Uncategorized
ZK Prover Code “Venus” Released to Dramatically Reduce L2 Fees and Scale Web3
San Francisco, USA
Open‑source software released today could significantly lower transaction fees on zero‑knowledge rollups, potentially ending the trade‑off between low cost and strong security that has long defined Ethereum layer‑2 scaling.
The code, named “Venus,” was published by Cysic, a verifiable compute network. It is a hardware‑optimized proving backend built on top of Zisk, an existing open‑source zkVM originally developed by the Polygon Hermez team (now independent). By accelerating the generation of zero‑knowledge proofs on GPUs, FPGAs, and eventually ASICs, the new code addresses the primary reason ZK‑rollups have remained more expensive than their optimistic counterparts.
The Persistent Cost Disadvantage of ZK‑Rollups
Optimistic rollups such as Arbitrum, Optimism, and Base offer low fees and fast throughput but require a seven‑day withdrawal window to allow for fraud proofs. ZK‑rollups provide instant finality and stronger cryptographic guarantees, but their proof generation has historically been computationally heavy – a cost passed on to end users.
As a result, many developers and users have opted for optimistic rollups, accepting slower withdrawals and weaker security models to save on fees. The industry has treated this as an unavoidable compromise.
Zisk Advanced ZK Proving – But Hit a Hardware Ceiling
Zisk, released by the Polygon Hermez team, represented a major leap forward. Its segmented proving model and RISC‑V‑based architecture achieved roughly 10‑fold faster trace generation than any other public zkVM at the time. For the first time, a general‑purpose zkVM could run on consumer hardware.
However, Zisk was designed as a CPU‑first system. Its GPU support was added through a hardware abstraction layer that treated the proving pipeline as an opaque black box. This resulted in high CPU‑GPU synchronization overhead, inefficient kernel launches, and no ability to globally optimize the proof flow. Support for FPGAs and ASICs – which can run ZK operations orders of magnitude faster – was absent.’
Zisk made ZK proving faster, but not fast enough to close the cost gap with optimistic rollups.
Venus: Hardware Acceleration for Zisk
To solve this problem, Venus replaces Zisk’s CPU‑first proving pipeline with a visible, composable computational graph. Instead of a single opaque prove() call, every operation – field arithmetic, hashing, Merkle tree updates – becomes a node that can be globally optimized.
Key features of the Venus release:
- Native CUDA Graph integration – Allows GPUs to run autonomously, eliminating CPU‑GPU synchronization delays. Early benchmarks on an RTX 5090 show a measurable throughput improvement over the Zisk baseline, with further gains expected.
- Full FPGA backend – Includes HLS kernels for Goldilocks field arithmetic, NTT, Poseidon2, Merkle trees, FRI, and expression evaluation, targeting AMD UltraScale+ and Versa l devices with HBM.
- Preliminary ASIC design – An initial implementation of a custom silicon path for zkVM proof acceleration.
- Open‑source licensing – Released under Apache 2.0 / MIT, the same permissive licenses as Zisk.
The code is under active development and has not yet been audited for production use. Cysic has invited the community to test, contribute, and build upon it.
What the Release Means for Layer‑2 Fees
If adopted by rollup operators and prover networks, Venus could significantly reduce the cost of generating ZK proofs. Lower proof costs would translate directly to lower transaction fees on ZK‑rollups – potentially making them price‑competitive with optimistic rollups for the first time.
More importantly, the release removes the economic penalty for choosing instant finality and cryptographic security over a seven‑day withdrawal window. Users would no longer have to choose between cheap and safe.
Industry observers note that widespread integration will take time. Rollup operators must update their prover infrastructure, and the code requires further testing and auditing. However, the release of Venus establishes a clear path toward a future where ZK‑rollups are no longer a premium product.
About the Release
Venus is available today on GitHub. It is built on Zisk, which remains under the stewardship of SilentSig Switzerland GmbH (the independent company spun out from Polygon Hermez). Cysic has contributed its hardware acceleration work back to the open‑source community under the same dual license.
GitHub: github.com/cysic-labs/venus
Quote from Cysic
“ZK‑rollups have always been the superior scaling solution – except for cost. Venus is our contribution to closing that gap. By open‑sourcing a hardware‑accelerated backend for Zisk, we are giving everyone the tools to run efficient provers. This is how we make ZK the default, not the premium option.”
Leo Fan, Founder of Cysic
About Cysic
Cysic is a verifiable compute network that combines custom hardware (ASICs, GPUs, FPGAs) with a decentralized prover marketplace to accelerate zero‑knowledge proof generation. The company’s infrastructure powers rollups, AI verification, and privacy‑preserving identity systems.
Website: cysic.xyz | Twitter: @cysic_xyz
Uncategorized
Schwab Earning Cash Management Concerns Weigh on Stock Despite Robust Results
WESTLAKE, TX
The Charles Schwab Corporation today reported strong quarterly financial results, showcasing solid revenue growth and continued client engagement. However, investor sentiment remained cautious as concerns surrounding the company’s cash management dynamics weighed on its stock performance.
The company delivered better-than-expected earnings, driven by higher net interest revenue, increased trading activity, and steady asset inflows. Schwab’s diversified business model and expanding client base contributed to overall resilience in a competitive financial services landscape.
Despite these positive indicators, analysts and investors expressed concerns over ongoing shifts in client cash allocation. The migration of client cash into higher-yielding alternatives has put pressure on Schwab’s net interest margins, a key driver of profitability. This trend, while industry-wide, has raised questions about the sustainability of earnings growth in the near term.
“Our performance this quarter reflects the strength of our client-first strategy and the durability of our business model,” said a Schwab spokesperson. “At the same time, we are actively managing through a dynamic rate environment and evolving client preferences.”
Schwab emphasized its proactive approach to balance sheet management, including adjustments to funding strategies and continued focus on expense discipline. The company also highlighted its commitment to delivering long-term value through innovation, technology investments, and client service enhancements.
Market reaction to the earnings release was mixed, with shares experiencing pressure following the announcement, reflecting investor sensitivity to cash sorting trends and margin outlook.
Looking ahead, Schwab remains focused on navigating macroeconomic uncertainties while capitalizing on growth opportunities across its core business segments. The firm reaffirmed its long-term strategy centered on client engagement, operational efficiency, and sustainable financial performance.
About The Charles Schwab Corporation
The Charles Schwab Corporation is a leading provider of financial services, offering a full range of brokerage, banking, and advisory solutions to individual investors and institutions worldwide.
Uncategorized
UK Testing Specialist Intertek Rejects EQT’s £9.3bn Takeover Proposal
London, UK
Intertek Group plc, a leading provider of quality assurance, testing, inspection, and certification services, today confirmed that it has rejected a £9.3 billion takeover proposal from EQT AB.
The Board of Intertek carefully reviewed the unsolicited proposal and determined that it significantly undervalues the company and its long-term growth prospects. As a result, the Board has unanimously decided to reject the offer, stating that it is not in the best interests of shareholders.
Intertek emphasized its strong market position, diversified global operations, and continued momentum across key sectors, including consumer goods, healthcare, and industrial services. The company highlighted its strategic initiatives and robust financial outlook as key drivers of sustainable value creation.
“We remain confident in Intertek’s strategy and future growth trajectory,” said a spokesperson for the company. “Our focus continues to be on delivering high-quality services, expanding our global footprint, and generating long-term value for our stakeholders.”
EQT AB, one of Europe’s largest investment organizations, has been actively pursuing opportunities in the testing, inspection, and certification sector as part of its broader investment strategy. However, Intertek’s leadership indicated that the current proposal does not adequately reflect the company’s intrinsic value or future potential.
Intertek reassured investors that it remains open to opportunities that align with its strategic vision and shareholder interests but reiterated that any proposal must fully recognize the company’s value.
About Intertek Group plc
Intertek Group plc is a leading Total Quality Assurance provider to industries worldwide. With a network of more than 1,000 laboratories and offices across over 100 countries, Intertek delivers innovative and bespoke assurance, testing, inspection, and certification solutions for customers’ operations and supply chains.
Uncategorized
Happy Horse 1.0 Surges to No.1 in Pure Visual Quality on Artificial Analysis Video Arena
Montgomery, Alabama
Happy Horse has surged to the top of the Artificial Analysis Video Arena — the trusted blind human-vote Elo leaderboard — in pure visual quality. While other models focus primarily on motion control and audio synchronization, Happy Horse 1.0 prioritizes what truly makes videos feel premium: hyper-realistic textures, film-grade lighting, rich color grading, and artistic consistency that gives every frame the polish of a $10 million Hollywood production.

Major players such as Kuaishou and ByteDance continue to deliver strong technical advances, yet HappyHorse 1.0 stands out as the model that prompts viewers to pause and say, “This actually looks real.” Tests conducted on the generator at happy-horse.art confirm next-level output, including soft volumetric lighting, intricate material details, natural skin tones, and exceptional character consistency across shots — all achieved without additional prompting.
Artificial Analysis Video Arena: What It Is and Why It Matters
The Artificial Analysis Video Arena remains the most respected blind-testing platform for video generation models. Real creators vote on side-by-side clips without knowing which model produced them. Elo scores update in real time based on thousands of preference votes, providing a transparent benchmark free of self-reported metrics or marketing claims.
It is especially effective at separating hype from actual performance in categories focused on pure visual beauty.
Current Leaderboard Snapshot (Apr 15, 2026) – Pure Visual Quality
|
Rank |
Model |
Elo Score |
Key Strength |
|
1 |
Happy Horse 1.0 |
1295 |
Photorealism & cinematic beauty |
|
2 |
Kling 3.0 |
1289 |
Motion Control & physics |
|
3 |
Dreamina Seedance 2.0 |
1225–1275 |
Multimodal + audio sync |
|
4 |
SkyReels V4 |
1141 |
Speed |
Source: Artificial Analysis Video Arena (live data). Happy Horse 1.0 also leads in multiple Image-to-Video and aesthetic subcategories.

What We Know About Happy Horse 1.0
All information in this section is drawn directly from the official Happy Horse platform at happy-horse.art. While leaderboard results and live demos are publicly available, comprehensive independent third-party verification of every technical specification remains ongoing as of April 15, 2026.
Core Claims (Architecture / Parameters / Functionality)
- Hyper-Photorealistic Visual Engine: Trained specifically for maximum aesthetic fidelity, the model delivers film-like lighting, volumetric god rays, intricate textures, realistic skin/hair/cloth physics, and professional color grading.
- Exceptional Character & World Consistency: It maintains industry-leading persistent identity across shots — consistent faces, bodies, clothing, and style with minimal drift. World coherence (lighting, shadows, atmosphere) remains stable even in multi-shot generations.
- Cinematic Quality at Scale: Outputs support 2K resolution, 5–15 second clips, and multiple aspect ratios. The model handles text-to-video, image-to-video, and reference-image control using natural language prompts.
- Artistic Intelligence: The system demonstrates strong understanding of mood, style references (cinematic, anime, documentary, etc.), and subtle emotional tones.
- Production-Ready Output: Generations are clean and highly detailed, typically requiring almost no post-production for visual polish.
What’s Still Unverified / The Gap
- Exact parameter count and full training methodology (the team provides minimal model cards).
- Depth of native audio and lip-sync features (visuals remain the primary focus; audio performance is solid but not the headline capability).
- Maximum clip length beyond 15 seconds without using chaining tools.
- Open-source status: the model is currently fully closed and proprietary.
Access Status
- Claims vs Reality: Fully live with generous free credits.
- Demo / Try Now: Available at the dedicated model page.
- API: Available for high-volume users.
- Open Weights / Local: No — cloud-only.
Builder Implications / What This Means for Creators and Teams
Creators looking for instant results can jump straight into the HappyHorse AI Video Generator and start generating with generous free credits—no credit card required. For the latest examples, technical details, and cinematic demonstrations, visit the official Happy Horse 1.0 model page.
For curious creators, the recommendation is to visit the HappyHorse AI Video Generator and test the free credits. The results immediately illustrate why the model ranks #1 in visual beauty.
For production users working on advertisements, music videos, short films, social content, or branding projects, Happy Horse AI is the preferred choice when premium visuals are essential. Its aesthetic quality can significantly reduce time spent on editing and color grading. Users can begin at Happy Horse 1.0 AI and scale through Pro plans as needed.
For developers, the robust reference system and strong consistency make the model well-suited for building visual-first AI tools and pipelines.
This distinction matters because most AI video tools still carry an identifiable “AI-generated” look. Happy Horse is among the first models that consistently allow audiences to forget they are watching AI-generated content — a development that transforms client work and audience engagement.

FAQ
Can users try Happy Horse 1.0 right now? Yes. New users receive free credits instantly at the Happy Horse 1.0 AI Video Generator. No credit card or waitlist is required, and the platform works on any device.
Is Happy Horse 1.0 truly better than Kling 3.0 or Seedance 2.0? In blind visual quality votes, yes. Kling 3.0 leads in motion control and Seedance 2.0 in multimodal and audio capabilities, but when jaw-dropping beauty and realism are the priority, Happy Horse 1.0 currently holds the #1 position for good reason.
Does it support commercial use? Yes. All outputs come with full commercial rights, making the model suitable for client projects, advertisements, and monetized content.
How strong is the character consistency and lighting? Best-in-class. Characters remain on-model across cuts, lighting feels natural and cinematic, and textures and materials appear tactile. Reviewers frequently note that the output does not look AI-generated.
What about longer videos or advanced motion control? The model is optimized for high-quality 5–15 second clips. Creators can use multi-shot mode or chain generations for longer sequences. Motion performance is strong but not the primary differentiator compared with dedicated motion-control models like Kling.
Is Happy Horse 1.0 open-source? No. It is a closed, proprietary model engineered for maximum visual performance.
How fast is generation? Most 2K clips complete in under 40 seconds on the web generator; API access offers even faster processing.
Bottom line: Happy Horse 1.0 is worth testing today and is production-ready when cinematic quality and visual beauty are the top priorities. Its #1 ranking in pure aesthetics appears well-deserved.
Further updates are expected as longer-context versions and additional features are released.
Related Articles
- Happy Horse 1.0 vs Kling 3.0: Visual Beauty Blind Test
- What Makes AI Video Look “Cinematic”? The Happy Horse Advantage
- Best AI Video Tools for Creators in 2026
HappyHorse AI provides the cleanest and fastest way to access the current #1 visual-quality AI video model. It is highly recommended for projects where stunning aesthetics are essential.
-
Uncategorized4 months agoEscape Timeshare Fees Releases Consumer Guidance on Interpreting BBB Profiles in the Timeshare Exit Industry
-
Uncategorized7 months agoKirill Dmitriev: Global Investment Strategist and Architect of International Partnership
-
Uncategorized9 months ago
Live with Purpose Ranked #1 Show in Binge Networks’ Top 10 for July
-
Uncategorized4 months agoTreasureNFT: Partnering with BlackRock Capital for a Major Upgrade – NOVA Platform Aims to Become the World’s Largest NFT Trading Ecosystem
-
Entertainment & Sports2 years agoRachael Sage Releases Powerful Reimagined Acoustic Album, Another Side
-
Business1 year agoAivista Quant Capital CEO Dr. Smith: Tariff Policies Trigger Wrongful Sell-Off in Quality Assets, ETH Below $1,400 Severely Undervalued, Targeting Over $4,500 by Year-End
-
Business2 years agoGlobal Academic Excellence with XI TING’s Professional Tutor Team
-
Politics2 years agoMusk Claims Trump Interview Targeted by Cyber Attack
