Silicon Valley AI startups are increasingly merging with major tech giants like Microsoft and Amazon. Due to financial constraints, many promising companies such as Inflection AI and Adept have seen key executives move to these tech giants through discreet deals. These transactions, often viewed as acquisitions, aim to bypass competition regulators.
Character AI and French startup Mistral struggle to secure the funding needed to remain independent. Even OpenAI, the creator of ChatGPT, is deeply tied to Microsoft through a $13 billion investment deal, ensuring exclusive access to its advanced models. Amazon has similarly invested in Anthropic to secure high-performing AI models.
The immense computing power required for developing generative AI, which can produce human-like content rapidly, necessitates substantial financial resources. As a result, many AI startups, founded by former leaders of major tech firms, rely on the support of large cloud providers to recreate the conditions of well-funded research labs. The shift like this one deviates from the traditional Silicon Valley startup narrative.
However, the consolidation of AI innovation under a few tech giants raises concerns about competition. Critics argue that aligning with these companies stifles creativity and innovation. Government regulators in the US, EU, and UK are scrutinising these deals, with recent actions indicating a growing regulatory interest in ensuring fair competition within the nascent AI industry.
Britain’s antitrust watchdog is examining Google-parent Alphabet’s partnership with AI startup Anthropic to assess its impact on market competition. The scrutiny comes amid growing global concerns about the influence of major tech companies on the AI industry following the AI boom sparked by Microsoft-backed OpenAI’s release of ChatGPT.
Regulators are scrutinising deals between big tech giants and AI startups, including Microsoft’s collaborations with OpenAI, Inflection AI, and Mistral AI, as well as Alphabet’s investments in companies like Anthropic and Cohere. Anthropic’s AI models, developed by former OpenAI executives Dario and Daniela Amodei, compete with OpenAI’s GPT series.
Last week, the UK’s Competition and Markets Authority (CMA) joined forces with US and the EU regulators to ensure fair competition in the AI sector. The CMA is now inviting public comments on the Alphabet-Anthropic partnership until 13 August before deciding whether to initiate a formal investigation. The CMA’s decision will be based on feedback during this initial consultation.
At SIGGRAPH, a major computer graphics conference, Nvidia presented new real-world applications of generative AI. Chief executive officer Jensen Huang highlighted the company’s role in AI development, emphasising their Nvidia Inference Microservices (NIM) platform. Nvidia has always prioritised advanced computing through a software-led approach.
Recent announcements showcased improvements in generative AI and 3D content generation. AI services and models are now available to accelerate humanoid robot development. Researchers can use devices like the Apple Vision Pro to teach robots various tasks. Collaborations with Getty Images and Shutterstock aim to improve the accuracy of AI-generated images matching text prompts.
Engineers now benefit from advancements in industrial design, visualisation, and advertising tools. A demo video displayed lifelike 3D worlds generated from simple text prompts. Coca-Cola and marketing agency WPP are among the early adopters of Nvidia’s generative AI art tools.
The importance of these developments extends beyond product outputs. Nvidia integrates AI into their own processes, aiding software debugging and chip design. The impact on the market has been substantial, contributing significantly to the S&P 500’s market capitalisation gains. The company’s efforts continue to shape the future of AI in various industries.
Samsung Electronics is making strides in developing memory chips essential for the AI market, narrowing the gap with rival SK Hynix. The company has recently received approval from Nvidia for its HBM3 memory chips and anticipates approval for its next generation, HBM3E, within months. The advancement follows months of setbacks, including development challenges and replacing the head of its semiconductor division.
Samsung’s efforts come as the demand for high-bandwidth memory (HBM) is expected to soar, driven by AI advancements. The HBM market is projected to grow from $4 billion in 2022 to $71 billion by 2027. Nvidia’s approval is crucial for Samsung to capitalise on this booming market and improve its revenue and market share despite still trailing SK Hynix.
Why does this matter?
The company has faced significant engineering challenges, particularly with the thermal management of the stacked DRAM chips used in HBM. Under the leadership of Jun Young-hyun, Samsung has focused on resolving these issues and enhancing its technology. The company has also reorganised its HBM team to boost innovation and collaboration.
As Samsung progresses, it aims to ramp up production and meet the growing demand for AI memory chips. With its financial resources and production capacity, the company is well-positioned to address market shortages and secure a significant share of the lucrative AI memory market.
The United Arab Emirates recently cancelled meetings between US Congressional staffers and the Emirati AI firm G42 following US lawmakers’ concerns about the potential transfer of US AI technology to China. The intervention was personally handled by the UAE Ambassador to the US, who stopped the meetings involving the House Select Committee on China.
The cancelled meetings suggest diplomatic tension as US lawmakers, particularly those wary of China, examine the implications of the AI technology transfer to the Middle East. The US State Department has not commented, while a G42 spokesperson directed questions to the Emirati government. The UAE embassy’s spokesperson cited a miscommunication around the visit since the embassy officials were only made aware of the staff delegation shortly before it was about to arrive.
The staffers’ visit was intended to discuss the transfer of advanced chips from companies like Nvidia to the UAE and Saudi Arabia and the broader US-China tech competition. Amidst these concerns, the Biden administration has defended the G42-Microsoft deal, noting that it led to G42 cutting ties with China’s Huawei.
Meta Platforms announced the launch of AI Studio, a tool enabling users to create and design personalised AI chatbots. The new feature allows Instagram creators to develop AI characters to manage direct messages and story replies, enhancing user interaction on the platform. These AI characters can be shared across Meta’s various platforms and are built using Meta’s Llama 3.1 model. This latest version of Meta’s AI model is available in multiple languages and competes with other advanced models like OpenAI’s.
Why does this matter?
Meta’s initiative follows OpenAI’s confidential project, code-named ‘Strawberry,’ aiming to showcase advanced reasoning capabilities. Introducing AI Studio marks Meta’s effort to offer cutting-edge AI tools to its vast user base, leveraging its Llama 3.1 model to provide powerful AI-driven features for content creators and users alike.
According to a recent research paper, Apple has opted to use Google-designed chips instead of Nvidia’s for two crucial components of its AI software infrastructure. This choice is noteworthy as Nvidia is widely regarded as the leading provider of AI processors. The paper detailed that Apple employed Google’s tensor processing units (TPUs) in large clusters, specifically 2,048 TPUv5p chips for AI models on devices like iPhones and 8,192 TPUv4 processors for server models.
The research paper did not mention any use of Nvidia chips, despite Nvidia dominating about 80% of the AI processor market through its graphics processing units (GPUs). Unlike Nvidia, which sells its GPUs directly, Google offers TPUs through its Google Cloud Platform, requiring customers to use Google’s platform for access.
Why does this matter?
Apple has begun introducing parts of its new AI suite, Apple Intelligence, to beta users. This recent publication only disclosed the full extent of Apple’s reliance on Google hardware, despite earlier reports hinting at this partnership.
Apple’s engineers noted the potential for even larger, more sophisticated AI models using Google’s chips. However, Apple’s stock saw a minor decline of 0.1% to $218.24 following the research paper’s release.
Amazon.com is pushing forward with the development of its own AI chips, aiming to reduce its dependence on Nvidia and lower costs for its customers. Engineers in Amazon’s Austin, Texas, chip lab are currently testing a new server design packed with these proprietary AI chips. Rami Sinno, director of engineering for Amazon’s Annapurna Labs, highlighted the increasing demand for more affordable alternatives to Nvidia’s products.
Amazon acquired Annapurna Labs in 2015 and has since focused on creating processors that can handle complex calculations and large data sets more economically. As competitors like Microsoft and Alphabet also develop their own AI processors, Amazon’s initiative is a strategic move to maintain its edge in the AI cloud business, which is the main growth driver for Amazon Web Services (AWS).
Amazon’s new AI tools are also part of its broader strategy to tackle misinformation and enhance service offerings. This development comes as Nvidia adapts to tightening US export controls by creating China-specific AI chips.
The National Institute of Standards and Technology (NIST) has re-released Dioptra, a tool designed to measure AI model risks, particularly from data poisoning attacks. The modular, open-source web-based tool, originally launched in 2022, aims to help companies and individuals assess and analyse AI risks. It can be used for benchmarking, researching models, and exposing them to simulated threats, offering a common platform for these activities.
NIST has positioned Dioptra to support government agencies and businesses in evaluating AI system performance claims. The tool’s release coincides with new documents from NIST and the AI Safety Institute that outline ways to mitigate AI-related dangers, including the generation of non-consensual pornography. This effort is part of a broader US-UK partnership to advance AI model testing, which was announced at the UK’s AI Safety Summit last year.
The development of Dioptra aligns with President Joe Biden’s executive order on AI, which mandates comprehensive AI system testing and the establishment of safety and security standards. Companies developing AI models, such as Apple, are required to notify the federal government and share safety test results before public deployment.
Despite its capabilities, Dioptra has limitations. It only works with models that can be downloaded and used locally, such as Meta’s expanding Llama family. Models that are accessible only via an API, like OpenAI’s GPT-4, are currently not compatible. Nonetheless, NIST proposes that Dioptra can highlight which types of attacks might degrade an AI system’s performance and quantify their impact.
Shanghai has announced a massive $13.8 billion investment to bolster its integrated circuit, biomedicine, and AI industries. The biggest city in China is making a significant push to solidify its position as a global tech leader.
The AI sector will benefit significantly, with investments directed towards intelligent chips, software, autonomous driving, and intelligent robots. The following initiative is part of Shanghai’s broader strategy to foster innovation and build a competitive edge in the global market.
Funds will support original innovation, enhancing Shanghai’s technological capabilities. This move is expected to accelerate the development of globally competitive enterprises within the city, driving growth and attracting talent.
By focusing on these key sectors, Shanghai aims to solidify its reputation as a leading commercial hub. The strategic investment underscores the city’s commitment to advancing technological innovation and economic development.
