How The Middle East’s Startup Scene and Gulf SWFs Are Reshaping Venture Capital
PLUS: Former Stripe Business Lead Building Something New
Episode of the Week
In Episode #115 of Geeks of the Valley, we sat down with Khalid Saad, a seasoned venture capitalist and FinTech entrepreneur with a strong track record in launching and scaling ventures across payments, remittance, crypto, and open banking. He is the Founder & Managing Partner of Bunat Ventures, a Bahrain-based venture builder VC investing in high-growth GCC startups. Khalid also serves on the boards of CoinMENA (a regulated crypto exchange) and Oqal Bahrain (the region’s largest angel network), advises 01 Systems, and sits on the Bahrain Chamber’s Finance, Insurance & Tax Committee.
Previously, he was the Founding CEO of Bahrain FinTech Bay, MENA’s largest FinTech hub, uniting around 100 public and private stakeholders and hosting over 50 FinTech companies. Earlier in his career, Khalid worked at the Bahrain Economic Development Board promoting financial services, at Ernst & Young conducting feasibility and market studies, and at SEI Investments in London managing UK and European equity portfolios.
This edition is brought to you by VARA Network
Built on the Gear Protocol - Vara Network is a revolutionary Web3 application platform, removing many of the barriers associated with building, deploying, and operating dApps. Vara’s computer science paradigms like persistent memory and actor model, along with parallel data processing, allows for deep scalability, new design patterns and features. Vara’s unique architecture makes it completely frictionless for users. Learn more at Vara.Network
Venture Radar
Meta is reportedly in talks to acquire Play AI, a fast-growing voice-cloning startup that enables ultra-realistic replication of human voices with minimal training data. While this news has yet to dominate headlines, the underlying implications signal a deeper strategic pivot for Meta: advancing from text-based AI interfaces toward fully multimodal, human-like communication systems.
At the core of Play AI’s technology is its proprietary small-footprint voice model, capable of cloning a person’s voice with just a few seconds of audio input. Unlike traditional text-to-speech systems that rely on extensive datasets and heavy compute, Play AI’s lightweight architecture allows for real-time inference on consumer devices. This unlocks seamless integration into existing Meta products like WhatsApp voice notes, Instagram Reels narration, or Messenger voice assistants.
The immediate application is clear: enabling personalised and scalable content creation. Imagine creators instantly generating multiple voiceovers for multilingual audiences or businesses deploying branded voice agents for customer support without the robotic intonation of legacy systems. For Meta, this acquisition could accelerates its race to deliver more immersive and humanised AI experiences across its apps.
However, the strategic implications go beyond user-facing features. Meta has been investing heavily in AI infrastructure, from open-sourcing its Llama language models to developing on-device AI chips. Play AI complements this push by adding native voice synthesis capabilities, allowing Meta to build an integrated stack for text, image, and audio generation under its control. This reduces dependence on external APIs while deepening its moat in social and creator ecosystems.
Voice cloning also carries controversial ethical and regulatory considerations. The risk of misuse for fraud, deepfake scams, or reputational attacks is high, particularly given Play AI’s minimal training data requirements. Meta’s acquisition will place it under scrutiny to implement robust watermarking, detection, and user consent frameworks to prevent abuse—a challenge the company has historically struggled with in content moderation.
Strategically, this move aligns with Meta’s broader vision of ambient AI agents embedded throughout daily life. Integrating realistic synthetic voices into AI assistants or AR/VR avatars will make interactions feel far more natural and personal, a critical step toward Meta’s metaverse ambitions. Combined with generative video and image models, Meta is effectively building the toolkit for fully synthetic yet human-like digital experiences.
In the near term, expect Meta to deploy Play AI’s models within its creator tools and messaging apps, offering enhanced voice features that drive engagement and platform stickiness. In the longer term, voice cloning becomes an infrastructural layer for Meta’s AI agent ecosystem—one that brings its products closer to human communication’s most natural medium: voice itself.
While the acquisition may appear as a feature upgrade today, its deeper meaning is structural. Voice is not just an input or output modality; it is a primary interface for how humans interact with each other—and soon, with AI.
Geeks of the Week
Startup Name: Papr
Geography: US
One-liner: Intelligent, blazing-fast memory and context retrieval API for AI agents.
Founder(s) Background: Senior PM at Apple.
Thoughts:
Addressing a foundational gap in AI agent reliability: Papr tackles one of the key limitations of current AI systems – their inability to retain, retrieve, and reason over large knowledge bases with speed and accuracy. While standard vector-based retrieval systems handle single-hop queries well, Papr’s focus on multi-hop RAG with graph reasoning enables agents to synthesise information across disparate data points, a core bottleneck in production-grade AI applications.
Unlocking practical deployment of AI agents at scale: By combining semantic vector search with graph-based knowledge structures, Papr delivers fast, contextually rich memory APIs. This is critical for use cases like customer support, document intelligence, and operational agents where reliability and accurate recall determine user trust and adoption.
Founder(s) building in stealth
This edition is brought to you in partnership with Stella Capital.