AI News August 18, 2026: Nestle's AI-Powered Snacks, Billions for Bots, and Underwater Data Centers
📅 August 18, 2026 · 5 min read · GuideGuru News Desk
Today, we dive into how AI is shaping everything from our diets to global tech infrastructure. Nestle is leveraging AI to create specialized foods for weight-loss drug users, while tech giants like Alibaba are making massive strategic pivots towards AI. The sheer financial growth of AI companies continues to astound, even as ethical questions arise from AI's data hunger.
⚡ Today at a glance
- Food giant Nestle is using AI to develop **new products for users of GLP-1 weight-loss drugs**, aiming to address side effects and nutritional needs.
- South Korea is pushing an innovative solution to the AI boom's energy demands by planning an **underwater data center** that uses seawater for cooling.
- Chinese tech giant Alibaba is **selling its gaming arm for $1.5 billion** to double down on its AI and cloud computing strategy.
- AI model developer Anthropic continues its rapid growth, with its **annualized revenue surging to an estimated $65 billion**.
- Amazon faces criticism as reports indicate it is **destroying rare physical books** to extract data for training its advanced AI models.
- Fobi AI unveiled **FORTRESS**, a new platform for secure, enterprise-owned AI.
- **Meta** faces "astronomical" legal consequences in a critical California lawsuit.
- **Alibaba** launched HappyShrimp, an AI model that generates music from text prompts.
- Running AI is getting cheaper, but the **infrastructure costs** for its growth are surging.
- **Amazon** is reportedly destroying rare books to use their content for AI training.
- Hollywood and TikTok's ByteDance reached the **first-ever AI copyright deal**.
- The **White House confirmed September 24** for its next major AI summit.
- AI labs are shifting focus, now **competing on price** rather than just raw intelligence.
- OpenAI rolled out a **safer ChatGPT version specifically for teens**.
- **Anthropic is eyeing a massive $2 trillion IPO**, showcasing incredible industry growth.
- **GPT-5.6 Sol** sees a **50% price cut**, making advanced models more affordable and fostering innovation.
- AI companies are reportedly **destroying antique books** after scanning them for training data, raising ethical concerns.
- **Predicate** launches a platform using **voice and vital signs** to help spot early health risks.
- **Hollywood and ByteDance** sign the **first AI copyright deal**, a landmark agreement for intellectual property in AI-generated content.
- The **top 1% of U.S. businesses** spend a median of **$7,400 per employee on AI**, highlighting a significant investment gap.
🍏 Nestle looks to develop new products to serve users of weight-loss drugs
New Launch
AI to help create specialized foods for a changing market.
Food and beverage giant **Nestle is using artificial intelligence and nutritional science to develop a new line of specialized products** for the millions of consumers taking GLP-1 weight-loss drugs like Ozempic and Wegovy. These drugs have caused a shift in consumer eating habits, leading Nestle to innovate. AI is playing a crucial role in analyzing research and identifying optimal nutrient combinations to address common side effects such as muscle loss and facial fat.
This move highlights how rapidly evolving health trends are driving AI adoption in traditional industries. Nestle is proactively adapting its business model, turning a potential threat to packaged food demand into a new market opportunity, demonstrating AI's power in understanding complex biological and nutritional data for product development.
What it means for you: Expect to see new food products on shelves specifically designed for people on weight-loss medications, formulated with AI to meet their unique nutritional needs and counter side effects.
🌊 South Korea plans underwater data centre as AI boom strains land and power
Business
An innovative solution to AI's massive energy and space demands.
In a groundbreaking move to address the escalating infrastructure demands of the AI boom, **South Korea is developing plans for a large-scale underwater data center** off the coast of Ulsan. This ambitious project aims to house up to **100,000 servers**, leveraging the surrounding seawater for efficient and sustainable cooling. The initiative comes as South Korea faces mounting constraints on available land, reliable power supply, and grid capacity, all of which are increasingly strained by the energy-intensive operations of AI data centers.
As AI models grow in complexity and usage, the need for data centers is exploding, posing significant environmental and logistical challenges. Underwater data centers offer a novel solution, not only for cooling but potentially for reducing land use and improving energy efficiency. This project underscores the creative engineering required to sustain the rapid growth of the global AI industry.
What it means for you: The digital infrastructure supporting your favorite AI tools is getting a radical redesign. Innovations like underwater data centers are critical for keeping AI running efficiently and sustainably as demand continues to skyrocket.
💰 Alibaba to sell gaming arm for $1.5 billion in boost to AI pivot
Business
Tech giant shifts focus and capital towards AI and cloud.
Chinese tech behemoth **Alibaba is undergoing a major corporate restructuring, selling its gaming arm for an estimated $1.5 billion** to intensify its strategic pivot towards artificial intelligence and cloud computing. Under CEO Eddie Wu, the company is divesting non-core assets to free up capital and focus resources on what it views as its top strategic priorities. Alibaba aims to achieve an ambitious **$100 billion in AI revenue within the next five years**, highlighting the immense importance of AI to its future.
This move signifies a broader trend among major technology companies globally: a concentrated effort to reallocate resources and capital towards the most promising and lucrative areas of AI development. For Alibaba, this means strengthening its position in cloud services and AI models, including its popular Qwen family of AI models, which recently hit 3 billion downloads.
What it means for you: Big tech companies are betting big on AI, even if it means shedding other profitable businesses. This intense focus could lead to faster innovation in AI tools and services that you use.
📈 Anthropic’s annualized revenue surges to $65B
Business
Another AI powerhouse shows explosive financial growth.
Anthropic, a leading developer of advanced AI models including the popular Claude series, continues to demonstrate remarkable financial success. Reports indicate that the company's **annualized revenue has surged to an astounding $65 billion**, adding an impressive $18 billion to its revenue run rate in just the last two months alone. This explosive growth underscores the significant demand for cutting-edge generative AI technologies across various industries.
The rapid revenue acceleration of companies like Anthropic showcases the immense value and utility businesses are finding in powerful AI models. As enterprises integrate AI into their operations, customer service, and product development, the companies providing these foundational technologies are reaping substantial financial rewards, further fueling investment and competition in the AI space.
What it means for you: The AI market is experiencing unprecedented growth, with companies like Anthropic capturing huge value. This means more competition and faster development of the AI tools you interact with daily.
⚠️ Amazon, which started off selling books, is destroying rare texts to train AI
Regulation
An ethical dilemma arises from AI's insatiable data appetite.
In a controversial development, Amazon, a company that began its journey selling books, is reportedly **destroying rare and valuable physical texts to extract data for training its large language models (LLMs)**. While LLMs have already processed most readily available digital content, rare books offer unique, undigitised insights and information crucial for pushing the boundaries of AI capabilities. The process involves physically dismantling the books to scan every page for data input, effectively rendering the originals unusable.
This practice raises significant ethical and cultural concerns about the preservation of historical artifacts and intellectual property. Critics argue that the pursuit of advanced AI should not come at the cost of irreplaceable cultural heritage. As AI models require increasingly vast and diverse datasets, the methods of data acquisition are coming under closer scrutiny, potentially leading to future regulations on how companies collect and utilize information, especially from copyrighted or rare sources.
What it means for you: The data behind your AI tools might be coming from unexpected, and sometimes ethically questionable, sources. This highlights the ongoing debate about AI's hunger for information and the balance between innovation and preservation.
🔒 Fobi AI Launches FORTRESS, a Sovereign AI Platform for Secure, Enterprise Owned Intelligence (Financial Post)
New Launch
Keeping your company's AI data safe and under your control just got easier with a new "sovereign AI" platform.
Fobi AI has introduced **FORTRESS**, a new Sovereign AI Platform designed to help enterprises keep their intelligence securely owned and managed internally. Launched on August 17, 2026, FORTRESS transforms a company's proprietary data into a secure intelligence layer, crucial for the emerging "Agentic AI Era" where AI agents work autonomously with sensitive information.
As businesses increasingly adopt AI for critical operations, concerns about data privacy, security, and intellectual property remain paramount. Platforms like FORTRESS aim to give companies full control over their AI models and data, reducing reliance on third-party cloud providers for sensitive computations. This approach is vital for industries with strict regulatory requirements, ensuring compliance and preventing data leakage.
What it means for you: If you work with sensitive company data, new "sovereign AI" tools mean your company can use AI without worrying about its private information leaving your control.
⚖️ Meta faces ‘astronomical’ consequences as legal fight reaches critical moment in California (CNBC)
Regulation
A major lawsuit in California could have massive implications for Meta and how tech companies manage user data.
Just weeks after a setback in New Mexico, **Meta** is now facing a far more significant legal challenge in California. This critical case, co-led by California's attorney general, could result in "astronomical" consequences for the tech giant. Details of the specific allegations were not fully disclosed in the headline, but previous lawsuits against Meta often involve data privacy, antitrust, or content moderation practices.
This legal battle is part of a broader trend of increased scrutiny on large tech companies, especially regarding their AI-powered platforms and data handling. The outcome could set important precedents for how social media companies operate, collect data, and develop AI features, potentially influencing regulations and user privacy protections across the industry. Such large-scale legal actions highlight the growing tension between rapid technological advancement and the need for robust oversight.
What it means for you: Big tech companies like Meta are under intense legal pressure. The results of these lawsuits could change how your data is collected and used by the AI tools you interact with every day.
🎶 Alibaba Launches AI Music Generation Model HappyShrimp in Beta (Yahoo Entertainment)
New Model
Alibaba's new AI can turn your text ideas into full musical tracks, putting a powerful creative tool in more hands.
Alibaba Group Holding Ltd. has released the beta version of its new artificial intelligence music-generation model, **HappyShrimp**. This innovative tool can transform a simple text prompt into a fully produced piece of music, offering a new avenue for creative expression for both professionals and hobbyists. This move deepens Alibaba's pivot into AI, further expanding its offerings beyond e-commerce.
AI music generation is rapidly evolving, making sophisticated music creation more accessible. HappyShrimp could empower content creators, filmmakers, and even everyday users to produce custom soundtracks without needing extensive musical knowledge or expensive equipment. This development highlights the increasing integration of AI into creative industries, potentially democratizing music production and inspiring new forms of artistic work.
What it means for you: You might soon be able to create custom background music for your videos, podcasts, or even just for fun, just by typing what you want.
💸 AI is getting cheaper to run, but powering its growth is getting pricier (Business Standard)
Business
While individual AI model runs are more efficient, the overall cost of building and maintaining AI infrastructure is skyrocketing.
A recent report indicates a curious dichotomy in the AI economy: while the cost of running individual AI models is becoming more efficient and thus cheaper, the overall cost of powering AI's explosive growth is increasing significantly. This surge in price is driven by the escalating demand for **data centers, advanced computing hardware, and electricity** needed to support ever-larger and more complex AI systems across the globe, including in India.
This trend suggests that while AI software development might become more affordable for developers, the foundational infrastructure required to host and scale these AI services will continue to be a major investment. This could lead to a concentration of AI power in companies with vast resources, potentially impacting accessibility and competition in the long run. It also puts pressure on energy grids and calls for more sustainable computing solutions.
What it means for you: AI services might get cheaper for you to use in some ways, but the massive energy and hardware needed to run them mean the AI industry itself is facing huge bills, which could affect prices down the line.
📚 Amazon, which started off selling books, is destroying rare texts to train AI (AI News & Artificial Intelligence | TechCrunch)
Business
Amazon is reportedly destroying unique physical books to extract their valuable content for training advanced AI language models.
In a surprising and controversial revelation, **Amazon**, a company that began its journey selling books, is reportedly destroying rare physical texts. The purpose is to digitize their content, which is then used to train large language models (LLMs). These rare books are highly valuable for AI training because their content is unique and not already widely available on the internet, providing novel data for sophisticated models.
While the demand for diverse and extensive datasets for AI training is immense, this practice raises significant ethical questions about cultural preservation, intellectual property, and responsible AI development. Critics argue that destroying rare physical artifacts for digital data, even for AI, is an irreversible loss. This story highlights the often-hidden and morally complex methods employed to fuel the rapid advancement of AI technologies.
What it means for you: The AI models you use are trained on vast amounts of data, sometimes sourced in controversial ways that raise ethical questions about preserving our cultural heritage versus advancing technology.
🤝 Hollywood Cuts First-Ever AI Copyright Deal with TikTok’s ByteDance
Regulation
A landmark agreement signals a new era for AI and creative content.
In a groundbreaking move, Hollywood studios and TikTok's parent company, **ByteDance**, have finalized the **first-ever AI copyright deal**. This agreement comes months after Disney issued a cease-and-desist letter in February, accusing ByteDance of illegally using its iconic characters like Star Wars and Marvel in AI-generated content.
This deal sets a crucial precedent for how original creative works will be protected and licensed in the age of generative AI. It indicates that content creators and AI platforms are beginning to find common ground, potentially paving the way for more partnerships and clearer guidelines on AI-generated content using copyrighted material.
What it means for you: Expect clearer rules on what AI can create using existing characters and stories, potentially leading to more officially licensed AI content and fewer legal battles.
🏛️ White House Confirms AI Summit Date for September 24
Regulation
Global leaders to converge on critical AI policy discussions.
The White House has officially confirmed that its highly anticipated AI summit will take place on **September 24**. This gathering is expected to bring together global leaders, tech executives, and policymakers to discuss the future of artificial intelligence, focusing on its impact on regulations, international competition, and technological innovation.
These summits are vital for shaping the global conversation around AI, ensuring responsible development and addressing concerns about safety, ethics, and economic impact. Decisions made here could influence everything from how AI is integrated into national infrastructure to standards for data privacy and algorithmic transparency.
What it means for you: Key discussions at this summit could lead to new policies and international agreements that affect how AI tools are developed and used globally, potentially enhancing safety and ethical guidelines.
💰 AI Labs Stop Competing on Smarts and Start Competing on Price
Business
The cost of powerful AI models is dropping dramatically, making them more accessible.
A significant shift is underway in the AI industry: labs are now focusing heavily on **competing on price** rather than just on who has the "smartest" model. According to Stanford’s AI Index, the cost of querying an AI model comparable to **GPT-3.5** plummeted from **$20 per million tokens in November 2022 to just $0.07 per million tokens by October 2024**. This represents an astounding over-280-fold reduction in just two years.
This dramatic decrease in cost is a game-changer. It means that powerful AI capabilities are becoming incredibly affordable, pushing them further into everyday applications and making them accessible to a much broader range of businesses and individual users. As AI becomes cheaper to run, we'll see it integrated into more products and services without a premium price tag.
What it means for you: AI tools will become much cheaper to use and integrate, leading to more AI-powered features in your everyday apps and devices without breaking the bank.
👨👩👧👦 OpenAI Launches a Safer ChatGPT for Teens
New Launch
A long-awaited version aims to provide age-appropriate AI interaction.
OpenAI has finally introduced a **safer version of ChatGPT specifically designed for teens**. This new offering includes age-appropriate safety measures, enhanced parental controls, and learning tools tailored to steer younger users away from harmful content and discourage academic dishonesty. This launch comes years after teens initially started using the platform, often without specific safeguards.
This move reflects a growing industry focus on responsible AI use among younger demographics. By providing curated content and protective features, OpenAI aims to foster a constructive environment for teens to explore and learn with AI, while also addressing parental concerns about exposure to inappropriate material or misuse for cheating.
What it means for you: If you have teens using AI, this new version offers greater peace of mind with built-in safety features and parental controls, promoting responsible use.
📈 Anthropic’s $2 Trillion IPO Dream Rests on Staggering Revenue Bet
Business
A major AI player seeks an unprecedented valuation, testing market confidence.
The artificial intelligence boom is about to face one of its biggest tests, as **Anthropic**, the company behind the popular Claude AI models, is reportedly projecting a staggering **$2 trillion valuation** for its upcoming Initial Public Offering (IPO). This ambitious target is built on incredibly aggressive revenue growth projections, placing Anthropic as one of the most highly anticipated tech debuts ever.
If successful, this IPO would solidify Anthropic's position as a titan in the AI landscape, potentially valuing it higher than many established tech giants. It signals immense investor confidence in the future of generative AI, but also highlights the speculative nature of current AI valuations, where future potential is weighted heavily over present-day earnings.
What it means for you: The incredibly high valuation for a major AI company like Anthropic shows just how much faith investors have in AI's future, which can fuel more innovation and development in the tools we use.
💸 GPT-5.6 Sol Pricing Cut by 50%
New Model
Get ready for more affordable and powerful AI.
**OpenAI** has announced a **50% price cut** for its flagship model, **GPT-5.6 Sol**, on the **Openrouter.ai** platform. This significant reduction brings the cost down to **$2.50 per million input tokens** and **$15 per million output tokens**, making one of the most advanced AI models considerably more accessible. The model also boasts an impressive **1,050,000 token context window** and a maximum output of **128,000 tokens**, alongside higher uptime from three providers.
This aggressive pricing strategy indicates a maturing AI market where competition is driving down costs for powerful models. For developers and businesses, this means it's now more cost-effective to integrate cutting-edge AI capabilities into their applications, potentially accelerating innovation across various sectors. Users can expect to see more sophisticated AI tools emerging as the barrier to entry for high-performance models lowers.
What it means for you: **Cheaper access to top-tier AI models** means more innovative AI tools will be available and easier to use for everyone.
📚 AI Companies Are Buying—And Destroying—Antique Books. Here’s Why.
Business
The hidden cost of AI training data might surprise you.
A recent report from **Forbes** reveals a concerning trend: AI companies are actively **buying millions of physical antique books**, scanning them for data to train their large language models, and then **discarding or destroying them** afterwards. This practice is driven by the insatiable demand for diverse and high-quality textual data, which is crucial for improving AI model comprehension, creativity, and accuracy.
While seemingly efficient for data acquisition, this raises significant ethical and cultural preservation questions. Many of these books are rare or unique, and their destruction represents a permanent loss of physical cultural artifacts. The push for data acquisition at any cost highlights the immense resources and methods AI companies are employing to advance their technology, often with unforeseen consequences for other sectors.
What it means for you: To train AI, companies are **permanently destroying valuable historical books**, raising questions about data sourcing ethics and cultural preservation.
🩺 Predicate Uses Voice & Vital Signs to Spot Health Risks
New Launch
Your voice could hold the key to your health.
Doctor-turned-entrepreneur Morris Nguyen has launched **Predicate**, a new platform that leverages **voice analysis and vital signs** to identify potential health risks. The startup, based in Wilmington, North Carolina, believes that subtle changes in a person's voice, combined with other physiological data, can provide early warnings for various medical conditions, offering a non-invasive way to monitor well-being.
This innovative approach to healthcare aims to make proactive health management more accessible. By continuously analyzing these unique health signals, Predicate hopes to empower individuals and medical professionals to detect issues earlier, potentially leading to more timely interventions and better health outcomes. It showcases the growing potential of AI to transform diagnostics and preventative care.
What it means for you: A new AI tool called **Predicate** uses your **voice and other health data** to help spot health problems early, offering a new way to monitor your well-being.
🎬 Hollywood cuts first-ever AI copyright deal with TikTok’s ByteDance
Regulation
The entertainment world just set a major precedent for AI content.
In a landmark agreement, **Hollywood** studios have signed the **first-ever AI copyright deal with ByteDance**, the parent company of TikTok. This groundbreaking pact follows a **February cease-and-desist letter from Disney**, which accused ByteDance of unauthorized use of characters from franchises like Star Wars and Marvel. The deal establishes new guidelines and compensation frameworks for the use of intellectual property in AI-generated content, particularly on platforms like TikTok.
This agreement marks a crucial step in defining the legal and financial landscape for AI in the creative industries. As AI tools become more sophisticated at generating images, videos, and music, concerns over intellectual property infringement have soared. This deal provides a template for future negotiations, signaling a shift towards formalizing how creators and copyright holders will interact with AI companies and platforms.
What it means for you: **Hollywood and ByteDance** have signed a **first-of-its-kind deal** about AI and copyright, meaning new rules for how AI can use famous characters and stories.
📈 Corporate America’s Top 1% Spend $7,400 Per Employee on AI
Business
The AI spending gap is widening in corporate America.
New data from the **Ramp AI Index**, released on August 12, highlights a stark disparity in AI investment across U.S. businesses. The **top 1% of companies are spending a median of $7,400 per employee on AI** in July alone, while the top 10% allocated around **$650 per employee**. In contrast, the median company spent a mere **$11.95 per employee** on AI, showcasing a significant investment gap.
This data suggests a widening chasm between AI leaders and laggards. While a select group of businesses is heavily investing to gain a competitive edge through advanced AI integration, the vast majority are either cautiously experimenting or falling behind. This trend could lead to a two-tiered economy where AI-savvy enterprises achieve disproportionate growth and efficiency, leaving others struggling to catch up.
What it means for you: The **richest companies are spending thousands per employee on AI**, while most others spend very little, creating a huge gap in who benefits from AI.