decentralization" to mask an absence of governance, ending up with neither efficiency nor real community power.

Helix chooses a more pragmatic path: starting foundation-led, then handing governance to the community in stages and verifiably as the network matures.

StartThe foundation leads key decisions, ensuring early iteration speed and directional consistency while keeping operations transparent.
TransitionGradually open parameter voting, letting HLX holders first take part on select issues, building the habits and mechanics of governance.
Self-governanceOnce the framework matures, key parameters and the treasury are decided by the community, with the foundation stepping back to a supporting role.

Treasury management will stay transparent, with major spending and parameter changes made in the open. We believe decentralization is a process that must be seriously built, not a state one can declare in a whitepaper. Helix's goal is to ultimately become a network truly owned and governed by its participants — but getting there takes a solid step at a time.

11

Risk Disclosure & Legal Disclaimer

Any serious project should face its risks honestly. This chapter lists the main risks Helix faces across technology, market, regulation and execution. We believe that discussing risks candidly is itself a sign of responsibility to participants.

Key risks

Technical risk

The implementation and stability of systems such as compute aggregation, smart routing and onchain settlement carry uncertainty; smart contracts may contain undiscovered vulnerabilities, and although audited, auditing cannot eliminate all risk.

Market risk

Supply, demand and pricing of AI compute are highly volatile; the decentralized compute field is fiercely competitive with several established players. Whether Helix can build and sustain a differentiated advantage is uncertain.

Supply-dependence risk

In the early stage, the network depends heavily on centralized third-party compute providers. Fluctuations in advanced GPU supply, price changes or provider policy shifts could all affect the network's ability to serve.

Regulatory risk

Regulation of crypto tokens, decentralized networks and AI is still evolving worldwide. Changes in policy could affect the network's operation, how HLX is used, or the eligibility of users in particular regions.

Execution risk

Delivering the roadmap depends on the team's execution, funding and ecosystem development. Phase goals may be delayed, adjusted, or unattainable as expected due to force majeure.

Cold-start risk

Launching a two-sided network is inherently difficult. Although Helix has designed an active-ignition cold-start strategy, whether the flywheel can successfully turn and form a self-reinforcing loop still requires market validation.

Legal disclaimer

Aggregate compute, empower creation — where supply and demand complete each other.

The DNA of AI · Helix Network

HELIX NETWORK · WHITEPAPER · DRAFT V0.1 · CH. 1–11 $HLX · Base · helixnetwork.app