Category: Podcast Episode

Nataraj welcomes Jeff Tatarchuk, Co-founder & CEO of TensorWave, to discuss the evolving landscape of AI compute and the rise of new cloud strategies. TensorWave is positioned as a key player in providing AI compute, exclusively working with AMD GPUs to address the growing demand.

From FPGAs to AMD GPUs: A Cloud Evolution

TensorWave started as an FPGA cloud business (VMXL), gaining valuable experience in deploying cloud infrastructure and managing data centers. Recognizing the shift towards AI and the GPU shortage, they pivoted to focus on AMD GPUs.

The decision to go all-in on AMD was driven by customer demand for alternatives to NVIDIA and the belief in AMD's vision and roadmap. TensorWave aims to be the premier AMD support, making it easy for customers to deploy AMD chips at scale.

The Challenges of Building Modern Data Centers

Access to power is the major bottleneck in building data centers today. The demand for power far exceeds the available supply, requiring creative solutions and strategic partnerships.

Building quickly and deploying rapidly is crucial, as customers are willing to pay a premium for fast access to compute. However, challenges such as permitting issues and workforce availability can impact timelines.

Financing and the Competitive Landscape

Raising capital is essential for building data centers, but it's only one piece of the puzzle. Cost of capital, customer credit ratings, and strategic partnerships are all critical factors.

The AI cloud market is becoming crowded, but TensorWave differentiates itself through its technical expertise, deep integration with AMD, and focus on solving the challenges of deploying AMD GPUs at scale.

Overcoming the CUDA Moat and Driving AMD Adoption

While NVIDIA's CUDA has been a significant advantage, most AI engineers use frameworks like PyTorch and TensorFlow, which allow for seamless code porting to AMD GPUs.

TensorWave is actively building an ecosystem around AMD, partnering with researchers and engineers to showcase the capabilities and performance of AMD GPUs. They also launched the 'Beyond CUDA' summit to promote innovation outside the CUDA ecosystem.

AMD GPUs offer architectural advantages, such as more memory (VRAM), making them suitable for hosting larger models. Their chiplet architecture also allows for greater flexibility and efficiency.

The Future of TensorWave: An AI Utility Company

TensorWave aims to be a resilient, secure, and performant cloud provider that customers can rely on for their AI training and inference needs. They want to abstract away the complexities of managing GPUs, allowing customers to focus on their core business.

Success for TensorWave means providing viable options for customers to buy compute, fostering a competitive market, and playing a significant role in democratizing access to AI infrastructure.

Nataraj welcomes Eilon Reshef, Co-Founder and CPO of Gong, to the Startup Project. Eilon's background includes co-founding Webcollage, a SaaS platform acquired in 2013, showcasing his experience in the tech industry. Gong leverages AI to analyze customer interactions and sales conversations, helping teams boost productivity and drive efficient growth. The conversation will explore Gong's origins, product-market fit, and the transformative impact of AI on their business.

The Genesis of Gong and the Power of Data

Eilon explains that Gong's vision has remained consistent since its founding in 2015: to bring data-driven workflows to the revenue space, which was traditionally treated as an art. The core idea was that AI could make sales processes more efficient, but only with access to quality data. Gong started by capturing conversations with customers, integrating with platforms like WebEx (later Zoom, email, and CRMs) to create a revenue graph. This graph would then be used to apply logic and help teams be more productive and leaders gain better intelligence.

In the early days, recording calls wasn't common, and many VCs were skeptical that people would want to be recorded. Gong had to develop its own recording technology, including a bot that joins calls. Initially, transcription accuracy was poor, but Gong quickly moved to a homegrown system. The focus was on building a robust, cloud-first system to capture and analyze these conversations.

Achieving Product-Market Fit

Eilon emphasizes the importance of focusing on a small niche to achieve product-market fit. Gong initially targeted software-as-a-service companies in North America selling in English over video conferencing. This narrow focus allowed them to refine their product and meet the specific needs of their early adopters. Even though transcription wasn't perfect, early adopters found value in search capabilities. This allowed them to track mentions of competitors or specific topics within conversations.

Collaboration tools were also key, allowing salespeople to easily tag colleagues and share relevant moments from calls. A pivotal moment occurred when 11 out of 12 design partners paid for the product after a beta period, signaling strong product-market fit. This early success validated Gong's approach and paved the way for future growth.

From Conversational Intelligence to Revenue AI Operating System

Gong has evolved from analyzing single conversations to providing a comprehensive revenue AI operating system. This includes pipeline management, forecasting, sales engagement, coaching, and enablement. By expanding its capabilities, Gong now serves a wide range of customers, from small companies to large enterprises, across various industries. The goal is to help everyone along the customer journey, from prospecting to post-sales, and even product managers.

Eilon discusses the idea of AI-driven coaching and training. While AI can't turn everyone into a top performer, it can move the average performers up the curve. Gong's AI trainer simulates customer interactions and provides coaching based on actual conversations, helping sales reps improve their skills and effectiveness.

The Future of AI Agents and Enterprise Software

Eilon believes that AI agents will gradually take on specific tasks within sales workflows, augmenting human capabilities rather than replacing them entirely. These agents can handle tasks like walking customers through proposals or answering common questions, freeing up sales reps to focus on more complex interactions. Gong is developing tools that allow companies to train AI agents based on historical data, ensuring they are effective and aligned with business goals.

Eilon also touches on the trend of commoditization of software and the exaggerated claims surrounding AI's capabilities. While AI can improve engineering productivity and enable rapid prototyping, it won't eliminate the need for skilled developers or comprehensive software solutions. He emphasizes the importance of focusing on real-world applications and delivering tangible value to customers.

Fundraising and Product Leadership

Eilon shares his approach to fundraising, emphasizing the importance of starting with a product and demonstrating product-market fit. Gong followed a traditional route, raising capital to build a team and acquire customers. The company also took advantage of favorable market conditions in 2021 to raise additional funds for future growth.

As Chief Product Officer, Eilon prioritizes customer centricity and close collaboration with design partners. He believes that understanding customer needs is crucial for developing valuable products. Gong's product development process involves extensive customer feedback and iteration, ensuring that every feature adds value.

Trends and Exciting Developments

Eilon discusses the importance of embracing AI to improve individual job performance. He encourages professionals to identify tasks that can be enhanced with AI, rather than fearing job displacement. He also highlights the potential of AI to transform enterprise software, but notes that the industry is still in the early stages of figuring out how to best leverage AI's capabilities.

Gong is launching an orchestration product that automates tasks and guides revenue professionals, helping them prioritize opportunities and drive change. Eilon envisions a future where companies can use AI to analyze their business processes and implement changes with the click of a button, enabling rapid optimization and personalization.

Nataraj welcomes Andrew Bialecki, CEO of Klaviyo, to discuss how the company has reshaped customer connections for B2C brands. Klaviyo started as an email marketing company and evolved into a platform integrating marketing attribution, SMS marketing, and AI capabilities. Today, Klaviyo is a public company valued around $9 billion, serving 180,000 brands globally and generating impressive revenue with substantial year-over-year growth.

From Database to Marketing Powerhouse

Andrew Bialecki explains Klaviyo's mission: to enable businesses to treat every customer as the most important one. Initially, Klaviyo started as a database business, aiming to replicate human thinking and information storage. The company then expanded into marketing automation after realizing customers were using the database for marketing purposes.

Klaviyo's platform now extends to customer service with its customer agent, aiming to empower businesses to deliver exceptional experiences at scale through technology. The goal is to help businesses act as their best sales representative and product expert through technology.

Finding Product-Market Fit and the Shopify Partnership

Andrew Bialecki knew Klaviyo was on the right track when a haberdashery customer promised to triple their spending if Klaviyo expanded into marketing. A pivotal moment was integrating with Shopify, which opened up a significant opportunity in the e-commerce space.

Klaviyo focused on a partner model for customer acquisition, emphasizing word-of-mouth, marketing agencies, and platform integrations. Shopify's forward-thinking approach to an app store allowed Klaviyo to help businesses with marketing and customer experiences.

Capital Efficiency and the Decision to Go Public

Klaviyo was notably capital-efficient, raising around $400 million but spending only a fraction before going public. Andrew Bialecki advises founders, especially those with technical backgrounds, that they often need less capital than they think. He emphasizes building a great product and acquiring customers, which attracts the best investors.

While many companies go public to fundraise, Klaviyo's decision was driven by factors like providing liquidity to investors and telling its story more directly. The company wanted to maintain its long-term vision and customer focus, believing its strong culture would withstand the transition.

The Rise of AI and the Future of Customer Engagement

Klaviyo had a successful holiday season, processing billions of data points and powering billions of customer experiences. A key trend is the evolution of retail brands into value-added services, offering expertise and support around their products.

Klaviyo is leveraging AI through its marketing and customer agents. The marketing agent automates the entire marketing process, from identifying trends to generating creative content. The customer agent uses customer context to answer questions and provide personalized recommendations.

Andrew Bialecki emphasizes that AI is increasing the overall amount of marketing and customer service engagement. He believes that AI will enable software to run itself, allowing businesses to focus on higher-level strategies.

AI's Impact on Company Culture and Hiring

Klaviyo's bootstrapped beginnings instilled a DNA of generalism and automation. The company seeks employees with a wide range of skills and encourages them to automate repetitive tasks. AI is seen as another iteration of this philosophy.

Andrew Bialecki emphasizes the importance of curiosity and an AI-first mindset. He encourages everyone to integrate AI into their workflows and values those who are actively building AI applications.

Nataraj welcomes Ilya Levtov, CEO of Craft, to the Startup Project podcast. The episode dives into the complexities of global supply chains and how Craft is helping governments and enterprises build more resilient systems. Ilya's background includes stints at Goldman Sachs and Venrock, providing a unique perspective on the challenges and opportunities in the supply chain space.

The Evolving Landscape of Global Supply Chains

Ilya notes that the pandemic exposed the brittleness of global supply chains, which had become overly reliant on outsourcing and just-in-time inventory management. While the pandemic brought attention to the issue, the focus has shifted to more secular concerns, recognizing the inherent complexity and dynamism of global supply chains.

He emphasizes that we are still in the early stages of truly mapping and understanding supply chains with data and intelligence.

Key Concerns and Risk Dimensions

Ilya highlights several key risk dimensions that are top of mind for supply chain leaders. These include cybersecurity, geopolitical risks (such as weaponization of the supply chain and foreign ownership concerns), ESG considerations, and traditional financial stability concerns.

He emphasizes that large enterprises often struggle to integrate disparate data sources and build comprehensive risk governance capabilities. Craft aims to solve this by delivering a holistic approach to supplier management and risk mitigation.

Ensuring Data Integrity and Proactive Risk Mitigation

Ilya explains that Craft addresses data integrity by using an “all of the above” strategy, incorporating multiple data sources for validation. This allows for comparison and flagging of potential inaccuracies, enabling targeted investigation.

Craft's platform helps companies identify and mitigate cyber risks by providing scores and qualitative attributions from leading cyber scanning companies. This allows companies to quickly identify vulnerable suppliers and implement mitigation efforts. Continuous monitoring ensures that even suppliers that appear safe are immediately flagged if their risk profile changes.

Craft's AI-Powered Intelligent Workspace

Ilya discusses how Craft leverages AI to enhance its platform, particularly through its Intelligent Workspace. This AI-driven tool provides procurement professionals with prioritized alerts on critical supplier events, such as financial distress, cyber vulnerabilities, and labor violations.

AI also significantly accelerates the creation of risk reports, reducing the time from days to seconds. Craft utilizes a range of existing LLMs, including OpenAI, Anthropic, and Google Gemini, and is exploring the development of its own supplier risk models to better cater to specific customer needs and risk tolerances.

Craft's Journey and Future Outlook

Ilya shares Craft's initial go-to-market strategy, which involved offering free company profiles to attract organic traffic and generate leads. This approach led to early adoption by government agencies and large enterprises.

Looking ahead, Craft aims to expand its supplier intelligence layer across the entire procurement and supply chain function, fostering collaboration and reducing silos within enterprises. The company also plans to enhance data egress and ingress, enabling seamless integration with other enterprise systems.

Nataraj interviews Andrew Feldman, CEO of Cerebras, about the company's mission to revolutionize AI compute. Cerebras is challenging established giants like NVIDIA with its wafer-scale engine, designed for unprecedented AI compute power and efficiency. The conversation explores the current AI landscape, the shift to open-source models, and the future of sovereign AI.

The Genesis of Cerebras

Founded in 2015, Cerebras aimed to address the emerging need for specialized AI compute. Feldman and his team, with a background in building chips and systems, recognized the potential for a new computer architecture tailored to AI workloads. They predicted that AI would usher in a new era of compute, similar to how cell phones and networking equipment created new demands and opportunities.

Despite NVIDIA's dominance at the time, Cerebras believed it could build a new type of computer that would be orders of magnitude faster. This vision led to the development of the wafer-scale engine, a groundbreaking innovation in AI hardware.

Building the World's Largest AI Chip

Cerebras embarked on a mission to build the largest chip in the history of the computer industry. This involved fundamental design, creativity, engineering, innovation, and invention. After three years and half a billion dollars, the company created its first wafer-scale engine.

The larger size allowed Cerebras to keep more data on the chip, reducing the need for frequent data movement and minimizing power consumption. This resulted in significantly faster performance compared to traditional chips. The initial period was challenging, with 15 months of high spending and no working chip, but the team persevered and eventually achieved a breakthrough.

The Advantage of On-Chip Memory

Cerebras's architecture utilizes SRAM, a fast type of memory, to overcome the limitations of traditional memory solutions. By stuffing the chip with SRAM, Cerebras achieved both capacity and speed, resulting in significantly faster performance. This is particularly beneficial for inference, where large amounts of data need to be moved quickly to generate results.

The memory bandwidth of GPUs is a known bottleneck, and Cerebras's design addresses this by placing the SRAM right next to the compute core. This allows for much faster data movement, leading to significant improvements in inference speed.

Product Strategy and Market Approach

Cerebras sells complete computer systems, not just chips, to provide comprehensive solutions for AI compute. These systems can be deployed on-premise or accessed through the cloud, either through Cerebras's own cloud or through partner clouds like Amazon Marketplace and Microsoft Marketplace.

The company also offers forward-deployed engineering services to help customers accelerate the delivery of AI solutions. This comprehensive approach sets Cerebras apart from other chip companies and allows it to cater to a wide range of customer needs.

The CUDA Challenge and Open Source

While CUDA has been a significant barrier for new chip companies, it is becoming less relevant in the inference space. Developers can easily switch to Cerebras's platform with minimal code changes, making it an attractive option for those looking to improve inference performance.

The rise of open-source models is further weakening CUDA's dominance, as developers can leverage these models without needing to know CUDA. This is creating new opportunities for innovation and making AI more accessible to a wider range of users.

Sovereign AI and Geopolitics

Cerebras is strategically partnering with sovereign institutions like G42 in the UAE to build AI infrastructure and develop models tailored to specific regions and languages. This reflects the growing importance of sovereign AI, as nations seek to build domestic AI capabilities.

Feldman believes that the US should sell chips and AI systems to its allies but not to its adversaries. This highlights the geopolitical considerations that are shaping the AI landscape and the importance of aligning AI development with national interests.