Tencent's AI chief tells staff outlasting rivals beats moving first
Two stories today about playing the long game: Tencent's cloud boss has written an unusually candid internal essay conceding his firm looks slow on AI — while in India, a voice-AI startup just convinced Peak XV that a phone call is only where the work begins.
Dowson Tong, who runs Tencent's Cloud and Smart Industries Group, has published an internal essay confronting the charge that Tencent is falling behind in AI — and arguing the race is a marathon, not a sprint. The piece appeared Tuesday evening in Zhi Dian, Tencent's internal annual journal, and Tong disclosed he co-wrote it with AI, describing models as collaborators in thinking rather than mere execution tools. He is candid about the mood inside the company: rivals are louder in the enterprise market, and compute shortages have dragged both the pace of Hunyuan model iteration and product rollouts, so internal anxiety is real. His bet is that the industry is still young enough — market structure and business models unsettled — that enduring matters more than being early, and that sustained investment aimed at genuine customer scenarios is what carries a company through the cycle.
His prescription leans on Tencent's existing strengths rather than raw model ambition. Algorithms set a product's ceiling, he argued, but engineering capability decides how quickly anything actually lands — and ownership of real-world scenarios is the company's core advantage, with long-termism and human-centered design named as the two constants through the change. The essay marks the third time this year Tencent leadership has answered the "slow" question: chairman Pony Ma told shareholders in May that the company's early AI foundations were weak but steadily improving, and Tong conceded at a June cloud conference that different business lines move at different speeds, failures included.
The takeaway: an admission that compute constrained a hyperscaler's model lab is rare anywhere, and rarer still in China's brag-heavy market. Patience is a defensible story only if Hunyuan starts shipping visibly better products — and Tencent's recent poaching of an xAI multimodal lead suggests Tong intends to buy speed as well as wait for it. We covered the hire when an xAI multimodal lead joined Tencent's Hunyuan model team.
Voice-AI startup Ringg has raised $10 million from Peak XV Partners, extending its Series A to $15.5 million as it pushes beyond simple phone automation. The company began life as a text-to-speech outfit called DesiVocal, found training its own speech models too expensive, and climbed up the stack into enterprise voice agents — signing fintech Cred as its first customer, then Flipkart, Practo, Groww, and PolicyBazaar. It now processes roughly 20 million call attempts a month and runs appointment booking across 1,200 clinics for Practo alone.
Co-founder Siddharth Tripathi told TechCrunch the early volume work — outbound calling, lead qualification, loan collection — was a price game with nothing sticky in it, so Ringg is deliberately chasing harder workflows: clinic scheduling, abandoned-cart recovery, and onboarding checks for fintech apps, plus chat, WhatsApp, and browser-based support for clients like Shell. The pitch is an agent platform measured by outcomes rather than a voice vendor measured by minutes — a meaningful distinction in a crowded field where ElevenLabs, Deepgram, Sarvam, and Smallest.ai are all fighting over the same enterprise workflows. The quiet thesis of the round: in voice AI, defensibility accrues to whoever owns the customer's outcome, not whoever owns the flashiest model.
What to watch: whether Tencent's next Hunyuan release lands fast enough to make the marathon metaphor stick — and whether Ringg's outcome-first pitch converts at Global Capability Centers, its chosen beachhead beyond India.
Is "we're slower but we'll outlast you" a strategy or a cope when your rivals have more GPUs? Tell us in the comments.
