A report by the Wall Street Journal said Meta had repeatedly delayed plans to make the API available to developers and, as of Tuesday, had not set a launch date. The report cited people familiar with the matter.

However, Meta disputed suggestions that the project had stalled. A company spokesperson said on Wednesday that testing is already underway with a group of early partners and that the company still expects to release the API later this month.

“The muse spark API will be coming soon,” Meta AI Chief Alexandr Wang announced in a post on X in April.

Meta AI Unveils Spark to Power Next-Gen AI across Platforms

The API would allow developers to integrate Muse Spark into their own software and services. An API, or Application Programming Interface, is a software bridge that enables different systems to communicate and work together.

Meta introduced Muse Spark in April as the first model developed under its Superintelligence Labs initiative, which was created to strengthen the company’s position in the competitive AI market.

The model is designed to narrow the gap between Meta and competitors including OpenAI, Anthropic and Google.

While Muse Spark is already available to consumers through Meta’s applications, users can currently access it only through built-in modes such as Instant, Thinking and Contemplating. Developers still do not have access to a public API, and Meta has yet to release documentation, pricing details, rate limits or eligibility requirements.

The lack of information has created apprehension among developers hoping to build products around the model. Without a public timeline, waitlist or technical documentation, companies interested in integrating Muse Spark are unable to plan deployments or assess costs.

The delays also come at a sensitive time for Meta. Investors have been monitoring the company’s AI strategy as it spends heavily on infrastructure, talent and product development.

Questions about execution have grown following reports of an Instagram security incident involving Meta’s AI-powered support system, which exposed weaknesses in automated account management processes.

Earlier on Wednesday, Meta unveiled a new AI agent designed to help businesses handle day-to-day tasks, showing that the company is going beyond consumer chatbots and into enterprise services.

The launch highlights Meta’s goal to compete more directly with OpenAI, Anthropic and Google across multiple areas of the AI market.

Muse Spark is expected to bolster that strategy. It is the first in what Meta has described as a new generation of advanced models from its Superintelligence Labs unit.

However, the repeated postponements have left analysts, developers and investors waiting for evidence that the company can translate its AI investments into products that are ready for global use.

Access is still currently limited to a small group of testing partners, while the developer community is waiting for Meta to open the platform to the public.

The company said it will start using internal detection systems from May 2026 to identify videos containing what it described as “significant photorealistic AI” content.

When creators fail to disclose that material, YouTube will now add the label automatically.

The update also changes where viewers see those warnings. Instead of hiding them inside video descriptions, YouTube will place labels directly below long-form videos and over Shorts, making them easier to spot.

YouTube has required creators since 2024 to disclose content made with AI tools when videos could realistically be mistaken for real people, places or events. However, content that was clearly fictional, animated or unrealistic did not need the same treatment.

Now, the company says it wants a more reliable system as AI video tools become harder to distinguish from real footage.

“We’ve heard consistently from our community that they value transparency when it comes to generative AI content,” YouTube said.

“That’s why since 2024, we’ve been labeling content when creators disclose they’ve used AI tools.”

The platform said the policy itself has not changed, but enforcement is becoming more active as AI-generated video quality improves.

The announcement follows the launch of Google’s Gemini Omni models at the company’s developer conference last week. Google said the models can generate highly realistic videos while showing an understanding of subjects including physics, science, history and culture.

Under the new system, creators will still be expected to disclose AI-generated content themselves. However, YouTube explained it will step in when its systems detect realistic AI content that has not been labelled.

“If a creator doesn’t specify whether or not they used AI, but our systems detect significant photorealistic AI use, we will now automatically apply a label,” the company said.

Creators who believe their content was wrongly flagged will be able to update the disclosure status through YouTube Studio. Still, YouTube said labels will remain permanent in some situations.

That includes videos produced using YouTube’s own AI tools such as Veo and Dream Screen. The same applies to videos carrying C2PA metadata showing they were fully generated with AI systems.

C2PA is an industry standard designed to help identify AI-generated and digitally altered media. Companies including OpenAI, Nvidia, Kakao and Eleven Labs have backed the standard in recent months.

YouTube is also changing how labels appear across the platform.

For long-form videos, labels will now be directly below the video player and above the description section. On Shorts, viewers will see them as overlays on the video itself.

The company said labels for unrealistic or lightly edited AI content will still appear only inside the expanded description section.

“By moving these labels on to the main stage, viewers get the context they need at a glance,” YouTube said.

The changes align with YouTube’s expansion of other tools aimed at detecting manipulated content. The company recently increased access to its AI deepfake detection system, allowing adults to scan the platform for videos that may contain their likeness.

At the same time, YouTube continues adding AI features across its services, including AI-generated video summaries, playlist tools for YouTube Music, interactive search functions and creation tools for creators.

Despite the labelling system, YouTube said the presence of an AI label will not affect recommendations or whether creators can make money from their videos.

“It’s important to note that a disclosure label alone does not change how a video is recommended or whether it’s eligible to earn money,” the company said.

The company announced the move on Wednesday during the ATx Summit in Singapore, where it also launched “OpenAI for Singapore”, a partnership with the country’s Ministry of Digital Development and Information (MDDI).

Under the initiative, OpenAI said it will commit more than S$300 million to Singapore and create about 200 technical roles over the next few years.

The company added that Singapore will become one of its global bases for Forward-Deployed Engineers, teams that work directly with businesses and public institutions to deploy AI systems.

The new lab will support projects tied to Singapore’s national AI priorities, especially in public services, healthcare, finance and digital infrastructure.

Denise Dresser, chief revenue officer at OpenAI, said the company sees Singapore as a key market because of its technical talent and long-term AI ambitions.

“We’re excited to partner with Singapore as it builds on its position as a global leader in AI,” she said.

“Singapore has strong technical talent, trusted institutions, and a clear ambition to use AI to drive long-term growth and improve people’s lives.”

She added: “Through OpenAI for Singapore, we want to help more organisations benefit from frontier AI, support the next generation of local AI talent, and widen access to these tools across the country.”

Singapore has spent the past few years positioning itself as a neutral and trusted centre for AI development in Asia. The government has steadily increased spending on AI research and infrastructure while encouraging global technology firms to expand operations in the country.

Authorities earlier pledged S$1 billion between 2025 and 2030 to strengthen public AI research capabilities. Tech giants including Google, Nvidia, AWS and Microsoft have also announced AI-related investments and partnerships in Singapore.

Alongside the OpenAI AI Lab deal, Singapore recently unveiled a National AI Partnership with Google focused on education, healthcare and enterprise innovation. Nvidia is also establishing a new AI research lab in the country to work with universities and government agencies.

The partnership with OpenAI will also include education and workforce programmes. OpenAI said it plans to work with Singapore’s Ministry of Education and GovTech on AI-powered learning tools, including support for Mother Tongue language learning.

The company will also launch a Singapore chapter of the OpenAI Academy, organise Codex hackathons for teachers and introduce a training programme for Forward-Deployed Engineers.

Singapore’s Permanent Secretary for Digital Development and Information, Chng Kai Fong, said the partnership shows the government’s drive to prepare its workforce and economy for AI adoption.

“With AI reshaping economies, businesses and the workforce, Singapore’s response has been deliberate: growing new sectors, anchoring global frontier companies here, and equipping our people with the skills to thrive in this new environment,” he said.

“This partnership with OpenAI reflects the Government’s commitment to developing Singapore’s AI capabilities, strengthening enterprise adoption of AI, and securing good jobs for Singaporeans.”

OpenAI said it also plans to support smaller businesses and startups through workshops, accelerator programmes and practical AI adoption initiatives.

Countries are currently competing to attract AI investment, talent and infrastructure. Singapore is not left out, standing alongside hubs such as London, Dubai and Silicon Valley to lead AI development.

Recent data from Slack’s Workforce Index showed that about 52% of workers in Singapore already use AI tools in their jobs, underlining how quickly adoption is spreading across the country’s economy.

The company said the feature works through a new command-line interface tool, now in beta, which connects Spotify with external coding agents such as OpenAI’s Codex, Anthropic’s Claude Code and OpenClaw.

Users can generate private audio briefings or podcast-style content from notes, schedules, articles and other personal material, then listen to them inside the Spotify app.

Spotify also explained that the podcasts will remain private and do not appear for other users on the platform.

“People are already starting to use their agents to create personal audio that guides their day: from summaries of class notes before an exam to briefings of what’s on their calendar. And they’re asking for a way to listen to it on Spotify, where they already listen to everything else,” the company said in a blog post.

The tool lets users create different types of audio content depending on what they need. Someone preparing for work can generate a short morning briefing with meeting reminders, weather updates and podcast suggestions for their commute.

A student studying philosophy could also create weekly audio lessons based on saved notes, articles and research material.

Users who already work with coding agents on desktop can install the beta tool through its GitHub page. After signing into Spotify through a browser, they can write prompts describing the kind of podcast they want and ask the agent to save it directly to their Spotify account.

For example, a user could request an audio session explaining the history of the World Cup, including major players, host countries and details about this year’s tournament. The system then creates the podcast and adds it to the user’s library with a Spotify link for playback.

The company further noted that the feature is available to eligible Free and Premium users globally, although limits may apply during the testing period.

Spotify plans to continue adjusting the personal podcasts experience as it receives feedback from listeners and expands the beta programme.

From what has emerged so far, Nvidia is weighing the largest cheque. The chipmaker, already central to OpenAI’s operations, is considering an investment of up to $30 billion. 

Microsoft, which has backed OpenAI from the start, is discussing a much smaller top-up of under $10 billion. Amazon, entering the picture for the first time, could commit well above $10 billion and possibly exceed $20 billion.

The discussions are said to be advanced. Term sheets are close, implying the talks have moved beyond early signalling into concrete commitments. None of the companies has confirmed the details. 

Amazon and Microsoft declined to comment, while Nvidia and OpenAI did not respond outside normal business hours.

This would form the core of a funding round that could reach $100 billion. That figure alone changes the scale of the story. At that level, OpenAI’s valuation would rise to around $830 billion, placing it ahead of every private company globally and within touching distance of the largest names on the S&P 500.

SoftBank is also circling. The Japanese group is reportedly in discussions to add up to $30 billion of its own, which would further tilt the round into record territory. 

Nvidia’s involvement helps lock in demand for its chips, which remain essential for training and running large models. Amazon’s interest goes beyond a simple equity stake. 

Its investment is tied to separate negotiations, including a possible expansion of OpenAI’s use of Amazon’s cloud servers and a commercial deal that would see OpenAI’s products sold through Amazon’s enterprise channels. 

For Microsoft, any fresh funding would strengthen an already tight integration between OpenAI’s technology and its own software and cloud services, even as regulators watch.

Behind the OpenAI funding round, the cost of training and operating large models keeps climbing, running into billions of dollars each year. Competition is also strengthening. 

Alphabet is pushing hard with its own systems, and competitors are striking partnerships to close the gap. 

The capital would give OpenAI room to expand and invest aggressively, but it also accentuates how much cash the business burns to stay at the frontier. 

It was the result of several innovations inside the app economy that finally lined up at once.

The $85 billion spent globally on non-game apps last year did not come from a surge in new users, because we see that downloads across mobile are largely reduced. 

But then, time spent has stabilised. So what changed was how people pay, and why they keep coming back.

Building habits

For years, while reach or downloads were used to describe how successful mobile apps were, games thrived because they could attract millions of casual players, monetise a small fraction of them, and repeat the cycle. Non-game apps didn’t match that efficiency before 2025.

That gap has now closed. Generative AI apps flipped the model and instead of focusing on new installs, they focused on becoming useful enough to open daily, sometimes dozens of times a day. 

The result is visible in Sensor Tower’s latest State of Mobile findings, which show global app spending rising 21% year-on-year. Sessions in AI apps crossed one trillion in 2025, growing faster than downloads. That tells us engagement is now the main engine.

This is a shift from scale-first to habit-first design.

Why AI assistants won, not just AI tools

Not all AI apps benefited equally. Assistants took over because they helped with multiple needs. Writing, search, coding help, image creation, planning, all in one place. That breadth reduced churn and increased willingness to pay.

ChatGPT’s $3.4 billion in in-app revenue is less important than how quickly it got there. No app has crossed $3 billion in annual consumer spending this fast. 

That speed is commendable because it shows that users accepted subscriptions and premium tiers without years of conditioning.

Others followed, with Google, Microsoft and X not just building similar features, but embedding assistants into daily workflows. Image and video generation became turning points, not side features. Once users could create, not just ask, time spent jumped.

Big tech’s return reshaped the field

Early AI growth came from smaller, fast-moving developers. That phase is over.

By 2025, OpenAI and DeepSeek controlled nearly half of all AI app downloads. Large technology firms expanded speedily, taking close to a third of the market. Together, they crowded out earlier competitors who lacked capital, distribution, or ecosystem access.

This concentration shows that AI on mobile is entering a maturity phase faster than previous app categories. Winners are pulling away early, leaving limited room for mid-tier challengers.

Mobile became the default AI gateway

One of the most underappreciated findings in the data is where AI usage happens.

More than half of AI assistant users in the United States now access these services only on mobile. A year earlier, that group barely existed. Phones are no longer secondary screens for AI, they are the main ones.

This has implications beyond apps. It explains why voice, camera input, and real-time image generation are advancing so quickly. Mobile limitations forced AI products to become faster, simpler, and more responsive.

Games did not collapse, they were overtaken

It is tempting to describe this as a loss for gaming. It isn’t.

Games still generate enormous revenue and attention. But their growth has slowed as user acquisition costs rose and playtime competed with social media, streaming, and now AI. Meanwhile, non-game apps learned how to monetise without friction.

Subscriptions, tiered access, and clear value exchanges worked. Users paid because they understood what they were getting back; saved time, better output, or creative control.

What this means for 2026

The mobile market has entered a monetisation-first era. Growth will not come from more downloads but from better use, clearer value, and products that are used in daily routines.

AI went beyond adding a new category to resetting expectations across the app ecosystem. Productivity, creativity, and even entertainment apps are now judged by how quickly they produce results, not how long they keep users scrolling.

This follows reports that CEO Sam Altman declared a “code red” inside the company in early December, halting side projects and pulling teams into a faster development sprint. 

The urgency was linked to Google’s latest innovations, which had placed Gemini 3 at the top of key performance rankings across reasoning, coding and multimodal tasks.

OpenAI says GPT-5.2 brings stronger general intelligence, better coding results, and far longer context handling. The company believes these improvements will help users complete more demanding work, particularly tasks that involve spreadsheets, complex documents, and project-heavy workflows. 

Interestingly, the new model stretches to handle up to a million tokens, a big difference from the previous model.

Google has been keen to highlight what Gemini 3 is capable of across text, audio, images and video, and analysts say its tight integration with Workspace and Android gives it an advantage with corporate users. 

Even with that, Altman played down the internal panic when he spoke on CNBC, saying: “Gemini 3 has had less of an impact on our metrics than we feared.” Google has not responded to requests for comment.

OpenAI is rolling out GPT-5.2 in three versions: Instant for quick responses, Thinking for slower but more reasoned answers, and Pro for enterprise-level performance. Paid ChatGPT users will receive them first. The company also states it will continue to support GPT-5.1, GPT-5 and GPT-4.1 on its API, giving developers more flexibility.

Away from the technical competition, OpenAI is also moving into entertainment. Disney has confirmed a $1 billion investment in the company and will allow its Sora video generator to use characters and worlds from Star Wars, Pixar and Marvel. 

This is one of the largest licensing deals yet between Hollywood and an AI firm, and it sets up OpenAI as a direct partner in digital content production. Microsoft, still OpenAI’s biggest backer with about $13 billion committed since 2019, continues to host the company’s models on Azure.

Industry forecasts show spending on cloud-based AI services is expected to rise sharply, with Gartner estimating it will exceed $723 billion next year. Many companies are already relying on GPT models for coding assistance, document processing and data insights. According to OpenAI, enterprise usage has climbed roughly 40% in the past year.

However, regulators in the US and Europe are examining safety standards, competition risks and copyright issues, with Disney’s licensing deal likely to draw even closer attention.

In Africa, approximately 40% of organisations are either experimenting with or deploying generative AI tools.

This adoption is already yielding measurable success: 51% of South African businesses believe generative AI has improved productivity and competitiveness.

Governance: The foundation for reliable innovation

To ensure this growth is reliable and responsible, organisations must build a foundation of trust. AI runs on data, and data runs on trust. Building a healthy data culture involves knowing what information is held, where it lives, who can use it, and for what purpose.

This is where governance comes in, providing structure and discipline. Governance establishes the standards and controls necessary to ensure information accuracy and security, as well as the accountability to uphold them.

Crucially, when governance works as it should, it doesn’t slow innovation – it makes it safer and faster. Clear rules give businesses the confidence to move quickly, use data creatively, and make better decisions.

African nations are proactively establishing strong legal foundations for trust in the digital space. For instance, South Africa mandates that boards are held responsible for managing data risk and ensuring lawful usage through its  Protection of Personal Information Act (POPIA) and the King IV Code. Similarly,  Nigeria’s Data Protection Act demands essential principles like transparency, consent, and human oversight in data handling. Meanwhile, in Kenya, the Ministry of Information, Communications and the Digital Economy (MICDE) is the primary governmental body actively driving cybersecurity and AI governance efforts.

Strategic security for sustainable growth

Africa’s AI market is projected for significant expansion, expected to reach around $6.4 billion by the end of 2025, supported by more than 2,400 AI-focused companies. This growth won’t be sustainable without strong foundations.

While the technology that makes work smarter can also be used by cybercriminals (e.g., forging images or cloning voices), organisations are implementing strategic solutions to mitigate these risks. Boards and executives must look at information governance as a strategic priority, not merely a technical one.

Technology offers robust support for this strategic focus: platforms can automate data protection, monitor activity, and simplify compliance, helping with the heavy lifting of security.

Empowering the human element

Cybersecurity, data governance, and training must all work together to maintain a secure system. Employees are essential, remaining the first and last line of defence.

Top down organisational culture is key to empowering employees with necessary skills. These essential skills include recognising manipulated voices, spotting deepfakes, avoiding suspicious links, and questioning urgent payment requests.

The safest way to work with AI is to treat data with the same care afforded to money or reputation. By integrating rules, oversight, and discipline, organisations keep the system honest.

The principle guiding this growth remains clear: progress is nothing without trust.

The new platform, developed in collaboration with NVIDIA, focuses on securing AI models, workloads, and agentic applications used in enterprise environments, without compromising performance.

The company confirmed that AI Cloud Protect is now available for on-premises enterprise use and has been validated on NVIDIA RTX PRO Servers. 

Built on NVIDIA’s BlueField technology, it enables organisations to secure AI model development and inference workloads both in data centres and cloud environments.

“As enterprises race to build AI-driven innovation, they can’t afford blind spots,” said Nataly Kremer, Chief Product Officer at Check Point. “With NVIDIA, we’re making AI factories secure by design—protecting models, data, and infrastructure without slowing innovation.”

The rise of AI has exposed enterprises to new and complex risks. According to Check Point data, one in every 80 generative AI prompts reveals sensitive information, while Gartner reports that nearly one-third of organisations suffered an AI-related security incident in the past year, ranging from prompt manipulation to infrastructure attacks.

AI Cloud Protect was built to address these vulnerabilities head-on. Running on NVIDIA BlueField-3 DPUs, it delivers full-stack protection without consuming CPU or GPU resources, putting an end to the common trade-off between security and performance. 

Its network-level defence prevents unauthorised access, data poisoning, and model exfiltration. At the host level, it leverages NVIDIA’s DOCA Argus framework for direct memory access, offering full visibility into active processes on AI nodes to detect and block malicious workloads, even within downloaded large language models.

“Security is essential for the next generation of AI infrastructure,” said David Reber, chief security officer at NVIDIA. “NVIDIA is working with Check Point to integrate BlueField acceleration and the NVIDIA DOCA Argus runtime security framework into the AI Cloud Protect platform to help enterprises deploy AI confidently.”

The solution will also extend to NVIDIA’s upcoming BlueField-4 DPU, which promises six times more compute power and double the network throughput, setting the stage for faster and more scalable AI operations.

Beyond AI Cloud Protect, Check Point’s AI security portfolio includes CloudGuard Web Application Firewall (WAF) with Lakera integration, designed to block prompt injection and jailbreak threats in enterprise AI applications. 

Another solution, Infinity GenAI Protect, helps organisations monitor generative AI usage, apply policies, and prevent sensitive data exposure across teams. Together, these solutions aim to deliver end-to-end protection across the AI ecosystem, from infrastructure to user level.

The company is currently piloting AI Cloud Protect with select financial services firms and partners like World Wide Technology (WWT), focusing on protecting data centres supporting large language model development.

“As enterprises build AI server factories at scale, the combination of Check Point’s AI Cloud Protect and NVIDIA BlueField acceleration delivers enterprise-grade protection for sensitive AI workloads from model training to inference without compromising the performance modern AI applications demand,” said Chris Konrad, vice president, Global Cyber, WWT.

Check Point is addressing data leakage, model manipulation, and infrastructure threats, while enabling organisations to innovate confidently in this phase of intelligent computing.

The company’s growth is being driven by surging enterprise adoption and a portfolio of cost-efficient AI models.

According to internal projections cited by two people familiar with the company’s financial outlook, Anthropic aims to close 2025 with an annualised revenue run rate of $9 billion, up from $5 billion in August and $7 billion this month. 

Per Reuters, management is targeting a base case of $20 billion, with an upper target of $26 billion for 2026, trusting in sustained enterprise demand.

The San Francisco-based company confirmed that revenue growth is accelerating but declined to comment on future targets. Anthropic and many other businesses are embedding generative AI tools into their operations, even though the sustainability of industry-wide investment levels is being questioned.

A key contributor to this growth is Anthropic’s enterprise product line, which now serves over 300,000 business and corporate users, a segment that contributes roughly 80% of total revenue. 

Its Claude Code product, a coding assistant introduced earlier this year, has already achieved a $1 billion annualised run rate, showing the appetite for automation tools that improve developer productivity.

In a move to reach more cost-sensitive clients, Anthropic recently launched Claude Haiku 4.5, the latest version of its smallest AI model. The model is priced at about one-third the cost of Sonnet 4, making it attractive for companies seeking high performance without steep infrastructure costs. 

Haiku 4.5 is optimised for low-latency applications, rapid prototyping, and multi-agent systems, core needs for businesses deploying AI at scale.

This expansion boosts Anthropic as one of OpenAI’s closest competitors. OpenAI’s annualised revenue exceeded $13 billion in August, with projections of reaching $20 billion by year-end, largely powered by ChatGPT’s 800 million weekly active users.

Investor enthusiasm has kept pace with Anthropic’s financial course. In a Series F funding round led by ICONIQ, the company raised $13 billion, pushing its valuation to $183 billion, a jump from $61.5 billion just six months earlier. 

Major backers include Google’s parent Alphabet and Amazon, both of which have strategic partnerships with the company.

Anthropic is also looking beyond the U.S. In August, the firm announced plans to offer its Claude model to the U.S. government for $1, a symbolic gesture signalling its commitment to public sector collaboration. 

The company also revealed plans to open an office in Bengaluru, India, in 2026, its first in Asia and second-largest market after the U.S. To support this expansion, Anthropic intends to triple its international workforce and grow its applied AI team fivefold this year.