Uber Questions AI Spending

AI Training Data Startup Secures $8.2 Million in Funding

Human Archive, an AI training data provider, has raised $8.2 million from top investors including Wing Venture Capital and OpenAI. The company collects real-world task data by partnering with gig economy workers, who use devices to record their activities. This data helps train AI systems, like robotic arms used in manufacturing. Currently, Human Archive works mainly with Indian gig platforms but is expanding globally through a cloud service. The startup also develops hardware, such as cameras and gloves, to enhance data quality. If successful, this approach could challenge other data providers and even synthetic data generators.

AI in Telecom Faces Data Challenges Despite High Adoption Rates

AI adoption in the telecom industry is booming, with 97% of executives using it for customer experience, network operations, and cost reduction. Yet, many initiatives struggle to scale beyond pilot projects. The issue isn't the technology itself-models are advanced, and processing power is ample-but rather the quality and accessibility of data. Telecoms generate vast amounts of information, but much of it remains fragmented or poorly organized, making it hard for AI systems to use effectively. The World Economic Forum highlights "data debt" as a major hurdle: outdated, inconsistent, or inaccessible data hinders AI's potential. This isn't just a technical problem-it amplifies existing organizational challenges. If your team can't efficiently manage data, AI won't magically fix it. Instead, it reflects and magnifies the chaos. To overcome these barriers, telecoms need to focus on building robust data infrastructure with unified access and semantic clarity. While many have adopted lakehouses, most still miss critical unstructured data like logs and contracts. Addressing this gap will unlock AI's full potential, helping companies navigate their unique operational landscapes and achieve meaningful scale.

Universal Music Group and TikTok Renew Licensing Agreement

Universal Music Group and TikTok renewed their licensing agreement. The agreement includes removing unauthorized AI-generated music from the platform. This matters because AI-generated music can mimic artists' voices and exploit streaming algorithms. Some AI-generated tracks have racked up millions of streams before being taken down. The music industry is worried about this issue. TikTok will work with Universal Music Group to improve artist and songwriter attribution and ensure they get paid. TikTok will continue to work on these issues in the future.

Perceptic Raises $12 Million for AI Drug Development Platform

Perceptic emerged from stealth and announced a $12 million seed funding round. The company is building an end-to-end AI platform for drug development. The company's software is already being used by multiple top-tier pharmaceutical companies. Perceptic says its system can compress scientific due diligence from weeks to hours. Perceptic is targeting areas like scouting external assets and clinical trial design. The company says its data foundation has produced a 50-fold increase in clinical data extractions. Perceptic will continue to develop its AI platform to speed up drug discovery.

Big Tech Pays Itself in Cloud Loop

Big Tech companies are paying themselves through their own cloud bills, sparking AI bubble fears. This is happening through a round-trip funding loop where a tech giant invests in an AI startup and the startup uses that money to pay for cloud services from the same tech giant. The numbers are significant, with OpenAI and Anthropic holding over half of the roughly $2 trillion in future cloud commitments from Microsoft, Amazon, Google, and Oracle. OpenAI's annual cloud bill has reportedly ballooned past $60 billion, while its actual revenue sits closer to $25 billion. This loop is legal under current accounting rules, but it raises concerns about the sustainability of the AI boom. As AI startups begin to operate outside of this protected loop, they will face budget meetings and real revenue expectations, which could lead to a burst in the AI bubble, and the impact will be seen in the future.