Tech Disruptors

“Microsoft is making a bet that we’re not going to need a single AI, we’re going to need many different AIs” Sebastien Bubeck, Microsoft’s vice president of generative-AI research, tells Bloomberg senior technology analyst Anurag Rana. In this Tech Disruptors episode, the two examine the differences between a large language model like ChatGPT-4o and a small language model such as Microsoft’s Phi-3 family. Bubeck and Rana account for various use cases of the models across various industries and workflows. The two also compare the costs and differences in compute/GPU requirements between SLMs and LLMs.

The cloud’s need for a high-performing AI infrastructure to support growing large language models has opened a Pandora’s box of performance bottlenecks. In this edition of Bloomberg Intelligence’s Tech Disruptors podcast, Ayar Labs CEO Mark Wade and VP of Commercial Operations Terry Thorn sit down with BI hardware and networking analyst Woo Jin Ho to discuss the company’s co-packaged silicon photonics that aims to disrupt conventional AI architectures and leapfrog the current wave of input/output technologies in the rapidly evolving computing market.

Nick Frosst, co-founder of Cohere, discusses the unique approach to model training like retrieval augmented generation, tokenization strategies, and challenges in debugging language models. They explore the significance of large language models with text prompts, focusing on efficient problem-solving in enterprise applications and the debate on model size vs synthetic data for AI development.

The workflow-management-solutions sector is evolving fast, yet remains fragmented and likely on the hunt for new revenue models. In this episode of the Tech Disruptors podcast, Thomas Scott, CEO of Wrike, joins Sunil Rajgopal, senior software analyst at Bloomberg Intelligence, to discuss how AI tools are used to boost client and internal productivity, the demand environment and the company’s product journey. The two also talk about Wrike’s horizontal market approach, future plans and potential pricing models as compute-intensive AI features and workloads grow.

“On average, a salesperson spends 30-45 minutes after a sales call on processing the sales call,” says Ketan Karkhanis, Salesforce’s executive vice president and general manager of Sales Cloud. In this installment of Tech Disruptors, Karkhanis sits down with Anurag Rana, Bloomberg Intelligence senior technology analyst, to examine the problem of time-consuming and unproductive tasks faced by salespeople, and how Salesforce and its AI solutions can help address this challenge. Additionally, the two discuss the trend of software consolidation, customer evolution on the Salesforce platform and the power of its developer ecosystem.

The rising capabilities of AI and progress in large language models are set to reduce communication barriers posed by language diversity. In this episode of the Tech Disruptors Podcast, Jaroslaw Kutylowski, founder and CEO of DeepL, joins Sunil Rajgopal, senior software analyst at Bloomberg Intelligence, to discuss how machine learning is reshaping language translation, along with competing language-translation tools and enterprise adoption. The two also talk about DeepL’s platform makeup, business model and the company’s expansion plans beyond Europe.

AI computing employs GPU/XPU clusters that are growing at unprecedented speeds, as well the bandwidth and power requirements of networks to support them. Near Margalit, Broadcom’s vice president and general manager of the Optical Systems Division, sits down with Bloomberg Intelligence’s senior semiconductor analyst Kunjan Sobhani in this Tech Disruptors podcast episode to discuss Broadcom’s innovations in optical networking, specifically on the role of its co-packaged optics (CPO) and how it could emerge as preferred solution for solving the bandwidth, power and cost challenges of AI infrastructure.

CTO of Precisely, Tendü Yoğurtçu, discusses data integration, location intelligence, and gen AI. They explore the challenges companies face in managing data, the importance of data quality, and Precisely's position in the market. Yoğurtçu highlights the need for AI-ready data and the product initiatives focusing on data integrity and customer communications.

“We’re seeing a big increase in GPU capacity demands and we expect the data centers over the next couple of years will need to support higher power densities giving us the ability to run very large clusters of GPUs,” says AWS Global Data Centers VP Kevin Miller. In this episode of Bloomberg Intelligence’s Tech Disruptors podcast, Miller talks with BI senior technology analyst Anurag Rana about all things cloud infrastructure, including data-center demand for AI, the timeline to construct new centers, the differences in the stress that training and inferencing place on the hardware and ways in which AWS is keeping the environment in mind as they approach their 2040 goal of carbon neutrality.

Expanded artificial-intelligence workloads are raising the power and cooling requirements of data centers, with individual rack consumption predicted to increase by 10x compared with traditional compute. Vertiv is a market-leading provider of UPS, liquid cooling and switchgear solutions catering to these needs. In this episode of the Tech Disruptors podcast, Vertiv CEO Giordano “Gio” Albertazzi sat down with Bloomberg Intelligence analysts Mustafa Okur and Woo Jin Ho to discuss the opportunities and challenges in the data-center infrastructure market.