Loading prices…
🔥BULLISH

Titan Network: Tencent, Alibaba Pay 75% Less for AI Compute

The pitch to consumers — 80% of enterprise revenue routed back to whoever's laptop or phone the work runs on — sidesteps the institutional-server focus of Akash and Aethir and targets 5% of Asia's AI…

Titan Network: Tencent, Alibaba Pay 75% Less for AI Compute
Titan Network: Tencent, Alibaba Pay 75% Less for AI Compute
Titan Network: Tencent, Alibaba Pay 75% Less for AI Compute
Titan Network: Tencent, Alibaba Pay 75% Less for AI Compute

Titan Network says two of the world's ten largest AI companies are running workloads on its crowdsourced computing network, paying as much as 75% less than they would on conventional cloud capacity. The infrastructure is built on roughly 4 million connected consumer devices worldwide, with about 1 million online at any given time, the company said.

Founder and chief strategy officer Konstantin Tkachuk, speaking at the Proof of Talk conference in Paris, framed the model as a direct pipeline from corporate AI spend to consumer hardware owners. Named clients include Tencent, Alibaba, and the AI video platform Kling AI. When those clients pay for data tasks — web scraping, data collection, content delivery — Titan routes 80% of the revenue back to the individuals whose devices and bandwidth did the work, paid out through a browser plugin or specialised software the user installs.

Why it matters

Titan sits in the DePIN category alongside Akash Network and Aethir, but its consumer-device focus is the structural difference. Where Akash and Aethir aggregate spare cycles on institutional servers, Titan's pitch is that every laptop, phone, and home internet connection sitting idle is a candidate node. The model only works if device operators trust the payment rail — which is where the 80% revenue share and the blockchain settlement layer come in. The company claims to have already captured roughly 5% of the AI data market in Asia, a small but non-trivial wedge against centralised hyperscalers.

Market impact

The AI infrastructure cost curve is the real driver. Training and inference workloads are still bottlenecked by GPU access and power, and any alternative that promises 75% savings gets a serious look from buyers. If Titan's numbers hold up under scrutiny, the read-through for the broader DePIN sector is bullish — consumer-side aggregation becomes investable once at least one project demonstrates enterprise-grade clients at hyperscaler-comparable cost.

Frequently asked questions

  1. What is Titan Network and how does it make money from AI?

    Titan Network aggregates unused computing resources from consumer devices via a browser plugin or specialised software, then rents that capacity to AI companies as a decentralized cloud. It claims clients pay up to 75% less than they would on conventional cloud providers, and Titan routes 80% of the corporate revenue…

  2. Who are Titan Network's named clients?

    Titan has named Tencent, Alibaba, and the AI video platform Kling AI as clients. Founder Konstantin Tkachuk also said two of the world's ten largest AI companies are using the network, without naming them.

  3. How is Titan different from Akash Network or Aethir?

    All three sit in the DePIN category of decentralised physical infrastructure networks, but Akash and Aethir aggregate spare compute from institutional servers. Titan's pitch is consumer devices — laptops, phones, and home internet connections — making the device operators the revenue recipients.

  4. How large is Titan's network today?

    Titan says it has onboarded roughly 4 million connected consumer devices worldwide, with about 1 million online at any given time. The company claims to have captured roughly 5% of the AI data market in Asia.

  5. What are the risks to Titan's business model?

    Coordinating 4 million consumer nodes is a long tail operationally, and the quality of compute available on idle consumer hardware is structurally lower than what an institutional server farm can deliver. The 75% cost-saving claim also depends on workload type — not every AI task is suitable for distribution across…

Source attribution
Aggregated from CoinDesk · Verified · Last refreshed 45d ago
Open original →